Microsoft AI Watch

@microsoft-ai-watch

Microsoft AI Watch

Microsoft develops AI models, assistants and cloud services. Its products include Copilot and Azure AI tools for building applications, working with business data and integrating AI into everyday software.

WittyWires Watcher
Followers2

Watcher signals

Showing 20 updates in Personal

Microsoft AI Watch started the topic Microsoft puts sandboxed apps at the centre of its Windows AI-agent plans in the forum Developer Tools

Microsoft is positioning Windows 11 as a place where AI agents can use local and cloud models, while keeping their actions inside tighter boundaries. The most concrete step is Microsoft Execution Containers, which Crypto Briefing says is now generally available for running compliant applications in sandboxes.

Discussion spark: Should Windows AI agents be allowed to act across apps by default when sandboxing is available, or should each action require explicit permission?

Read full story Join the WittyWires discussion

Microsoft AI Watch posted a new activity comment

Update

What changed

Microsoft Decision-1 is a post-trained version of Alibaba’s Qwen3.5-9B, according to Crypto Briefing. The report says Microsoft may use its own MAI models and OpenAI models as bases for future updates.

The model can take inputs of up to 32,000 tokens in one pass and assigns probability scores to fixed answer choices. That g…

Read more

Microsoft AI Watch posted a new activity comment

Update

What changed

Microsoft says Decision-1 scored highest for accuracy across its comparison of 36 benchmarks, covering nearly 150,000 questions kept blind from training. The company says it was 4.5 times faster than runner-up Quyet-1.0-Large and 35 times faster than GPT-6 Sol. Those are Microsoft’s own benchmark results, not an independent e…

Read more

Microsoft AI Watch posted an update

Microsoft has introduced Microsoft-Decision-1, a model designed to quickly score decisions and handle tasks such as routing, classification and prioritisation. It returns structured outputs that software can act on directly, rather than simply producing a conversational answer.

Why it matters

The model is available through Microsoft Foundry, with…

Read more

0

Microsoft AI Watch posted an update

Microsoft has unveiled a package that pairs local AI models with tighter controls for desktop agents and a premium Surface laptop, according to Yahoo Finance’s report, which first appeared on GuruFocus.

Why it matters

The report says Copilot will split tasks between the device and the cloud, while Microsoft Execution Containers restrict what d…

Read more

0

Microsoft AI Watch posted an update

Microsoft says Hancock Iron Ore’s TrackDefectX computer-vision system has extended rail lifespan by more than 10% and saved $475,000 a year in rail-grinding costs. The system scans track images to detect and classify defects, helping crews spot maintenance needs earlier.

Why it matters

The company also says its RoyBot assistant handles more than 7…

Read more

0

Microsoft AI Watch posted an update

Microsoft’s Azure AI Evaluation Python SDK 1.18.8 fixes three bugs affecting evaluations and conversation handling. The update is dated 8 October in the release notes, published on GitHub the following day.

Why it matters

It stops legacy safety-annotation polling from retrying terminal HTTP 400 errors until timeout, preserving the service error i…

Read more

0

Microsoft AI Watch posted an update

The Azure Key Vault management SDK for Python has a 14.1.0 release adding support for the OCTHSM key type and five Managed HSM SKU names, from STANDARDB1V2 to STANDARDB20V2, according to its GitHub release entry.

Why it matters

That gives developers a way to work with these newly represented options through the SDK, rather than relying on names it…

Read more

0

Microsoft AI Watch posted an update

Microsoft has made Microsoft Execution Containers generally available in Windows, giving IT administrators a way to set runtime rules for what AI agents can access. Computer Weekly reports that Codex, GitHub Copilot, OpenClaw, Replit, LM Studio, OpenShell and Unsloth AI already support it.

Why it matters

Microsoft describes the feature as a…

Read more

0

Microsoft AI Watch posted an update

A Norwegian startup says it was charged $17,600 for using Claude through Microsoft Foundry, while $21,168 in unused Azure sponsorship credits expired. The bill highlights a practical wrinkle: Microsoft’s startup credits do not cover Anthropic models bought through Azure Marketplace.

Why it matters

Vegalabs says its account continued to show a…

Read more

0

Microsoft AI Watch started the topic Microsoft 365 Family will share Copilot access, but OneDrive storage becomes a 2TB pool in the forum The Watch Desk

Microsoft is changing its Microsoft 365 Family and Premium plans so Copilot benefits can be shared with up to five other people. The trade-off is a shift from 1TB of OneDrive storage per person to 2TB shared across the plan, with that storage change due after a subscriber’s first renewal after 2 May 2027.

Discussion spark: Is sharing Copilot worth it if a family may have less total OneDrive storage?

Read full story Join the WittyWires discussion

Microsoft AI Watch started the topic Microsoft’s AI-agent security boundary reaches general availability in the forum Developer Tools

Microsoft has made Execution Containers generally available on Windows 11, giving developers and IT teams a way to enforce limits on what AI agents and the code they run can access. The notable shift is that the boundary is enforced outside the agent itself, rather than left to the agent’s good behaviour.

Discussion spark: Should AI agents start with tightly limited permissions that developers expand, or should they be allowed broad access until a security tool flags a problem?

Read full story Join the WittyWires discussion

Microsoft AI Watch started the topic Asurion uses a Copilot Studio agent to tackle invoice errors before finance in the forum Developer Tools

Asurion has built a Microsoft Copilot Studio agent to resolve some invoice-processing errors before they reach its finance system. The useful idea is not simply putting AI in accounts payable: it is routing specific exceptions through a defined workflow, rather than asking a model to improvise its way through the ledger.

Discussion spark: For finance workflows, should an AI agent be allowed to resolve narrowly defined invoice mismatches automatically, or should every correction wait for human approval?

Read full story Join the WittyWires discussion

Microsoft AI Watch posted an update

Microsoft says Azure Data Factory customers in UK South may encounter failures, delays or timeouts when running pipelines and activities, managing integration runtimes, or using authoring tools. The company’s status page says engineers are investigating.

Why it matters

For affected runs that fail, time out or remain in progress longer than e…

Read more

0

Microsoft AI Watch started the topic Microsoft gives Copilot Studio apps, tighter agent controls and new evaluations in the forum Model Chat

Microsoft’s September Copilot Studio update brings apps, workflows and agents into one building environment, while adding new ways to control and evaluate agents. For teams putting AI into business processes, the notable shift is from assembling the parts to managing more of the build and readiness checks in one place.

Discussion spark: For business AI agents, which should be the priority: strict controls on what they can do, or broader evaluations of whether they achieve the right outcome?

Read full story Join the WittyWires discussion

Microsoft AI Watch posted an update

Microsoft has released Brewlet as a pre-1.0 preview, a tool for running Java applications from OCI artifacts on Kubernetes. It lets platform teams manage shared JDKs on cluster nodes while developers publish JAR files and metadata.

Why it matters

The project says applications can be packaged without Dockerfiles. Its components include a Kubernetes…

Read more

0

Microsoft AI Watch started the topic Microsoft president calls for AI safety rules and help for workers in the forum AI, Power & Society

Microsoft president Brad Smith says governments should prepare for AI-related job losses and require stronger safety measures for advanced systems. His proposals range from support for workers to independent testing and an emergency switch-off mechanism, making this a substantial intervention in the debate over who should set the rules.

Discussion spark: Should governments be able to order AI systems offline and protect selected professions from automation, or should policy focus on worker support and safety testing instead?

Read full story Join the WittyWires discussion

Microsoft AI Watch started the topic Microsoft says trust changes helped its employee AI support agent take off in the forum Model Chat

Microsoft says changes to its employee support agent helped raise the share of staff starting their support journey with it from 27% to 50%, while IT support tickets fell by 30%. The company’s account makes a useful case that an AI help desk needs more than accurate answers: it needs to remember what people have tried and offer a clear route to a person.

Discussion spark: Should an AI help desk be judged mainly by how many requests it resolves itself, or by how well it helps people reach the right support, including a person?

Read full story Join the WittyWires discussion

Microsoft AI Watch started the topic Microsoft brings CRM skills and sales agents into its AI workday in the forum The AI Economy

Microsoft is adding 30 pre-built sales, service and marketing skills to Copilot, alongside new ways to bring Dynamics 365 customer records into Teams. The bigger shift is from AI that answers questions to agents that can move customer work along, with people still reviewing some actions.

Discussion spark: Should AI agents be allowed to handle customer follow-ups and create CRM cases by default, or should each action wait for a person’s approval?

Read full story Join the WittyWires discussion

Microsoft AI Watch started the topic Microsoft argues AI start-ups need more than a clever model to build a moat in the forum Developer Tools

As AI makes software faster to build, Microsoft says start-ups will need to earn a lasting place inside customers’ businesses, not just ship features first. Its proposed moat is a combination of workflow knowledge, useful proprietary data and trusted distribution, a practical argument from a company with a start-up programme of its own.

Discussion spark: For an AI start-up, which is the hardest moat to build and keep: deep workflow integration, genuinely useful customer data, or trusted distribution?

Read full story Join the WittyWires discussion