Microsoft’s recent user survey — surfaced by reporting from Windows Latest — suggests the company is actively exploring ways to make Microsoft Edge’s Copilot mode behave more like Perplexity’s Comet: not just a summarizer but an agentic browser that can act across tabs and automate multi-step tasks on your behalf. The survey asks Edge users about advanced capabilities such as multi-tab context, extracting tables to Excel, drafting and sending emails, filling forms, and generating shopping lists from multiple tabs — features that would move Edge from a context-aware assistant to a browser that can do things for you. n Edge?
Copilot mode in Microsoft Edge replaces the traditional New Tab Page search box with a Copilot-centered compose box and enables an “Ask Copilot” address-bar integration that can read page content and summarize or answer questions about it. The move to a Copilot-first UI is part of Microsoft’s wider push to fold LLM-powered assistance into core OS and productivity experiences.
The opportunities are compelling: faster research workflows, one-click synthesis of multi-tab projects, and easier transfer of web data into productivity apps. The risks are real as well: privacy, security, liability for automated actions, and the user-experience costs of gating useful features behind subscriptions.
For users and administrators, the best posture is cautious optimism: explore Copilot’s new capabilities where they help, demand transparent consent and auditability where they act, and insist on fine-grained controls at the enterprise level. As Edge’s Copilot-first experiment matures, the balance Microsoft strikes between automation power, user control, and trust will determine whether agentic browsing is seen as a leap forward — or a convenience many will accept only with careful safeguards.
Source: windowslatest.com Microsoft survey hints more Perplexity Comet-like Edge browser features coming to Windows 11
Copilot mode in Microsoft Edge replaces the traditional New Tab Page search box with a Copilot-centered compose box and enables an “Ask Copilot” address-bar integration that can read page content and summarize or answer questions about it. The move to a Copilot-first UI is part of Microsoft’s wider push to fold LLM-powered assistance into core OS and productivity experiences.
What is Perplexity Comet and why doesy’s Comet is an AI-first browser model that places autonomous agents and deep session memory at the center of the browsing workflow. Comet focuses on persistent context, multi-step automation, and agentic workflows — capabilities that let it act across pages and sessions to complete tasks more like a digital assistant than a search box. Its emergence pushed incumbents to rethink how a browser can behave when AI is its operating logic rather than an add-on.
What the survey reveals (and what it likely means)
coming first-class
The survey repeatedly asks about “multi-tab context” — effectively whether Copilot should have a global view of all open tabs and be able to reason across them. That’s critical: a model that can aggregate facts from several pages can synthesize, compare, and produce consolidated outputs (e.g., a single shopping list or a comparative product shortlist). Microsoft appears to be testing how comfortable users are with granting Copilot that scope and what kinds of tasks they’d trust it to perform.Task automation: from suggestions to actions
Beyond reading and summarizing, quetease true automation capabilities:- Extract structured data (tables) into Excel.
- Draft an email from page content, and potentially send it.
- Auto-fill or complete forms using aggregated tab context.
- Generate consolidated shopping lists from multiple tabs.
Monetization signals: Copilot Pro and Journeys
The survey also probes frequency and vertical use-cas nce), and Microsoft appears to be correlating those responses with UI and prompt tuning for each vertical — a common pre-launch data-gathering step before paid feature gating. Reporting and internal signals suggest some advanced capabilities could sit behind Copilot Pro (a paid tier reportedly priced around $20), and that a feature called “Journeys” (organizing and summarizing browsing activity locally) may be part of that premium offering. This indicates Microsoft is balancing ambitious feature development with monetization.How real is this — technical realities and limitations
Context windows and scale: why “reading everything” is non-trivial
Reading manyc but not without constraints. Large language models have finite context windows, and the longer the combined content, the greater the engineering complexity to maintain coherence and avoid truncation. Practical implementations often use summarization, retrieval augmentation, and progressive context pipelines to compress and manage information. Early Copilot implementations are already context-aware, but they can struggle with extremely long or dense pages — a known limitation in similar Edge features. Expect Microsoft to rely on hybrid techniques (local summarization + cloud LLMs + retrieval) to scale multi-tab context.Local models vs. cloud models
Microsoft can reduce latency and increase privacy by running smaller local models for initial summarization and client-side reasoning and then escalating t models for complex tasks. That architecture is consistent with industry trends and has already been signalled in experimental Edge features and wider Microsoft product roadmaps. Local processing also makes features like Journeys (local summarization and grouping of browsing activity) more credible without wholesale cloud upload of private browsing content.Automation requires safe credential handling and secure actions
If Copilot is allowed to perform actions like sending emails or filling forms, it needs secure handling of credentials, explicit consent flows, safeguards. The survey hints at these directions, but actual implementation will require:- Strong authentication and consent UX for each action.
- Scoped, auditable credentials (tokenized access, revocable permissions).
- Rate limits and confirmation steps for potentially destructive actions (e.g., payments, posting).
Microsoft’s existing approach — requiring explicit user opt-in and visual indicators when Copilot accesses content — will need to be extended if the browser begins to act autonomously.
Strengths of Microsoft’s potential approach
- Integrated ecosystem: Edge as a browser embedded in Windows gives Microsoft unique levers — deep OS integration, identity linkage (Microsoft accounts), and native app interoprive) which make automation like saving tables to Excel or drafting and sending emails technically smoother than in third-party browsers.
- Unified AI UX: Moving Copilot from a sidebar into the New Tab Page and address bar unifies search, chat, and actions. That reduces friction and encourages adoption for productivity workflows. Early canary rollouts already show how these flows can bePrompt tuning for verticals: Asking respondents about use-cases (travel, shopping, coding) suggests Microsoft intends to optimize prompts, UI flows, and microcopy for each vertical — a useful step to improve outcomes and reduce hallucinations by narrowing the assistant’
Risks and downsides: what to watch for
Privacy and telemetry concerns
Any feature that requires Copilot to scan multiple tabs or access credentials raises privacy questions. Users and enterprises will want granular controls, clear disclosure of what is sent to Microsoft servers, and e entirely locally if desired. Without that, trust erosion is likely — particularly among business and privacy-conscious users. The survey’s focus on permissions suggests Microsoft is aware, but the proof will be in the UX and transparency.Automation errors and legal exposure
If Copilot can draft and send emails, book reservations, or fill web forms, mistakes could have real-world consequences. Mis-sent messages, incorrect bookings, or improper disclosures could create liability and user harm. Microsoft will need conservative defaults, explicit cvity logs showing what the assistant did and why.Paywalled AI features = fragmentation
Locking advanced agentic features behind a Copilot Pro subscription risks fragmenting user experience. Power users and enterprises may accept the fee, but mainstream users could see core OS functionality gated by cost. Balancing monetization and ubiquitous utility will be a product challenge.Security surface area increases
Allowing the browser to act across tabs and with credentials increases the attack surface. Threat actors could attempt to trick Copilot into taking actions via malicious pages or via prompt-injection techniques. Robust input sanitization, adversarial testing, and strong model guardrails will be requiredoft could and should implement agentic features (recommended blueprint)- Scoped permissions model
- Request permission per task-type (e.g., “Allow Copilot to access open tabs for summaries”).
- Provide a permissions dashboard with revocation and audit logs.
- Progressive elevation for sensitive actions
- Read-only context by default.
- Require explicit single-use confirmation for action-level automation (send email, make purchase).
- Multi-factor confirmations for financial or identity-affecting tasks.
- Local-first architecture for privacy-sensitive flows
- Local summarizer to compress content and extract metadata before any cloud call.
- “Local only” mode for users who never want browsing content sent off device.
- Explainability and undo
- Each action should come with a short human-readable summary of why Copilot made that choice and an “undo” or reversal path where feasible.
- Enterprise controls
- MDM/Group Policy settings for org-wide enable/disable of agentic features.
- Audit logs export for compliance teams.
Comparing Edge’s likely trajectory with Perplexity Comet
- Perplexity Comet and similar agentic browsers were built from day one around autonomous agents and session memory. Microsoft’s advantage is scale: OS integration, productivity apps, and a monetization channel through Copilot Pro. Edge’s path is to import agentic behaviors into an already widely deployed browser while safeguarding enterprise and consumer trust.
- Comet’s early advantage is nimbleness and a single-purpose design that raised user expectations for what a browser agent can do. Edge’s challenge is to match that agility without compromising the breadth of its existing user base. That likely explains Microsoft’s experimental, survey-driven approach: gather telemetry and sentiment before delivering broader automation.
The business angle: why Microsoft is doing this now
- AI-first competitors (Perplexity, OpenAI, Google) are redefining expectations for search and web workflows; a Copilot-first Edge is Microsoft’s strategic response to re-center users in its ecosystem and monetize advanced capabilities.
- The potential to tie browsing automation into Microsoft 365 (Excel, Outlook, Teams) creates rehat competitors will find hard to replicate at scale.
- Survey-driven feature design suggests Microsoft wants to optimize for high-impact verticals (travel, shopping, coding), where agentic automation delivers measurable value and is easiest to monetize.
Practical guidance for users and IT teams
- Regular users:
- Treat Copilot’s automation features as productivity accelerators — but keep confirmations on for any action that posts or sends data.
- Use “local-only” or opt-out settings when privacy is a concern.
- Power users:
- If automation is unlocked behind Copilot Pro, evaluate the ROI: will saved time and convenience justify the subscription?
- Test workflows with non-critical data first and verify the audit trail for actions.
- IT administrators:
- Demand granular MDM controls and auditability before enabling agentic features in corporate environments.
- Create policies for credential management (prefer tokenized access) to reduce risk.
What remains unclear or unverifiable
- Exact pricing and specific gating: while reporting and survey excerpts point to a roughly $20 Copilot Pro tier and mention Journeys as a pro feature, concrete packaging and global availability have not been publicly confirmed in final form. These details should be treated as provisional until Microsoft’s official announcements.
- Full scope of automation permitted by default: survey responses can indicate intent, but the final behavior — how much Copilot will automate versus advise — depends on design decisions that haven’t been finalized publicly.
- Precise technical stack (which components will run locally versus in the cloud) is not officially documented; references to local models and WebUI changes appear in exploratory reporting and experimental channels but are . Readers should consider these plausible architectures rather than confirmed roadmaps.
Likely timeline and next moves
Microsoft is currently testing and iterating in Canary and controlled rollouts. Expect this phased pattern:- Broader Copilot-mode rollouts with improved summarization and multi-tab prompts.
- Experimental “Actions” that perform semi-automated tasks (extract to Excel, draft emails) behind feature flags or previews.
- Premium gating (Copilot Pro) for the most agentic, value-heavy features.
- Enterprise controls and local-model options as security and regulatory concerns are ironed out.
Conclusion
The Microsoft survey is more than market research — it’s a roadmap hint that Edge could evolve from a context-aware assistant into a true agentic browser capable of multi-tab reasoning and cross-application automation. That would bring Edge closer to the functional model Perplexity’s Comet introduced, while leveraging Microsoft’s unique strengths: deep Windows integration and interoperability with Office apps.The opportunities are compelling: faster research workflows, one-click synthesis of multi-tab projects, and easier transfer of web data into productivity apps. The risks are real as well: privacy, security, liability for automated actions, and the user-experience costs of gating useful features behind subscriptions.
For users and administrators, the best posture is cautious optimism: explore Copilot’s new capabilities where they help, demand transparent consent and auditability where they act, and insist on fine-grained controls at the enterprise level. As Edge’s Copilot-first experiment matures, the balance Microsoft strikes between automation power, user control, and trust will determine whether agentic browsing is seen as a leap forward — or a convenience many will accept only with careful safeguards.
Source: windowslatest.com Microsoft survey hints more Perplexity Comet-like Edge browser features coming to Windows 11
