AI Applications

A major Spanish bank has made significant strides in improving customer service by leveraging artificial intelligence. BBVA’s contact centers in Italy and Germany have seen a more than 15% reduction in average handling times for common customer inquiries, thanks to the implementation of generative AI assistants.

The two fully digital banks’ teams developed these specialized tools using BBVA Assistants, the Group’s corporate platform for creating and scaling generative AI. This platform enables employees to build customized assistants that can transform business processes by combining OpenAI models with proprietary orchestration architecture.

BBVA customer service agents in Italy and Germany previously spent considerable time consulting multiple internal documents and product manuals specific to each market when responding to nearly 100,000 customer inquiries per month via phone, email, or the bank’s messaging channel. This process often required searching through various sources to find relevant information for each case.

To address this challenge, BBVA’s Digital Banks teams identified an opportunity to create AI assistants that would bring together all necessary knowledge in a single place. Agents can now submit customer requests in natural language, and these AI copilots generate structured response proposals within seconds, which are then reviewed and adapted by the agents before being sent to customers.

The tool has been used daily by over 260 customer service agents across both countries without accessing sensitive data. It is designed primarily for answering frequently asked questions about products, services, and general procedures, such as increasing transfer limits or tracking card deliveries. However, when a response depends on specific circumstances, the agent still analyzes the case and consults internal systems to provide personalized support.

The AI assistants were developed using BBVA Assistants, which enables employees to create specialized tools that can streamline business processes. This platform combines OpenAI models with proprietary architecture for access control, knowledge management, and supervision mechanisms that ensure compliance with security and governance standards.

In June alone, customer service agents in both countries exchanged over 40,000 messages with their AI copilots while preparing responses to customer inquiries. This makes them among the most widely used assistants on BBVA Assistants, demonstrating how bringing generative AI capabilities closer to business teams can lead to tangible improvements in customer experience.

The development and implementation of these AI assistants were guided by BBVA’s AI Assistant Governance Framework, an internal governance model that establishes supervision and control measures for all AI assistants developed within the bank. This framework includes human oversight mechanisms, reviews of prompt quality and knowledge sources used, as well as impact assessments to mitigate risks associated with generative AI.

The framework ensures that these specialized tools are designed to alert agents about potential data usage issues and incorporates technical controls to prevent inappropriate or unsupported information from being generated. All outputs are reviewed before sharing them with customers, further ensuring the accuracy of responses.

Training has also been crucial in mitigating these risks. Customer service agents receive dedicated training sessions where they learn how to formulate prompts effectively, provide context, and make the most of the assistants’ capabilities. This training contributes to more accurate responses, improves user experience for agents, and accelerates adoption of the solution.

The success of this project is attributed to its bottom-up approach, with BBVA’s customer-facing teams identifying a challenge in their daily work and leveraging AI to turn that idea into a tangible impact on both business operations and customer satisfaction. According to Elena Alfaro, Head of Global AI Adoption at BBVA, ‘the need came directly from the teams who interact with our customers every day.’

The implementation of these generative AI assistants marks a significant improvement in customer service for BBVA’s two fully digital banks. By speeding up information retrieval and response preparation, these tools enable agents to dedicate more time to complex cases that require deeper analysis and personalized support.

Google DeepMind researchers have unveiled a family of AI models designed to power humanoid robots. These new series, called Gemini Robotics 2, allows multiple autonomous machines to collaborate on complex tasks involving hundreds of steps.

The Gemini Robotics 2 series is built to automate chores that demand human-like dexterity and problem-solving skills. Humanoid robots usually rely on a dual-system AI architecture with two separate models working together: one crafts the high-level plan for how to perform a task, while the other translates those instructions into low-level commands for the host robot’s motors.

The new Gemini Robotics 2 series boasts an improved embodied reasoning algorithm called Gemini ER 2. This tool lets users describe tasks in natural language so humanoid robots can understand and execute complex actions even when they involve hundreds of steps or take several minutes to complete.

One key feature of ER 2 is its ability to split lengthy chores among multiple robots speeding up task completion. Users also have a tool-calling feature that gives them access to external cloud services like Google Search helping the model clarify parts it doesn’t understand when interpreting tasks.

When tasks don’t go as planned, humans and robots alike need to redo them without manual intervention. To address this challenge DeepMind’s engineers equipped ER 2 with two features: it can track task progress using footage from the host robot’s cameras and identify mistakes made along the way.

If a mistake is made ER 2 identifies the last step completed correctly and picks up where it left off allowing tasks to resume smoothly. This means robots have more time for other work and don’t get stuck on one task forever.

Gemini ER 2 also makes AI tools more efficient at determining when a task is complete enabling humanoid robots’ AI to tick off tasks faster. With this capability, robots can move on to the next step sooner and adapt in real-time situations.

Google engineers Steven Hansen and Peng Xu explained their work ‘By watching continuous video feeds robots can now track their own progress adapt if something goes wrong and know exactly when to move on to the next step.’ Their expertise has improved humanoid robot capabilities through ER 2’s integration of AI making a significant impact in this field.

After ER 2 generates a plan for task execution it sends instructions to one of two other models in the Gemini Robotics 2 series. These VLA algorithms translate action plans into low-level instructions for the host robot: the first model is called Gemini Robotics 2 and has several notable improvements over its predecessor.

Gemini Robotics 2 can control all components of a humanoid robot not just its hands allowing it to optimize the center of gravity minimizing falls in the process. This algorithm supports a broader range of robotic hands than earlier software from DeepMind giving robots more flexibility and precision when interacting with their environment.

The second VLA model is called Gemini Robotics On-Device 2 and runs directly on humanoid robots’ onboard computers. Google says this model can be adapted to new robots in just a few hours of training making the integration process smoother for developers who want to implement ER 2 in their projects.

Developers have several options when it comes to accessing ER 2: they can use Google Cloud the Gemini API or Google AI Studio. The company has also rolled out a new embodied AI safety benchmark called ASIMOV-Agentic Benchmark which evaluates human robots’ ability to avoid collisions and other risks.

This benchmark will help developers refine their models to ensure safe interactions between humans and machines. Safety is an essential consideration in robotics development, and this move by Google DeepMind sets a high standard for the industry as a whole.

Scheduling the operations of agile earth observation satellites has become a complex task, requiring sophisticated algorithms to optimize resource allocation and minimize costs. Researchers have now developed an innovative approach that leverages Large Language Models (LLMs) to automate the design of critical heuristics for this process.

The new method, called evALNS, uses LLMs to generate novel and efficient scheduling strategies, which can significantly outperform traditional algorithms in complex scenarios. This is particularly important for tasks like environmental monitoring and resource management, where accurate and timely data are crucial.

However, the researchers recognized that LLMs lack the specialized domain knowledge required for such a complex problem. To address this limitation, they introduced an innovative Dual-Population Co-Evolutionary Computing Framework (DPEC). This framework sets up two evolving populations of heuristics: one generated by the LLM and another seeded with proven, expert-designed rules.

The DPEC framework allows the LLM to learn from the expert strategies and fuse that knowledge into its own creative solutions. Additionally, a Functional Natural Language Embedding (FNLE) strategy is used to help the LLM better understand the logic and purpose of the heuristics it evolves.

Extensive experiments showed that the evALNS algorithm, powered by the DPEC framework, significantly outperformed traditional Adaptive Large Neighborhood Search (ALNS) using only expert-designed operators. The new approach achieved an average profit improvement of 8.48% across various complex scenarios.

This work not only provides a more effective solution for satellite scheduling but also establishes a powerful new paradigm for using LLMs to automate the design of sophisticated algorithms for other complex real-world optimization problems.

VC3, a managed IT and cybersecurity services provider, has announced the launch of its purpose-built artificial intelligence (AI) assistants designed specifically for municipalities. The new offering aims to help local governments accelerate service delivery and operational efficiency by providing a structured way to deploy, use, and govern AI across internal operations and public-facing services.

The VC3 managed AI offering is tailored to meet the unique needs of municipal organizations, which often require specialized tools that can be easily integrated into their existing systems. According to David Bridges, President of VC3, ‘municipal leaders need AI that fits how local governments actually operate, not generic tools that require significant configuration or internal expertise to manage.’

The offering includes four purpose-built municipal AI assistants: the Council Meeting Minutes Assistant, Policy Assistant, and New User Onboarding Assistant. These tools transform how staff access information and complete daily tasks internally by providing instant answers to common questions and streamlining administrative processes.

For instance, the Council Meeting Minutes Assistant enables staff to quickly search decades of council meeting records, while the Policy Assistant helps navigate policies and bylaws. The New User Onboarding Assistant simplifies the process of onboarding new employees by providing essential information in seconds rather than hours.

Municipalities also benefit from a website AI assistant deployed directly onto their public website, giving residents instant access to answers about permits, bylaws, services, and contacts. This reduces inbound call volume and service requests, allowing staff to focus on more critical tasks.

Secure Chat is another key component of the VC3 managed AI offering. It provides municipal employees with secure access to multiple leading AI models through a single environment while addressing governance and data privacy risks associated with unsanctioned AI tools used for official work.

The Secure Chat solution enables faster drafting, research, summaries, and handling of complex multi-step administrative tasks by giving staff an approved and governed alternative. This eliminates the need for security tradeoffs and reduces the risk of sensitive data exposure or compliance concerns.

Every engagement begins with a comprehensive AI readiness assessment that provides municipalities with actionable recommendations on where they’re ready to implement AI and where gaps may exist. This enables them to move forward securely and confidently, leveraging the full potential of their new AI tools.

The VC3 managed approach ensures that AI is not only deployed effectively but also supported over time through ongoing guidance, oversight, and optimization built into the service. The company’s experienced team provides expert support to ensure municipalities get the most out of their AI investments.

VC3’s managed AI offering is now available to municipal clients, with additional enhancements planned for later in the year. These include expanded website and AI assistant capabilities that will further enhance the efficiency and effectiveness of local government operations.

For more information or to schedule a demonstration, interested parties can visit VC3’s official website. With its comprehensive range of services, including managed IT, cybersecurity, and AI solutions, VC3 is well-positioned to support municipalities in their digital transformation efforts.

A new era of in-car experience is arriving for Waymo riders, thanks to the introduction of Gemini, a conversational AI assistant designed to make every journey more productive and personalized. This innovative feature is part of the Ojai’s redesigned cabin, which reimagines the vehicle as a flexible, premium space for riders to enjoy.

The Ojai was created from scratch with ride-hailing in mind, offering a unique blend of comfort and functionality. As Waymo prepares to welcome more public riders on board, it has unveiled new features aimed at enhancing every aspect of the journey. At its core is Gemini, an AI assistant that empowers riders with cabin controls, journey information, and everyday knowledge at their fingertips.

Riders can now interact with Gemini using voice commands or by tapping on the screen icon. This allows them to personalize their ride, access a wealth of information about the world around them, and even control various aspects of the vehicle’s environment. For instance, they can ask Gemini to adjust the temperature or find nearby attractions.

However, it’s essential to note that Gemini operates independently from the Waymo Driver, which remains in charge of navigating the physical world outside the vehicle. This separation ensures rider safety while providing an additional layer of convenience and control within the cabin.

One of the key benefits of Gemini is its ability to provide riders with a wealth of knowledge at their fingertips. Whether they want to learn about local history or find the nearest coffee shop, Gemini’s live assistant makes it easy to access information without needing to manually search for it.

Rider privacy remains a top priority for Waymo, and Gemini is designed to remain inactive unless explicitly engaged by the rider. This means that riders have complete control over their interaction with the AI assistant, ensuring that their personal space is respected at all times.

The Ojai’s cabin has also undergone significant redesign, introducing an entirely new screen interface that supports a relaxed ‘lean-back’ ride experience. The dynamic screens display key information and streamline music playback, while an intuitive system menu makes it easy to adjust cabin controls or access media options.

One of the standout features of this redesigned interface is its ability to adapt to different riders and their needs. For example, if only one rider is in the right rear seat, they will see full controls on their screen, while other screens may display a simplified trip status view or ambient media player. This choreographed tri-screen experience allows for new capabilities like Calm Mode, which dims the screen and displays minimal trip information.

Waymo continues to refine its Ojai rider experience, with plans to enhance Gemini’s voice assistant functionality over time. As this innovative technology evolves, riders can expect even more personalized and productive experiences on board.

A new initiative by OpenAI aims to accelerate scientific discovery and research using its advanced AI tools. The company is offering free access to its frontier models, including GPT-5.6 Sol Pro, to 100,000 researchers at selected academic institutions. This move marks a significant step towards democratizing access to powerful data analysis tools for the research community.

The program, called ChatGPT for Academic Researchers, is part of OpenAI’s commitment to support external scientific research and discovery. The company has pledged over $250 million through 2027 to fund various initiatives, including NextGenAI, which supports research institutions, and its work with the Department of Energy’s Genesis Mission to bring frontier AI to researchers at national laboratories and universities.

OpenAI believes that scientific progress accelerates when more researchers have access to powerful tools. The company is putting its frontier models in the hands of 100,000 scientists, mathematicians, and engineers, allowing them to pursue their own ideas without being limited by resource constraints. This approach marks a shift towards empowering researchers rather than trying to solve all problems themselves.

The program will start with 10,000 researchers this summer, with access already available at institutions such as the Institute for Advanced Study (IAS) and École normale supérieure (ENS). OpenAI plans to expand to 100,000 researchers through 2027. Participants will have access to its frontier models, including GPT-5.6 Sol Pro, which is capable of tackling complex scientific and mathematical problems.

Researchers using these tools can expect a range of benefits, from improved productivity to accelerated discovery. They will be able to use more than 75 life science skills spanning genetics, genomics, sequencing, single-cell analysis, protein modeling, and drug discovery. Connectors support research across disciplines, providing access to scientific literature, public genomic and clinical databases, satellite imagery, computational notebooks, data platforms, and reference managers.

Codex can help write and debug code, analyze datasets, and build reproducible workflows. ChatGPT Work can support longer projects such as finding funding opportunities, preparing grant applications, reviewing literature, drafting manuscripts, and creating materials to communicate results. Researchers will have different starting points with AI, from using these tools for the first time to developing advanced applications in their fields.

The program offers training tailored to different levels of experience, from getting started and improving workflows to advanced research applications. Participants will have access to specialists familiar with research workflows, including help integrating these tools into their work. As researchers discover new insights, OpenAI plans to create more opportunities for them to share practical approaches, learn from one another, and highlight uses across disciplines.

Researchers using AI most intensively are also taking on more ambitious work. Those in the top 20% of AI usage within their field are almost twice as likely as their peers to ask AI to take on tasks estimated to require four hours or more: nearly 7% of their requests, compared with 3.5% among other researchers in the same field.

The initial program is open to qualifying researchers at selected academic institutions. Eligible institutions must be recognized, degree-granting colleges or universities with a high level of research activity. Applicants will need to verify their institutional affiliation and provide information about their active research and intended scientific use. Approved researchers may invite up to four collaborators from their institution.

Applications are open today for the ChatGPT for Academic Researchers program. Eligible institutions can review the eligibility criteria and apply here. OpenAI’s commitment to supporting external scientific research and discovery reflects its belief that scientific progress accelerates when more researchers have access to powerful tools.

Financial technology firm Intuit Credit Karma is using artificial intelligence to optimize recurring financial decisions, helping users make informed choices about their money. The company has developed a unique approach that pairs financial modeling with generative AI, resulting in personalized recommendations and explanations for consumers navigating complex financial situations.

In contrast to firms that treat debt, refunds, and paychecks as separate products and layer on generic chatbots, Intuit Credit Karma takes a more targeted approach. Its AI assistants are designed for specific moments, such as paying down debt, managing tax refunds, and allocating paychecks. These assistants help consumers navigate each decision by generating personalized recommendations and explaining the reasoning behind them.

The company’s Debt Assistant, Refund Assistant, and Paycheck Assistant work together to provide a comprehensive financial context for each consumer. Using this shared context, they generate tailored advice that takes into account individual circumstances, including debt profiles, balances, interest rates, monthly payments, and stated goals. This approach ensures that recommendations are mathematically sound and based on actual calculations, rather than relying solely on an LLM’s best guess.

Gurpreet Singh, Head of Product at Intuit Credit Karma, emphasizes the importance of getting the underlying technological foundation right when building AI-powered financial tools. He notes that in situations like debt consolidation, accuracy is paramount, as incorrect numbers can have serious consequences for consumers. To address this challenge, the company’s architecture splits the decision-making process into two distinct stages.

The first stage involves a deterministic financial optimization engine built on event-based simulations and linear optimization. This engine evaluates different scenarios to determine the mathematically optimal path for debt consolidation, guaranteeing that savings numbers are based on actual calculations rather than an LLM’s estimate. The second stage is powered by generative AI (GenAI), which provides personalized recommendations and explanations for consumers.

When asked why not let the LLM handle recommendations on its own, Singh explains that in situations where accuracy matters most, such as debt consolidation, relying solely on an LLM can be problematic. He notes that ‘the stakes are too high’ to risk getting numbers wrong, highlighting the importance of a robust and reliable financial optimization engine.

The company’s approach marks a significant departure from traditional AI-powered chatbots, which often fail to provide personalized advice due to their generic nature. By pairing financial modeling with generative AI, Intuit Credit Karma is able to offer consumers more effective support in navigating complex financial decisions.

Scientific computing is at the heart of modern research in academia and industry, but its software has struggled to keep pace with the rapid growth of data. Many widely used tools were initially developed as code accompanying research papers by small academic teams with limited engineering expertise and no time for packaging, testing, or long-term support. As a result, scientific infrastructure often relies on slow, fragile workflows that require constant maintenance, hindering the pace of discovery.

The situation is about to change thanks to AI agents like Codex and Claude Code. By reducing the costs associated with engineering work and taking over tedious implementation tasks, these agents can help researchers prototype ideas more quickly, pursue projects previously deemed impractical, and maintain software over time without breaking their stride. This shift enables scientific software to become more efficient and better maintained, freeing up researchers to focus on discovery.

The field report explores eight agent-assisted computing projects in the life sciences, five using Codex alone and three combining Codex with Claude Code. The case studies bring together contributions from teams behind each project, highlighting recurring themes and patterns. These projects range from routine maintenance and targeted optimization to large-scale language migrations and GPU-native redesigns.

One of these projects involved modernizing a widely used library for parsing genomic data called cyvcf2. GPT-5.5 replaced the legacy build and packaging system with a unified process designed to make installation, testing, and release easier. The result was a significant improvement in usability and maintainability.

Brent Pedersen noted that while agents can accelerate development, there’s still a need for expert guidance, understanding, taste, and care when it comes to scientific computing. He emphasized the importance of human oversight in validating an AI agent’s output, which often depends on human judgment.

Across case studies, agents handled specific requests effectively but struggled with judging whether their work was scientifically valid or met expectations. In fact, they sometimes expressed confidence even when containing clear errors. Human reviewers therefore needed to find reliable ways to validate the results, such as using an external reference or measurable acceptance target like exact output agreement.

Another recurring theme in these projects was that agents often produced initial implementations quickly but resolving edge cases and subtle numerical differences took much longer. Completing the ‘last mile’ of implementation required significant effort, highlighting the importance of iterative refinement rather than one-shot approaches.

The case studies suggest that agents are enabling researchers to spend less time on implementation and more on directing scientific work. People define goals, break down complex projects into manageable chunks, and judge whether results are scientifically valid. By easing engineering constraints, agents expand what researchers can build while freeing them to focus on the scientific questions and decisions that matter most.

The maintenance gap in research software has long slowed iteration and limited reproducibility and reliability. Published studies have found that published software often fails to properly install or run as documented, forcing researchers to spend substantial time on configuration and debugging. Even routine improvements can save researchers time and reduce computing demands, while performance-based refactoring and rewrites can deliver larger gains.

However, lower implementation costs also make it easier to produce many similar rewrites, fragmenting users and spreading the expert attention required to keep any one tool reliable. This makes long-term stewardship and attribution essential for mature scientific software, which carries undocumented conventions, compatibility requirements, and user trust that cannot be reproduced by translating source code alone.

The case studies illustrate several possible paths forward. Changes to MHCflurry and cyvcf2 were incorporated into their original upstream projects, while rustar-aligner moved under new community stewardship because the original project had been abandoned. Where coordination with existing maintainers is available, it should begin as early as possible. When a separate implementation is necessary, it needs a clear owner and a credible maintenance plan.

The field report highlights that coding agents like Codex can significantly lower the cost of maintenance, migration, optimization, and new implementations. Their long-term scientific value still depends on human decisions around what to build, how to verify it, and who will maintain it. The deeper change is not simply that researchers can produce more software but that they can focus more effort on defining, validating, and stewarding the tools.

These case studies show that agents can already accelerate iteration in scientific computing. As coding agents improve, researchers will be able to spend less time keeping analysis pipelines running and more time advancing their fields.

A financial services company has introduced an artificial intelligence-driven feature designed to help users prepare for regular expenses. Oportun’s Smart Bills tool automatically sets aside money for recurring payments, including rent, utilities, insurance, and subscriptions. This new addition builds on the company’s existing Set & Save platform, which uses AI analysis of income and spending patterns to allocate funds towards savings goals.

Since its launch in 2015, the Set & Save platform has helped users accumulate over $12.8 billion in savings, with an average annual saving of $1,800 per user. The new Smart Bills feature takes this a step further by calculating ‘safe-to-save’ amounts and reserving funds ahead of upcoming payment dates. It can also identify recurring expenses that users may want to track.

Early data from Oportun suggests that users are taking advantage of the tool’s capabilities, with those using Smart Bills saving 48% more each month compared to those relying on Set & Save alone. The feature has collectively saved its users over $14 million so far, with car payments, rent, and utilities being among the most commonly added expenses.

The launch reflects a broader trend in consumer finance, where FinTech companies are increasingly using artificial intelligence to provide personalized budgeting tools. Rather than simply tracking spending, newer AI-powered solutions aim to help consumers anticipate expenses and make proactive financial decisions.

According to Annie Ma, Oportun’s head of savings, ‘Set & Save makes it easier for people to build healthy financial habits in a way that works for their everyday lives.’ With Smart Bills, the company is helping users connect their savings to recurring priority expenses, allowing them to better budget and plan ahead.

Military personnel at Marine Forces Special Operations Command (MARSOC) are seeking to streamline their armory operations by leveraging cutting-edge technologies. The command has partnered with the SOFWERX innovation hub on an initiative aimed at automating its armories using computer vision, among other tools.

The goal is to develop a secure and efficient system that can track and manage thousands of serialized weapons, optics, lasers, night vision devices, communications equipment, and other controlled assets. Current processes rely heavily on manual inventories, handwritten forms, repetitive data entry, and multiple independent records, which are labor-intensive, prone to human error, and consume significant man-hours.

The MARSOC armory maintains a vast inventory of serialized items that transfer custody among operators, maintenance activities, and storage locations. To address this challenge, the command wants to collaborate with innovative companies to design an automated system that relies exclusively on optical recognition and computer vision.

A key aspect of this initiative is the development of a passive, image-based inventory system where weapons, gear, and serial numbers are automatically tracked, verified, and logged by taking digital photos or scanning optical images as items enter or exit the facility. This approach will eliminate slow paper-based workflows and human-error bottlenecks while ensuring absolute property accountability.

The use of RFID, active Bluetooth, or ‘radiating’ digital transmitters is prohibited to maintain operational security. The command wants a system that guarantees absolute property accountability without exposing its facilities to electronic detection.

MARSOC and SOFWERX are scheduled to host a collaboration event in Tampa, Florida, in September where solution providers will meet with military personnel to discuss their needs. Vendors who make it past a downselect later this year will be invited to an assessment event slated for December where their tech will be evaluated.

The evaluation process could potentially lead to other transaction agreements, FAR-based contracts or other deals with industry. The deadline for vendors to request to attend the collaboration event in September is August 16.

Singapore’s DBS Group has upgraded its artificial intelligence (AI) powered virtual assistants, enabling them to support everyday banking tasks across retail, wealth, and corporate banking. The enhancements apply to two separate assistants: DBS Joy, which serves corporate and small-to-medium enterprise customers, and DBS digibot, catering to retail banking clients.

The upgrades bring the total number of users supported by these AI assistants to over 10 million in Singapore, Hong Kong, and Taiwan. Each month, they are expected to handle more than one million customer chats. This marks a significant shift from AI being used primarily as an information assistant to becoming a digital banking assistant capable of carrying out customer-requested tasks through a single conversation.

DBS Group Chief Operating Officer Derrick Goh emphasized the true value of AI lies in delivering meaningful outcomes for customers at scale. He stated, ‘The next leap forward is moving from helping customers find information to helping them get things done.’ This vision has driven DBS’ more than decade-long commitment to AI innovation.

For corporate and SME clients in Singapore, DBS Joy has already been upgraded with agentic AI capabilities. Instead of directing users to navigate different sections of the bank’s digital banking platform, the assistant can now retrieve relevant information and complete selected banking requests within a single conversation.

DBS customers can ask natural language questions about recent payments, transaction history, or banking fees, with the assistant analyzing account and transaction data to provide requested information directly. The capabilities will soon be extended to corporate clients in Hong Kong from September before expanding to other DBS markets.

The enhancement addresses some of the most common requests received by the assistant, including payment tracking, transaction enquiries, and guidance on using the bank’s corporate banking platform. Additionally, DBS Joy is being expanded with knowledge of SME banking products and programmes, allowing businesses to seek information on banking solutions alongside servicing requests.

New live chat functionality will enable customers to connect with customer service officers when required, while transitions from chat to phone support are planned for the fourth quarter of 2026. This upgrade aims to streamline interactions between customers and bank representatives.

On the retail banking side, DBS has enhanced its digibot assistant to respond more naturally to customer enquiries relating to cards, remittances, refunds, and fee waivers. Beginning in August, it will also be integrated into DBS’ digiWealth platform to help customers navigate wealth-related enquiries and connect with their wealth planning managers when needed.

Later this year, DBS plans to further upgrade digibot with agentic AI and voice AI capabilities. These enhancements will enable customers to complete selected servicing requests without leaving the conversation. The rollout represents DBS positioning conversational AI as a means of simplifying customer interactions while reducing the effort required to complete routine banking tasks.

The AI assistants will execute only customer-initiated instructions, ensuring that customers requiring assistance with more complex issues can continue to access human customer service officers.

Packaging machinery manufacturers are redesigning equipment to help producers tackle labor shortages, rising sustainability requirements, and increasingly complex production demands. The industry is undergoing significant changes driven by four major forces: labor availability, smaller portion sizes, increasing product variety, and sustainability requirements. Manufacturers are being asked to produce more packs, more SKUs, often from the same factory footprint with the same or fewer skilled operators. Retailers, regulators, and consumers are placing greater emphasis on reducing waste and improving sustainability performance.

Simon Hill, group product innovation manager at TNA Solutions, highlights these pressures driving demand for packaging solutions that can deliver higher throughput, faster changeovers, lower waste, and greater flexibility without adding complexity to operations. He asserts that operational efficiency and agility are now critical differentiators for food manufacturers as they respond to changing consumer and market demands.

Sumeet Arora, senior VP Engineering Business Marketing at UFlex, argues that the packaging machinery industry has evolved significantly over the years. For many years, it was driven by three fundamental objectives: speed, flexibility, and accuracy. Manufacturers focused on increasing machine speeds, improving precision, and enabling quick product changeovers. However, customer expectations have changed, and manufacturers are now looking beyond machine performance.

They want higher productivity from the same manufacturing footprint, maximum equipment uptime, lower operating costs, and greater process predictability. Packaging machinery is increasingly expected to become an intelligent manufacturing asset rather than simply a production tool. Arora also notes that consumer preferences are changing rapidly, product portfolios are expanding, and packaging formats continue to evolve.

This has led to the growing need for versatility in machines. They must be capable of handling multiple pack formats and a wider range of packaging materials with minimal changeover time. Smart factory solutions are playing a crucial role in this transformation, as automation and digital technologies help manufacturers reduce manual intervention and improve accuracy.

Stephen Tagg, global application sales manager for Software at Markem-Imaje, explains that AI inspection is a valuable application in packaging. It can check if the right code, label, image, or product feature is present without relying on manual inspection or older rule-based systems. CoLOS 7 from Markem-Imaje supports fully inline verification through its Vision and Verification module, reducing manual coding errors by up to 20%.

Andrea Solfa, product manager for End of Lines and Tunnels at Sidel, adds that connected platforms enable manufacturers to monitor performance in real-time, gain greater visibility across production lines, and make faster decisions. This supports predictive and proactive maintenance strategies, minimizing downtime while maximizing equipment availability. Digital technologies are continuing to evolve, with data-driven insights playing a crucial role in enhancing efficiency and productivity.

Automation is becoming essential for overcoming labor constraints and improving overall equipment effectiveness. The next generation of packaging machinery is designed to simplify operation, reduce the risk of operator error, and support faster decision-making through greater visibility of machine and line performance. Markem-Imaje’s Tagg expects automation to grow due to labor pressures.

Many manufacturers are finding it harder to attract people into factory roles. Deloitte and the Manufacturing Institute estimate that US manufacturers will require 3.8 million new workers by 2033, with nearly half of these roles potentially remaining unfilled if current labor gaps persist. Similarly, Deloitte reports that 62% of European manufacturers face recruitment and retention challenges expected to continue through 2030.

Automation is not about replacing people but empowering them to focus on problem-solving, quality, and decision-making while machines handle repetitive tasks. Elias Hashemian Nik, product manager sterilization at Syntegon, highlights the increasing significance of packaging machinery automation amid labor shortages and productivity demands, alongside regulatory factors in pharmaceutical production.

Pharmaceutical manufacturers operate in a highly regulated environment where new requirements directly influence equipment design. The EU Good Manufacturing Practice (GMP) guidance continues to shape major shifts in particular. Manufacturers are facing growing pressure from product diversity and the need for faster changeovers. They are looking for equipment that ensures compliance, improves productivity, and operational flexibility.

Syntegon argues that automation is the solution to these challenges in liquid fill-finish operations. The fully automated SynTiso fill-finish line removes human interference with gloveless isolator technology and a fully robotized operation, including contactless container transport. It ensures benchmark-level regulatory compliance and up to 50% faster changeovers.

Digital technologies are becoming equally important as customers increasingly expect transparent production data for predictive maintenance, performance optimization, and data-driven decision-making. Solutions like Synexio provide visibility by connecting equipment, monitoring process performance, and helping manufacturers continuously improve production over the entire lifecycle of their equipment.

Tobias Richter, chief sales officer and group president at Multivac, highlights that increasing cost pressure is shaping the packaging machinery market. Manufacturers must reduce packaging material use while maintaining high levels of efficiency, output, and pack quality. Machinery must enable reliable processing of thinner films and alternative materials.

Multivac supports sustainability targets by helping customers reduce material consumption, energy use, and process-related waste in industrial production. The company develops equipment that can reliably process recyclable monomaterials and alternative packaging materials while maintaining efficiency, output, and pack quality. Its cloud-based Smart Services combine a Consumption Monitor with advanced analytics functions to enable detailed insights into energy and resource consumption.

Sidel’s Solfa expects continued growth in technologies combining sustainability, automation, and operational streamlining. Advanced bottle technologies like Sidel’s Returnable PET — Still Water bottle support circular packaging objectives by enabling bottle-to-bottle recyclability at the end of its service life. Integrated production solutions that streamline multiple processing stages into a single system will continue to gain traction as manufacturers seek greater productivity and flexibility.

TNA’s Hill adds that TNA supports consumers’ environmental sustainability goals by helping reduce product giveaway, minimizing film waste, and improving production efficiency. Equipment must provide the flexibility to accommodate new packaging materials and support future sustainability initiatives without compromising performance. Manufacturers are looking for sustainability improvements that also deliver measurable operational and commercial benefits.

Dairy producers are no strangers to embracing new technology, but the concept of artificial intelligence (AI) may still be unfamiliar to many. However, a growing number of companies in the industry are exploring ways to integrate AI into their operations, and with good reason – this emerging technology has the potential to revolutionize decision-making processes and improve overall efficiency.

A key player in this movement is Shari van de Pol, founder and CEO of CATTLEytics, an ag-tech company dedicated to using technology, particularly AI, to help farmers make better data-driven decisions. Van de Pol’s expertise lies at the intersection of computer engineering and large-animal veterinary medicine, making her uniquely qualified to discuss the potential applications of AI in dairy production.

One area where AI can have a significant impact is in data analytics. Producers often find themselves overwhelmed by vast amounts of information, but with AI, they can analyze and interpret this data to inform their decisions. This not only streamlines daily operations but also enables producers to make more informed choices about nutrition and feed rations, breeding schedules, disease diagnosis, and treatment.

AI’s benefits extend beyond cattle management, however. Van de Pol notes that it has shown promise in helping producers with employee management as well. By utilizing AI technology, dairy farms can develop onboarding processes, protocols, and trainings for employees, making the transition to new staff smoother and more efficient. Additionally, AI can assist in creating employee work schedules and even translate information into another language within the context of their operation.

One of the most significant advantages of AI is its ability to level the playing field between large-scale operations and smaller dairies. Van de Pol suggests that this technology can be a game-changer for producers who may not have access to the same resources or expertise as larger farms, enabling them to compete more effectively in the market.

When it comes to implementing AI on their farm, van de Pol recommends starting small – selecting one aspect of operations they’d like to work on and asking AI for ideas. She also encourages producers to explore companies that understand both the industry and tech, finding people who can make the technology work for them rather than against them.

Van de Pol emphasizes the importance of trust in experts and professionals within the agriculture industry when using AI tools. While this technology has limitations, it can help fill gaps and generate ideas, making it an invaluable resource for producers looking to improve their operations. However, she cautions that producers must still rely on human expertise – such as nutritionists or veterinarians – where necessary.

The learning process itself is not a difficult one, van de Pol assures us. She suggests starting by simply opening a browser and using platforms like ChatGPT, Copilot, or Perplexity to generate ideas and suggestions based on large sets of data. By practicing with these tools, producers can develop the skills needed to effectively utilize AI in their operations.

Another key component is learning how to ask the right questions – something van de Pol stresses as crucial for success when using AI tools. She advises producers not to be discouraged if they encounter difficulties or unclear results; instead, challenge the platform again and refine its answer within the context of dairy or with expert input.

The potential benefits of integrating AI into dairy operations are vast, from improved decision-making processes to enhanced efficiency and productivity. While it may take time for producers to become familiar with this new technology, van de Pol is confident that they hold the power to make it an effective tool for their business.

A major hurdle in manufacturing automation is the challenge of bin picking, a task that has remained largely manual despite advances in technology. For decades, automating bin picking has required custom projects, involving months of design and expensive vision specialists, system integrators, and costly iterations once it reaches the plant floor to achieve cycle times. This model is broken, hindering reshoring efforts, greenfield builds, and brownfield modernization across manufacturing in a constrained labor market.

The acceleration of reshoring, driven by factors such as labor shortages and supply chain disruptions, has highlighted the need for more efficient automation solutions. However, traditional approaches to bin picking have failed to scale fast enough to meet manufacturing’s demand for automation. This is where Apera AI comes in with its innovative approach to autonomous bin picking.

Apera VuePod is a turnkey bin-picking cell that ships fully proven on customers’ parts before it arrives at their facility, eliminating the integration project, vision and system design expertise tax, and validation risk. This shift from ‘projects to products’ changes the game for manufacturing companies looking to automate bin picking tasks.

The physical AI difference lies in autonomous vision and real-time decision-making that eliminates months of iteration and on-floor tuning. Apera’s solution is not just a custom build but a proven product that can be deployed across multiple stations and plants, making it an attractive option for manufacturers seeking rapid ROI and scalability.

Jamie Westell, Director of Engineering at Apera AI, brings extensive experience in software development and engineering management to the company. With over 15 years of expertise in imaging devices and computer vision, Jamie has played a key role in driving technological advancements and overseeing large-scale development teams. His notable leadership experience includes managing a team of over 60 engineers at Motorola Solutions.

Sina Afrooze, CEO of Apera AI, is a pioneer in the field of automation, with over 20 years of experience in artificial intelligence, machine learning, and systems engineering. Sina’s track record of bringing transformative vision technology to scale includes his tenure as a founding engineer at Avigilon, where he contributed to technology that became the surveillance industry standard before Motorola’s acquisition for $1.2 billion.

At Apera AI, Sina is focused on solving automation problems that have resisted conventional solutions: bin picking of reflective and transparent parts, handling high part variability in unstructured environments, and reducing expertise and on-site engineering required to deploy vision-guided systems in plants. His conviction that ‘most manufacturing tasks can be automated if robots can see like humans’ drives Apera’s suite of 4D Vision solutions.

The same proven solution brings autonomous bin-picking and rapid ROI across the entire spectrum of manufacturing industries, from automotive, pharma, aerospace to automation OEMs and system integrators. This shift in buying model means that manufacturers no longer need to invest months of time and resources into custom projects but can instead deploy a proven product that meets their specific needs.

The webinar will cover why now is the right time for this shift, highlighting the acceleration of reshoring efforts and labor shortages as key drivers. It will also delve into the physical AI difference that makes Apera’s solution so effective in eliminating months of iteration and on-floor tuning. Additionally, it will explore how this new buying model changes bin-picking automation ROI, deployment speed, and scalability across multiple stations and plants.

The webinar is a must-attend for manufacturing companies looking to automate their bin picking tasks efficiently and effectively. By attending the webinar, manufacturers can learn about Apera’s innovative solution and how it can help them overcome the challenges of traditional approaches to bin picking.

A law firm’s collective knowledge is its most valuable asset, but tapping into it can be tricky. DeepJudge, developed by former Google researchers, uses advanced search technology to catalog data from internal sources like SharePoint and email systems.

The heart of the platform lies in making a firm’s institutional knowledge searchable. By indexing vast amounts of information, DeepJudge equips law firms with tools for refining their research processes and informing better decisions. For example, it allows cross-matter analysis that can be a real game-changer for lawyers dealing with complex cases.

Security is critical when working with sensitive legal data, so DeepJudge’s design prioritizes security from the start. It integrates seamlessly with existing systems, making it an attractive option for law firms looking to upgrade their knowledge management capabilities. Companies like Freshfields and Homburger have already adopted DeepJudge, reporting high adoption rates and substantial time savings.

While primarily designed for mid-to-large law firms, the platform’s benefits can extend beyond its intended scope. By using LLM capabilities to analyze data, law firms can uncover new insights that improve their overall performance. A firm’s quality and volume of internal data will significantly impact DeepJudge’s effectiveness – in other words, it won’t work well if they don’t have a lot of useful information to draw from.

The results speak for themselves: this AI platform stands out as a highly practical solution for the legal industry’s knowledge management challenges. By solving the problem of fragmented institutional knowledge, DeepJudge empowers lawyers to work more efficiently, make better decisions, and deliver higher-quality client work.

Home improvement retailer BAUHAUS has partnered with robotics company XYZ Robotics to automate its entire goods receiving process at the central warehouse in Krefeld, Germany. The system uses mobile manipulation robots called RockyOne and RockyOne SE to handle container unloading, sorting, and palletizing in a single continuous workflow.

The automation solution is designed to improve operational consistency and free up staff from manual labor. BAUHAUS handles thousands of inbound shipments daily across over 10,000 stock-keeping units (SKUs), with cartons varying widely in shape, weight, and stacking. Conventional automation systems struggle to cope with this complexity.

XYZ Robotics’ end-to-end system addresses these challenges by using RockyOne robots to autonomously enter shipping containers and unload loose cartons of varied sizes and orientations. A barcode scanning system identifies each carton for real-time sorting, while the RockyOne SE robot builds pallets according to customer-defined patterns.

The automation process has significantly reduced manual labor involved in container unloading, which was previously a repetitive task with high turnover and injury risk. Operators now supervise the system through a tablet interface, requiring minimal training. The system runs reliably around the clock without manual intervention, allowing staff to focus on higher-value work.

BAUHAUS Vice President of Operations (Warehousing & Transport) Torsten Winter praised the partnership, saying ‘Our team can now focus on more fulfilling tasks.’ He also mentioned that BAUHAUS is exploring further applications across its network. The company has found a reliable and forward-thinking partner in XYZ Robotics, which provides full lifecycle support.

Following the successful deployment at Krefeld, XYZ Robotics is expanding its European footprint to serve more warehouse and distribution operators. The company develops mobile manipulation robots (MMR) that pioneer automation in truck loading, unloading, and case picking. Founded in 2018, XYZ Robotics serves industry-leading customers worldwide with branches in Germany, the United States, China, Japan, and South Korea.

BAUHAUS has a long history of innovation, introducing self-service brand-name products from various specialty categories under one roof as early as 1960. The company has established over 160 specialty centers in Germany and more than 290 locations across Europe in 19 countries. Each location is divided into 15 departments, offering high-quality products at competitive prices.

The partnership between BAUHAUS and XYZ Robotics marks a significant step towards automating warehouse operations and improving efficiency. As the demand for automation solutions continues to grow, companies like XYZ Robotics are poised to play a crucial role in streamlining logistics processes.

Dutch-based Deliverect, a global restaurant technology platform, has formed a partnership with US-based SoundHound AI. The collaboration aims to bring fully automated voice ordering capabilities to restaurants worldwide.

The integration links SoundHound’s Smart Ordering system directly to Deliverect’s order and menu management platform. This allows SoundHound’s voice AI agents access real-time menu information from Deliverect, enabling them to take orders through various channels, including phone calls, drive-throughs, and kiosks.

Once an order is placed, it will be sent straight to the kitchen via Deliverect’s certified POS connections. This streamlined process eliminates the need for manual updates across different ordering channels, reducing errors and increasing efficiency.

The partnership also enables menu information synchronization in real-time from Deliverect, ensuring that all voice AI agents have access to up-to-date menus. This removes the need for manual updates, further streamlining operations.

According to Don MacMillan, head of US and Canada partnerships at Deliverect, ‘voice is the oldest ordering channel in the restaurant business.’ He notes that until now, it has been the least automated, but this partnership changes that.

The integration allows every spoken order to flow directly into the kitchen with the same accuracy and speed as a digital order. This means operators can focus on hospitality rather than handling orders manually.

Deliverect’s platform supports more than 100 languages, making the partnership particularly beneficial for international restaurant chains looking to extend their ordering infrastructure across channels and markets.

In an interview, Mike Lauricella, vice-president of channel partnerships at SoundHound AI, emphasized that ‘the future of restaurant hospitality relies on automation and convenience.’ He believes this partnership will help deliver a seamless customer experience by reducing front-of-house strain through automated order placement.

A new study published in the Journal of the American College of Radiology suggests that large language models (LLMs) can automate quality control processes in radiology departments. The research found that LLM-based systems can streamline the processing of radiology report checks, with potential applications beyond breast imaging alone.

The study focused on identifying variability in breast ultrasound reports and used a dataset of 735 patients from 60 hospitals in China. Researchers compared the performance of an LLM to manual reviewers who converted free text reports into standardized BI-RADS-based structured reports. The results showed that the LLM outperformed human reviewers in evaluating several key characteristics, including margins and echo patterns.

The model’s accuracy persisted even when dealing with complex reports involving multiple lesions, and it improved as the level of suspicion increased. In fact, the LLM demonstrated particularly strong performance in reports containing more suspicious findings. This suggests that LLMs could be a valuable tool for automating quality control processes in radiology departments.

One key advantage of using an LLM is its ability to process complex clinical narratives with accuracy comparable to manual reviewers. The model’s extensive pretraining on medical texts and strong contextual understanding enable it to identify errors, inconsistencies, and omissions in reports. Additionally, the LLM does not suffer from fatigue like human reviewers do, ensuring stable and consistent performance.

The study also highlighted a significant workflow benefit of using an LLM. In comparison to manual reviewers who took an average of 213 minutes to complete quality control for 50 reports, the model completed the same task in just 13 minutes. This substantial time savings could be particularly valuable for radiology departments looking to streamline their workflows.

While the study’s findings are promising, researchers acknowledge that there are concerns related to data privacy and the need for frequent updates with new information. They suggest that future studies should evaluate user acceptance, infrastructure requirements, and long-term model stability in real-world settings.

OpenAI has launched its new ‘Health’ feature for users of its popular chatbot, ChatGPT. The feature allows users to securely connect their health information and use it to inform conversations with the AI assistant. This marks a significant development in the integration of healthcare data into conversational AI systems.

The Health feature is currently available to logged-in US-based users aged 18 and older on web and iOS platforms, across all plans including Free, Go, Plus, and Pro. Users can access the feature by opening ChatGPT’s sidebar or More menu and selecting ‘Get started’.

According to OpenAI, more than 300 million people use ChatGPT each week with health-related questions, often scattered across patient portals, medical records, apps, and wearables. The new Health feature aims to provide a comprehensive view of users’ health information by allowing them to connect relevant data from various sources.

With the user’s permission, ChatGPT can draw on connected health information to help compare lab results with prior tests, summarize changes since their last appointment, or explore how sleep and activity relate to their routine. This can reduce the need for repeated gathering of details, making users feel more informed and empowered in managing their healthcare.

OpenAI has implemented layered privacy and security safeguards to protect connected health information. Conversations that use this data are not used to train its foundation models or target ads, regardless of the model-training setting chosen by the user. Connected medical records and Apple Health information receive additional encryption protections.

The company continues to improve its models behind health conversations in ChatGPT through dedicated training. GPT-5.6 Sol is OpenAI’s strongest model yet for handling complex health-related tasks, building on progress made with stronger performance on reasoning across details, clear communication, and careful judgment.

Physicians extensively tested the Health feature before release to measure real-world model performance and safety with connected health data. While these results represent meaningful progress, ChatGPT can still make mistakes and does not replace qualified medical professionals’ care and judgment. Users are encouraged to verify important information and discuss decisions with their healthcare provider.

The user has control over when ChatGPT uses the connected health information. By default, the AI assistant asks for permission before using this data to personalize a response. Users can approve each request or choose to always allow access, which turns off these permission prompts.

Health conversations are stored in the left sidebar of ChatGPT, where users can manage their accounts and review recent data and trends. Connected information may not be complete or current, so it’s essential for users to check important details against the original source.

When connecting health information, users should ensure that sharing is turned on with any wearable, fitness, or nutrition app linked to Apple Health. ChatGPT can use available information from these apps when allowed by the user.

Once connected, users can review conditions and medications, remove irrelevant data, and add details such as family health history. It’s crucial for users to tell ChatGPT about changes in their health status and verify important details against original sources.

The Health feature allows users to ask questions anywhere within ChatGPT, with the AI assistant using connected health information based on user permissions. Users can change these settings at any time in Settings > Plugins > Health or add @Health to a message to explicitly request health context.

OpenAI emphasizes that its primary goal is not to replace medical professionals but rather support users’ healthcare decisions by providing accurate and relevant information. The company continues to test safeguards through dedicated red-team exercises focused on connected information in the Health feature, using findings to strengthen protections.

The challenges facing modern researchers are unprecedented. From simulating the intricate dynamics of fusion plasma to exploring vast search spaces for new materials and making sense of exabytes of data from experimental facilities, scientists today face extraordinary scale and complexity. Frontier AI can help address these challenges while accelerating groundbreaking scientific discoveries.

In December, Google DeepMind announced its commitment to the White House’s Genesis Mission – a national effort to harness AI and double the pace of American scientific discovery within a decade. The company has since provided early access to AI for science tools at 17 Department of Energy National Laboratories through an in-kind program. Today, at the DOE Genesis Mission Summit 2026, Google DeepMind is expanding this commitment by pledging $40 million in AI tokens and cloud credits for researchers supporting the Genesis Mission.

The expanded commitment provides two key benefits: first, it gives DOE’s Genesis Mission awardees access to Google DeepMind’s frontier AI for science portfolio. This includes AlphaEvolve – a Gemini-powered coding and discovery agent designed for advanced algorithms; AlphaFold 3 – a model predicting protein structure and interactions; AlphaGenome – a tool understanding DNA variation and its impact on biology and disease; WeatherNext – state-of-the-art weather forecasting models; and AlphaEarth Foundations, an AI model mapping our planet in unprecedented detail.

Secondly, Google DeepMind will provide Gemini for Government seats and tokens to tens of thousands of users across the DOE National Laboratories’ operations, research, and management teams. This secure platform supports work from the research bench to specialized user facilities serving the entire scientific community, providing a single foundation that the DOE mission can depend on.

The practical impact of the Genesis Mission is already coming to life across laboratory ecosystems. At Pacific Northwest National Laboratory (PNNL), senior scientist Dr. Henry Kvinge uses AlphaEvolve to map out massive mathematical systems too complex for humans to explore by hand. The AI uncovers hidden connections automatically, fast-tracking discoveries that would normally take years.

Dr. Kvinge notes the importance of abstraction in modern math and how combinatorics offers concrete models making geometry and algebra easier to grasp. He believes AlphaEvolve is perfect for this search due to its ability to automate exploration using LLMs’ broad mathematical knowledge, ‘automate the exploration of countless angles.’ The discoveries are already shaping future research.

At the National Laboratory of the Rockies (NLR), researchers utilize Gemini to fundamentally change how they interact with physical laboratory hardware. Dr. Steven R. Spurgeon leads a pioneering program in autonomous materials discovery using Gemini instruments. He explains that deploying Gemini has cut microscope calibration time from over 90 minutes to about 13 minutes, reducing manual steps needed to focus an image from as many as 50 down to two.

This has given back time and attention to the science itself, enabling genuinely autonomous workflows observing, reasoning, and deciding in real-time. This capability helps explore parts of the material design space that could not be reached through manual operation alone. The Genesis Mission represents a chance to transform research and science across America by providing access to advanced AI tools.

By accelerating breakthroughs across critical energy, security, and scientific challenges, Google DeepMind aims to help scientists overcome these hurdles. To learn more about how these AI capabilities can support your research initiatives, join the upcoming Google Public Sector Summit in October.