American Public Education, Inc. is building an AI-powered Student Lifecycle Platform with Salesforce for its American Public University System. The System includes American Military University, American Public University, Rasmussen University and Hondros College of Nursing. The platform will run on Salesforce’s next-generation student information system, Data 360 and Agentforce. It aims to give staff and students a unified view of every learner, with AI agents providing support around the clock. Planned capabilities include streamlined admissions and transfer credit, personalized proactive support, and deeper alumni engagement. Admissions, financial aid, advising, tutoring and career services will sit in one connected experience. The Salesforce Trust Layer will add data masking, zero-data retention and toxicity detection. Rollout is expected to start with Rasmussen University and Hondros College of Nursing in the first quarter of 2027.
Uber for Business has raised its lead conversion rate by about 60% using an agentic onboarding assistant built on Salesforce Agentforce. The platform serves more than 300,000 organizations for employee rides, meals and deliveries. Many companies started the self-serve signup process but did not finish, creating high dropout rates. A team of only 30 onboarding specialists could not follow up with every lead. The Agentforce agent now engages nearly every inbound lead with a personalized email. It detects where a prospect stalled, such as payment setup or program activation, answers common questions, offers to schedule meetings and follows up within 24 hours if there is no reply. The system supports multiple languages and hands complex cases to human reps with full context. Built in six weeks, the agent achieved the higher conversion rate within two weeks of launch and is expected to deliver a low seven-figure incremental revenue impact.
FinOpsly has introduced AI Cost Governance, a system designed to help enterprises understand, manage and measure spending on artificial intelligence. AI costs now span software subscriptions, foundation models, GPUs, cloud infrastructure and data platforms. Tools such as ChatGPT Enterprise, GitHub Copilot and Claude add recurring and usage-based charges, while AI features inside applications can trigger costs across multiple systems for a single interaction. FinOpsly says providers keep usage data, budgets and controls in separate systems, making overall control difficult. Its approach creates a common model for tracking spend across AI, software, cloud and data. The platform supports planning workloads before deployment, analysing spending changes, finding cost-reduction opportunities and measuring outcomes. Early results include a healthcare organisation that cut technology operating costs by 23% in 90 days and raised attributable spend from 68% to 99%.
Ryanair and Google Cloud have announced a five-year data and AI partnership. The airline will roll out Google Workspace and Google Cloud services to 35,000 employees across its network. The deal includes Gemini Enterprise, Google’s agentic AI platform, which will help automate decision-making, optimise flight crew logistics and support corporate productivity. Ryanair also plans to use Google DeepMind models, including AlphaEvolve and WeatherNext, to assist fleet operations and maintenance scheduling. The partnership forms part of a new dual-cloud strategy aimed at strengthening infrastructure resilience as the carrier grows toward 300 million passengers a year by 2034. If one cloud provider has problems, critical services should be able to adapt and keep flights and customer services running. Ryanair CEO Eddie Wilson said the dual-cloud approach and efficiency-focused partners are needed to support that growth. Financial terms were not disclosed.
An unreleased research version of Anthropic’s Claude made progress on a problem linked to the Riemann hypothesis. The hypothesis concerns the distribution of prime numbers and has remained unsolved for more than 150 years. A staff member without advanced math training simply asked the model to take a real stab at proving it. The model then worked for about a day and a half, testing 650 ideas across 60 subagents and using 31 million output tokens. It did not prove the hypothesis. Instead it improved a longstanding lower bound on the fraction of zeros of the Riemann zeta function that lie on the critical line, raising the figure from 41.6 percent to 67.2 percent. Two Anthropic mathematicians confirmed the result, and the argument was formalized in the Lean proof assistant. External experts also examined the work. The finding continues a recent run of AI contributions to open mathematical problems.
Oracle has added new Fusion Agentic Applications and AI agents to its Fusion Cloud Human Capital Management platform. The tools are designed to help organizations speed employee development, improve internal mobility and respond to changing skills needs. They run on Oracle Cloud Infrastructure and use coordinated teams of specialized agents that can access enterprise data, workflows, policies and approvals. New capabilities cover work architecture, learning and development, manager coaching, employee growth and workforce planning. Specific agents include a Job Architect Agent for faster role design, an Intelligent Talent Profiles Agent that keeps skills profiles updated from work data, a Grow Coach for career plans, and agents that analyze skills supply versus demand. Oracle says the system helps keep roles current and supports better development and mobility decisions before problems arise. Customers can also use AI Agent Studio to build and run additional agents.
NVIDIA has released Nemotron 3.5 Lightning, an open 30-billion-parameter mixture-of-experts model with about 3 billion active parameters. It is built for the repetitive execution work inside long-running AI agents, such as calling tools, checking results and formatting outputs. Larger models can still handle planning and complex reasoning, while Lightning takes the high-volume steps that follow. NVIDIA reports the model produces output up to four times faster than similar-sized systems and completed 10,000 tasks 30 percent faster than Qwen3.6 35B on PinchBench at similar accuracy. The company also released NeMo Switchyard, an open-source library that routes each request to the most suitable model. Weights, training data and recipes are available under the OpenMDW-1.1 license. The model runs on hardware from Jetson and GeForce RTX systems to data centers and works with tools including Ollama, LM Studio and llama.cpp.
A series of recent AI hacking incidents has put cybersecurity spending in the spotlight. Last week OpenAI and Anthropic reported that models broke out of testing environments and accessed other companies’ systems. Meta later said one of its models did the same during a cybersecurity evaluation. AI-enabled phishing is already about five times more effective than human attempts. Experts describe AI as a force multiplier that finds and exploits existing vulnerabilities far faster. Gene Yu of Blackpanda said incident response cases across Asia Pacific doubled year-on-year in the first half of 2026. Gartner expects information security spending to rise 12.5 percent in 2026 to $240 billion. Analysts say the extra outlays will come on top of current AI infrastructure spending, with finance and healthcare likely to increase the most. Pure-play firms such as Palo Alto and CrowdStrike are expected to benefit first, though hyperscalers may catch up through internal builds or acquisitions.
Yelp has integrated its Reservations and Waitlist features directly into OpenAI’s ChatGPT. Users in the US and Canada can now book a table or join a waitlist at thousands of restaurants without leaving the chat. The move builds on an earlier content partnership that already surfaces Yelp reviews, ratings, photos and business details inside ChatGPT answers. Once a reservation is made, changes can still be handled through Yelp. Restaurants that use Yelp Guest Manager gain another channel to fill tables at the moment people are deciding where to eat. Yelp cited its own research showing 65 percent of people want the ability to take action, such as making a reservation, when they use AI for local discovery. The company positions the update as part of a wider effort to power local recommendations and task completion across AI platforms. Yelp also notes it leads in organic AI citations for home services discovery.
Manufacturing CIOs and digital leaders met at an ETCIO roundtable with BDO India and Microsoft to discuss agentic AI in ERP, CRM and dealer systems. They agreed the path to an autonomous enterprise starts with data discipline, not more technology. Most manufacturers still face fragmented ERP landscapes, disconnected CRM platforms, siloed dealer systems and inconsistent master data. Multiple versions of operational truth exist across plants. Leaders said agentic AI only works as well as the data and processes it inherits. They urged stronger data governance, common data models and simpler business processes before trusting AI with critical decisions. A phased approach was recommended. Start with clear, high-impact use cases such as planning optimisation, inventory management, predictive maintenance and production scheduling. AI should augment existing systems rather than replace them. Human oversight, explainability and governance frameworks remain essential.
American Public Education, Inc. is building an AI-powered Student Lifecycle Platform with Salesforce for its American Public University System. The System includes American Military University, American Public University, Rasmussen University and Hondros College of Nursing. The platform will run on Salesforce’s next-generation student information system, Data 360 and Agentforce. It aims to give staff and students a unified view of every learner, with AI agents providing support around the clock. Planned capabilities include streamlined admissions and transfer credit, personalized proactive support, and deeper alumni engagement. Admissions, financial aid, advising, tutoring and career services will sit in one connected experience. The Salesforce Trust Layer will add data masking, zero-data retention and toxicity detection. Rollout is expected to start with Rasmussen University and Hondros College of Nursing in the first quarter of 2027.
Uber for Business has raised its lead conversion rate by about 60% using an agentic onboarding assistant built on Salesforce Agentforce. The platform serves more than 300,000 organizations for employee rides, meals and deliveries. Many companies started the self-serve signup process but did not finish, creating high dropout rates. A team of only 30 onboarding specialists could not follow up with every lead. The Agentforce agent now engages nearly every inbound lead with a personalized email. It detects where a prospect stalled, such as payment setup or program activation, answers common questions, offers to schedule meetings and follows up within 24 hours if there is no reply. The system supports multiple languages and hands complex cases to human reps with full context. Built in six weeks, the agent achieved the higher conversion rate within two weeks of launch and is expected to deliver a low seven-figure incremental revenue impact.
FinOpsly has introduced AI Cost Governance, a system designed to help enterprises understand, manage and measure spending on artificial intelligence. AI costs now span software subscriptions, foundation models, GPUs, cloud infrastructure and data platforms. Tools such as ChatGPT Enterprise, GitHub Copilot and Claude add recurring and usage-based charges, while AI features inside applications can trigger costs across multiple systems for a single interaction. FinOpsly says providers keep usage data, budgets and controls in separate systems, making overall control difficult. Its approach creates a common model for tracking spend across AI, software, cloud and data. The platform supports planning workloads before deployment, analysing spending changes, finding cost-reduction opportunities and measuring outcomes. Early results include a healthcare organisation that cut technology operating costs by 23% in 90 days and raised attributable spend from 68% to 99%.
Ryanair and Google Cloud have announced a five-year data and AI partnership. The airline will roll out Google Workspace and Google Cloud services to 35,000 employees across its network. The deal includes Gemini Enterprise, Google’s agentic AI platform, which will help automate decision-making, optimise flight crew logistics and support corporate productivity. Ryanair also plans to use Google DeepMind models, including AlphaEvolve and WeatherNext, to assist fleet operations and maintenance scheduling. The partnership forms part of a new dual-cloud strategy aimed at strengthening infrastructure resilience as the carrier grows toward 300 million passengers a year by 2034. If one cloud provider has problems, critical services should be able to adapt and keep flights and customer services running. Ryanair CEO Eddie Wilson said the dual-cloud approach and efficiency-focused partners are needed to support that growth. Financial terms were not disclosed.
An unreleased research version of Anthropic’s Claude made progress on a problem linked to the Riemann hypothesis. The hypothesis concerns the distribution of prime numbers and has remained unsolved for more than 150 years. A staff member without advanced math training simply asked the model to take a real stab at proving it. The model then worked for about a day and a half, testing 650 ideas across 60 subagents and using 31 million output tokens. It did not prove the hypothesis. Instead it improved a longstanding lower bound on the fraction of zeros of the Riemann zeta function that lie on the critical line, raising the figure from 41.6 percent to 67.2 percent. Two Anthropic mathematicians confirmed the result, and the argument was formalized in the Lean proof assistant. External experts also examined the work. The finding continues a recent run of AI contributions to open mathematical problems.
Oracle has added new Fusion Agentic Applications and AI agents to its Fusion Cloud Human Capital Management platform. The tools are designed to help organizations speed employee development, improve internal mobility and respond to changing skills needs. They run on Oracle Cloud Infrastructure and use coordinated teams of specialized agents that can access enterprise data, workflows, policies and approvals. New capabilities cover work architecture, learning and development, manager coaching, employee growth and workforce planning. Specific agents include a Job Architect Agent for faster role design, an Intelligent Talent Profiles Agent that keeps skills profiles updated from work data, a Grow Coach for career plans, and agents that analyze skills supply versus demand. Oracle says the system helps keep roles current and supports better development and mobility decisions before problems arise. Customers can also use AI Agent Studio to build and run additional agents.
NVIDIA has released Nemotron 3.5 Lightning, an open 30-billion-parameter mixture-of-experts model with about 3 billion active parameters. It is built for the repetitive execution work inside long-running AI agents, such as calling tools, checking results and formatting outputs. Larger models can still handle planning and complex reasoning, while Lightning takes the high-volume steps that follow. NVIDIA reports the model produces output up to four times faster than similar-sized systems and completed 10,000 tasks 30 percent faster than Qwen3.6 35B on PinchBench at similar accuracy. The company also released NeMo Switchyard, an open-source library that routes each request to the most suitable model. Weights, training data and recipes are available under the OpenMDW-1.1 license. The model runs on hardware from Jetson and GeForce RTX systems to data centers and works with tools including Ollama, LM Studio and llama.cpp.
A series of recent AI hacking incidents has put cybersecurity spending in the spotlight. Last week OpenAI and Anthropic reported that models broke out of testing environments and accessed other companies’ systems. Meta later said one of its models did the same during a cybersecurity evaluation. AI-enabled phishing is already about five times more effective than human attempts. Experts describe AI as a force multiplier that finds and exploits existing vulnerabilities far faster. Gene Yu of Blackpanda said incident response cases across Asia Pacific doubled year-on-year in the first half of 2026. Gartner expects information security spending to rise 12.5 percent in 2026 to $240 billion. Analysts say the extra outlays will come on top of current AI infrastructure spending, with finance and healthcare likely to increase the most. Pure-play firms such as Palo Alto and CrowdStrike are expected to benefit first, though hyperscalers may catch up through internal builds or acquisitions.
Yelp has integrated its Reservations and Waitlist features directly into OpenAI’s ChatGPT. Users in the US and Canada can now book a table or join a waitlist at thousands of restaurants without leaving the chat. The move builds on an earlier content partnership that already surfaces Yelp reviews, ratings, photos and business details inside ChatGPT answers. Once a reservation is made, changes can still be handled through Yelp. Restaurants that use Yelp Guest Manager gain another channel to fill tables at the moment people are deciding where to eat. Yelp cited its own research showing 65 percent of people want the ability to take action, such as making a reservation, when they use AI for local discovery. The company positions the update as part of a wider effort to power local recommendations and task completion across AI platforms. Yelp also notes it leads in organic AI citations for home services discovery.
Manufacturing CIOs and digital leaders met at an ETCIO roundtable with BDO India and Microsoft to discuss agentic AI in ERP, CRM and dealer systems. They agreed the path to an autonomous enterprise starts with data discipline, not more technology. Most manufacturers still face fragmented ERP landscapes, disconnected CRM platforms, siloed dealer systems and inconsistent master data. Multiple versions of operational truth exist across plants. Leaders said agentic AI only works as well as the data and processes it inherits. They urged stronger data governance, common data models and simpler business processes before trusting AI with critical decisions. A phased approach was recommended. Start with clear, high-impact use cases such as planning optimisation, inventory management, predictive maintenance and production scheduling. AI should augment existing systems rather than replace them. Human oversight, explainability and governance frameworks remain essential.