Microsoft is building a direct line from its neurodivergent employees to its product teams, and early feedback from that channel has already prompted the company to explore new ways to keep Copilot from overwhelming users with too much information. The initiative, detailed in a July 16 Inside Track blog post, marks a shift from ad-hoc accessibility reviews to a structured co-design process that prioritizes cognitive ease alongside traditional accessibility criteria.
What sets this apart is the focus: testers weren’t checking for screen-reader compatibility or color contrast. They were evaluating something more subtle—how the software feels to use—measuring cognitive load, clarity, and whether the product helps you recover when you hit a dead end.
A New Feedback Channel, Backed by Lived Experience
The first round of feedback, which ran during Neurodiversity Celebration Week, connected 150 neurodivergent employees with 10 products across Microsoft. Testers worked through real-world scenarios, recorded their screens, and described pain points as they happened. Then they filled out usability surveys that rated each experience.
Those surveys didn’t focus on classic accessibility.
Instead they homed in on factors that conventional checklists miss: how much mental energy each task required, how easy it was to navigate, whether the instructions made sense at a glance, and—crucially—what happened when something went wrong. Karl Niblock, an architect in Microsoft’s Engineering and Architecture Group Security team, described a common pattern: “dead-end workflows, where a user follows the process exactly as instructed but ends up stuck, with no obvious next step.”
For someone with dyslexia or autism, decoding similar-looking instructions over and over can be exhausting. But, as Niblock noted, clearer recovery paths and plain-language guidance reduce that friction for everyone.
Microsoft’s chief accessibility officer, Neil Barnett, framed the initiative as a move beyond compliance: “Accessibility is strongest when it is informed by lived experience, and initiatives like our product feedback sessions make sure those voices directly influence how our products evolve.”
The program gives product teams one unified path to engage with neurodivergent testers—replacing the old model of informal requests sent through employee inclusion networks. A tester pool is being built, and guidelines are being drawn up to make sure testing scenarios reflect real user needs, not hypothetical edge cases.
Where Copilot Fell Short and What’s Being Explored
Among the most concrete product findings: testers said Microsoft 365 Copilot can throw back so much information that it becomes hard to process. The issue wasn’t accuracy or capability; it was control.
Audrey Aday, a UX researcher on Microsoft’s AI Design team, said testers made it clear that “the issue wasn’t capability, it was control.” Copilot’s detailed, multi-paragraph responses—often helpful in one context—could feel overwhelming when a simple answer was all someone needed.
In response, product teams are now exploring what control might look like. That includes custom instruction menus that let users tell Copilot how much detail they want, smarter defaults that adapt to the type of query, and adaptive personalization that learns an individual’s preference over time—without requiring extra configuration on the user’s part.
None of this has been released, and Microsoft stressed that these are exploratory efforts, not committed features. Still, the fact that the AI design team is publicly discussing such controls suggests it’s being taken seriously.
What This Means for Everyday Users and IT Pros
If you’re a regular Microsoft 365 user, nothing changes today. Copilot still responds the same way. But the conversation around detail controls signals that Microsoft is paying attention to how people actually process AI-generated information—not just whether the answers are correct.
For power users and admins, this is a leading indicator. Enterprise software is full of dense settings pages, ambiguous error messages, and workflows that assume the user will infer the next step. Those pain points don’t only affect neurodivergent employees; they trigger help-desk tickets and slow down adoption. A formal feedback program aimed at cognitive friction could, over time, reduce the support burden that comes with confusing interfaces.
Microsoft’s shift towards “building with, not just testing for” neurodivergent employees also means product teams may catch cognitive bottlenecks earlier in the development cycle. That could lead to fewer post-release patches and a smoother rollout experience for everyone.
From Bug Bash to Structured Co-Design
This didn’t start as a corporate mandate. The program grew out of a grassroots idea from software engineer Jordan Cowe, who suggested an annual neurodiversity bug bash—a focused event where neurodivergent employees could stress-test products before launch. Cowe, who works on the Copilot Engineering team, observed that “if something affects a neurodivergent employee, it likely also affects many people who don’t identify as neurodiverse.”
The first bug bash proved so valuable that Microsoft decided to make it permanent, folding it into a broader co-design model. Now, instead of product teams reaching out to inclusion networks on an ad-hoc basis, there’s a repeatable pipeline: a tester pool, a standard feedback format, and a direct line to product groups.
That model mirrors work already happening in Microsoft’s Inclusive Tech Lab, where design teams partner with disabled advisors early in a product’s life. Sarah Heinzen, a senior designer, said the best time to engage is “when employees are starting something new or revising old designs."
No Immediate Copilot Changes – But Here’s What You Can Watch For
Since no Copilot features have been announced, there’s no new setting to toggle or policy to configure. If you’re interested in influencing future design, your best bet is to submit feedback through the built-in Feedback Hub or Microsoft 365’s feedback channels—the same routes that typically feed into product planning.
For IT administrators, keep an eye on the Microsoft 365 roadmap and the Admin Center for any items related to “custom instruction menus” or “adaptive personalization” in Copilot. Those are the terms the AI design team used, and they may surface as private previews before a wider rollout.
In the meantime, you can look for existing controls that already help manage information density in Microsoft 365. Some apps, like Outlook and Teams, offer focused or simplified views that strip away non-essential UI elements. While not directly tied to Copilot, they reflect the same principle of reducing cognitive load.
The Road Ahead for Inclusive AI
This program is just one ripple in Microsoft’s broader push to embed neurodiversity into product design, building on its Neurodiversity Hiring Program and investments in inclusive employee experiences. The fact that Copilot was a focus area signals that AI interfaces—where the output can vary wildly in length and complexity—are ripe for personalization.
Don’t expect a sudden Copilot overhaul. But do expect incremental changes: clearer next-step prompts, less repetitive instructions, and, eventually, a way to tell Copilot, “Just give me the highlights.” When that happens, it won’t just be good news for neurodivergent users. It’ll be a better experience for anyone who has ever felt buried by an AI’s overeager reply.