Microsoft's Copilot now orchestrates GPT and Claude together, Claude adds scheduled tasks, and health AI tools proliferate with real-world impact questions.
Today in AI Products
 Microsoft Copilot |
Mar 30 |
Copilot's Researcher Agent Now Blends GPT and Claude
Microsoft's 365 Copilot introduced a Researcher Agent that strategically uses both OpenAI's GPT and Anthropic's Claude models in tandem. GPT drafts initial responses while Claude critiques and refines them, creating a complementary workflow. This represents a pragmatic approach to leveraging strengths of different models rather than betting on a single provider. Source →
Designer's Takeaway: Notice how this design treats different AI models as specialized collaborators rather than competitors. Apply this pattern by considering which tasks benefit from multiple AI perspectives, and design UI that makes model selection transparent without overwhelming users.
Pattern: Collaborative AI
 Claude |
Mar 29 |
Claude Adds Scheduled Tasks Feature
Claude rolled out scheduled tasks that allow users to set up recurring AI workflows without manual intervention. This addresses a gap that other AI assistants left unresolved, letting users defer work to specific times or trigger actions based on conditions. The feature moves Claude beyond reactive chat toward proactive task automation. Source →
Designer's Takeaway: Consider how your AI product handles temporal workflows. Design clear affordances for scheduling that don't require users to understand cron syntax or complex automation logic, making automation accessible to non-technical users.
Pattern: Predictive Anticipation
| Microsoft Copilot Health / Amazon Health AI |
Mar 30 |
Health AI Tools Multiply But Effectiveness Remains Unclear
Microsoft launched Copilot Health allowing users to upload medical records and ask health questions, while Amazon expanded its Health AI beyond One Medical members. The expansion is rapid, but real-world validation of accuracy and safety lags behind deployment. This creates a critical UX challenge around trust and appropriate confidence signaling. Source →
Designer's Takeaway: Apply heightened trust calibration design when users depend on AI for health decisions. Include clear disclaimers about model limitations, confidence indicators for medical claims, and always surface when professional consultation is necessary rather than optional.
Pattern: Trust Calibration
 Perplexity Computer |
Mar 28 |
Perplexity Computer Delivers on AI Assistant Promises
Early users report that Perplexity's Computer interface finally delivers the ambient, always-available AI assistant experience that has been promised but rarely achieved. The interface handles continuous multi-turn interactions without the cognitive friction of traditional chat interfaces. It demonstrates that rethinking the interaction model fundamentally improves user experience. Source →
Designer's Takeaway: Notice how moving beyond turn-based chat creates space for more natural interaction patterns. Consider designing AI interfaces that anticipate context across sessions and reduce the ceremony of starting new conversations.
Pattern: Conversational UI
 LiteLLM |
Mar 30 |
LiteLLM Severs Ties with Delve After Security Incident
LiteLLM, a popular AI gateway that abstracts multiple model APIs, dropped its security partnership with startup Delve following a credential-stealing malware attack. The incident exposed risks when third-party services handle authentication for multi-model access. It highlights the security surface area that expands with gateway architectures. Source →
Designer's Takeaway: Consider the trust and security implications of gateway abstractions in your architecture. Design clear visibility into which third-party services handle authentication, and provide users audit trails showing what credentials are stored where.
Pattern: Privacy-First Design
Today's Takeaway
Multi-Model Orchestration and Health AI Force New Design Patterns
Two distinct trends are reshaping AI UX: first, products are moving beyond single-model reliance toward orchestrating multiple AI providers strategically (Claude and GPT working together), which demands UI patterns that make model selection and reasoning transparent. Second, health AI's rapid expansion into consumer products reveals that deployment speed outpaces validation, creating an urgent need for designers to implement trust calibration, confidence visualization, and clear escalation pathways when AI advice has real health consequences. Both trends require designers to think beyond conversational chat toward more sophisticated interaction patterns.
Want to learn more about the patterns mentioned today?
Explore All 28 Patterns →