• Product Tapas
  • Posts
  • Microsoft Pulls Bing APIs, OpenAI's Parallel Coding Agent, Google's I/O Blitz

Microsoft Pulls Bing APIs, OpenAI's Parallel Coding Agent, Google's I/O Blitz

Plus: Designing for mental health, bending the curve of luck, and navigating platform risk

We track Product so you don't have to. Top Podcasts summarised, the latest AI tools, plus research and news in a 5 min digest.

Hey Product Fans!

Welcome to this week’s 🌮 Product Tapas!

If you’ve been forwarded this or just stumbled upon it, you’re in for a treat. For the best reading experience, check out the web version and sign up for future editions here.

What’s on the menu this week? 🧑‍🍳

📰 Not Boring – Microsoft pulls the plug on Bing Search APIs while Netflix's ad strategy proves surprisingly successful. OpenAI launches Codex, a cloud-based coding agent with parallel task capabilities, as Microsoft's Build 2025 event outlines a vision for an open agentic web. Google had a massive week at I/O. Meanwhile, iPhone 17 Air specs leaked, Klarna brings back human customer service, and MasterClass tests AI clones of experts. And one classic ironic twist… Anthropic doesn't let you use AI when applying for a job there.

⌚️ Productivity Tapas – This week's tools include an AI agent for visual communication that designs dynamic content for you, an affordable self-hosted workflow automation solution, and a platform that transforms any content into rich, visual sites perfect for proposals and presentations.

🍔 Blog Bites – Discover how Headspace designed Ebb, a trustworthy AI mental health companion that balances innovation with human-centred design principles. Plus, learn a structured four-step framework for analysing qualitative usability test data from Nielsen Norman Group, and explore how teams can "bend the curve of luck" by strategically tackling uncertainty early in projects.

🎙️ Pod Shots – This week's PodShot takes a different angle, breaking down a roundtable discussion on the collapse of Synapse—a middleware provider that powered banking and payments for dozens of fintechs. Even if you're not in fintech, the lessons about platform risk, vendor oversight, and regulatory blind spots are relevant to anyone building products in regulated or complex environments.

Plenty to get stuck into - off we go! 🚀

StartEngine’s Blockbuster Year: Learn How You Can Join the Action

StartEngine is connecting accredited investors with top pre-IPO companies like OpenAI, Perplexity, and Databricks.

How’s it going? They just posted record Q4 revenues and 2x revenue YoY.

Better yet, the window is open for you to support this growth directly. Join 50K+ who have invested $84M+ in StartEngine.

Reg A+ via StartEngine Crowdfunding, Inc. No BD/intermediary involved. Investment is speculative, illiquid & high risk. See OC and Risks on page.

📰 Not boring

  • Microsoft pulls plug on Bing Search APIs

  • If you hate ads on Netflix, I have bad news: they’re doing really well

  • OpenAI launches Codex, a cloud based coding agent that can parallel task; not sure what this means for it’s purported purchase of WindSurf 🤔 

  • Study shows emergence of social norms between AI Agents

  • Microsoft held its Build 2025 event - outlining a vision for an open agentic web

  • Google has had a massive week announcing a stack of stuff at its I/O event

    • Here’s a NotebookLM that covers everything announced; some highlights included Google's 'universal AI assistant' prototype, real-time AI camera sharing in Search, a new way to hold virtual meetings, and Gmail smart replies. Honestly it was massive. Kudos to Google.

    • They also rolled out its suite of creative AI models: Veo2 (video), Gemini 2.0 (image creation & editing) Imagen3 for photorealistic visuals

  • iPhone 17 Air specs leaked (light and thin; some concerns over battery performance)

  • Klarna changes its AI tune and again recruits humans for customer service

  • Shein & Temu ramp up advertising in UK and France as US tariffs hit

  • MasterClass is beta testing a product where subscribers can talk with AI clones of experts

  • Replit launches an interesting Notion integration

  • Anthropic doesn’t let youuse AI when you apply for a job there

⌚️ Productivity Tapas: Time-Saving Tools & GPTs

  • PageOn: Cursor for visual communication. AI Agents understand, research, and design dynamic visuals for you. Tell them your goal—they’ll plan, research, and design dynamic, stunning pages. It’s visualised Notion for storytellers. Build like Lego, visualise with AI

  • Railway: deploy self hosted n8n for $5 per month. Bargain.

  • Faces: Turn any content into rich, visual sites that impress. Perfect for proposals, presentations, or ideas that need to land

Remember, as a Product Tapas Pro subscriber you can access the full time saving tools database for fast approaching 400 time-saving tools relevant for product managers and founders 🔥.

Check the link here to access.

🍔 Blog Bites - Essential Reads for Product Teams

UX & AI: How Headspace Designed Ebb, a Trustworthy AI Mental Health Companion

In a recent Figma blog, Matt Alagiah explores how Headspace created Ebb, an AI companion for mental health support that balances technological innovation with human-centred design principles. The case study reveals how the team navigated the challenges of building an AI tool for the sensitive domain of mental health. Read the full article here.

💡 "We didn't want AI to be invisible. We wanted a member to know they were interacting with AI, not a human." This commitment to transparency from Product Design Manager Priyanka Marawar highlights how Headspace prioritised ethical design principles in creating Ebb.

Key Takeaways:

Trust Through Transparency: Headspace deliberately made Ebb's AI nature clear to users, avoiding deception while still creating an empathetic experience.

Collaborative Design Process: The team used FigJam workshops to align stakeholders and "playgrounds" to explore visual identities, bringing together designers, copywriters, and product teams.

Thoughtful Personification: Ebb was visualised as a "friendly blob" with careful attention to emotional expression—present but not overly cheerful—to create a safe space for vulnerability.

Inclusive Naming Strategy: The team deliberately avoided feminine names (common for AI assistants) to prevent reinforcing stereotypes about caregiving and mental health.

Safety-First Approach: User controls like the ability to exit or delete conversations were prioritised, along with clear crisis resources and transparent data practices.

Iterative Testing: Figma prototyping enabled continuous user testing of different aspects of Ebb, from visual identity to conversation design, ensuring the experience felt safe and supportive.

Consistent Brand Experience: A centralised "story kit" in Figma ensured visual and messaging consistency across all touchpoints, building trust through coherence.

Figma Blog

UX Research: Analysing Usability Test Data in 4 Steps

Here’s an interesting piece from Maria Rosala (Nielsen Norman Group) where she presents a structured framework for analysing qualitative usability test data, moving beyond simple issue cataloging to develop meaningful insights from complex research scenarios. Read the full article here.

💡 "Analysing data from qualitative usability testing is often more complex than it's portrayed. This type of data is rich, nuanced, and messy." This insight highlights why a systematic approach to analysis is crucial for extracting reliable conclusions from user research.

Key Takeaways

Four-Step Analysis Framework: The article outlines a systematic approach to usability data analysis: collect relevant data, assess for accuracy, explain the data, and check for good fit between explanations and evidence.

Analysis vs. Synthesis: Effective research requires both breaking down information (analysis) and recombining it into meaningful insights (synthesis), with researchers often moving back and forth between these activities.

Contextual Assessment: Not all data points should be weighted equally—researchers must consider factors like study design, participant recruitment, and facilitation techniques when evaluating observations.

Hypothesis Testing: The framework encourages researchers to generate multiple possible explanations for observed behaviours, then test these explanations against the full dataset to identify the most reliable interpretation.

Iterative Refinement: The process isn't linear—researchers typically form initial explanations based on memorable data points, then refine these explanations as they test them against the broader dataset.

Embracing Uncertainty: Sometimes analysis will raise more questions than answers, and it's acceptable to conclude that more research is needed rather than forcing definitive conclusions.

Triangulation: Complex research questions require gathering multiple data points, weighing them against study design factors, and combining information from different sources to provide trustworthy answers.

Maria Rosala

Research: How to bend the curve of luck

Finally this week, in a recent Trigger Strategy piece Tom Kerwin and Corissa Nunn explore how teams can increase their chances of success by strategically tackling uncertainty early in projects, creating an "asymmetry of adaptiveness" that turns potential disasters into opportunities. Read the full article here.

💡 "Antifragile prioritisation means this: Your project grows more likely to succeed when you prioritise getting signals from your areas of greatest uncertainty."

The article illustrates these principles with real examples, including how a recruitment-tech startup avoided months of wasted work on a Calendly integration by discovering early that customers weren't interested, and how another team quickly pivoted when they discovered API documentation was incorrect.

Key Takeaways

Early Uncertainty Exploration: By identifying and investigating major unknowns at the beginning of projects (like API capabilities or customer interest), teams can discover potential roadblocks before investing significant resources.

Multiverse Mapping: This collaborative technique helps teams visualise customer journeys and identify critical uncertainties in just 30-40 minutes, creating shared understanding without extensive research.

Build-Measure-Learn Simultaneously: Rather than following a linear process, teams can work on all three aspects concurrently by reordering their existing work to prioritise areas of greatest uncertainty.

Practical Validation: Simple tests (like emailing customers with an offer) can quickly validate assumptions before committing to months of development work.

The Asymmetry of Adaptiveness: Discovering problems early creates an asymmetric advantage—teams have ample time to adapt without the constraints of previous decisions or sunk costs.

Human Psychology Leverage: The approach works because once uncertainty is explicitly acknowledged, people have a natural drive to resolve it, turning avoidance into action.

Manageable Uncertainty Chunks: Breaking down the "vast fog of uncertainty" into relevant, bite-sized puzzles makes it easier for teams to take concrete action.

Tom Kerwin Corissa Nunn

🎙️ Pod Shots - Bitesized Podcast Summaries

Remember, Product Tapas Pro subscribers get access to an ever growing database of all the top Podcast summaries we’ve ever created. 

Check it out here

💥 Synapse's Collapse: What Founders and Product Leaders Need to Know

This week's podshot is a little different. Instead of the usual deep dive on product tactics, I'm breaking down a roundtable discussion on the collapse of Synapse—a middleware provider that powered banking and payments for dozens of fintechs. If you're not in fintech, don't tune out: the lessons here about platform risk, vendor oversight, and regulatory blind spots are relevant to anyone building or scaling products in regulated or complex environments.

Fintech Business Weekly

 🎙️ Listen to the full episode here

📆 Published: May 9th, 2025

🕒 Estimated Reading Time: 3 mins. Time saved: 41 mins🔥

⁉️ What happened?

Synapse was a key infrastructure player, connecting fintechs, banks, and end users. When it filed for bankruptcy, funds were frozen for thousands of consumers and businesses, exposing gaps in oversight, accountability, and operational resilience. The fallout has triggered industry-wide soul-searching and policy debates that go far beyond banking.

🏦 Accountability in Complex Partnerships

A central theme: when you build on third-party infrastructure, you can't outsource ultimate responsibility. In Synapse's case, banks like Evolve contracted out critical services but still held the legal and ethical obligation to end users. The same logic applies if you're relying on any external platform or vendor for core product functionality—whether it's payments, cloud hosting, or data processing.

Key Takeaways:

  • You can delegate operations, but not accountability. If your product depends on a partner, your users will hold you responsible when things go wrong.

  • Complex, multi-party relationships multiply risk and make it harder to assign blame or fix issues quickly.

  • This is true in any sector: SaaS, healthtech, logistics, or marketplaces—if you're building on someone else's rails, you need to understand the risks and have contingency plans.

🕵️‍♂️ Regulatory and Oversight Gaps

The Synapse collapse exposed how regulators struggled to keep up with new business models and technology stacks. Even when they had the authority to intervene, they often didn't act until it was too late. This isn't just a fintech problem—regulatory lag is a challenge in sectors like health, mobility, and AI.

Key Takeaways:

  • Don't assume regulators (or your partners) will catch problems before they impact your users.

  • If you operate in a regulated space, stay proactive about compliance and risk—don't wait for the rules to catch up.

  • Transparency and clear communication with users about how their data, money, or assets are handled is critical, especially when using third parties.

🔄 Policy Overreactions and the Danger of One-Size-Fits-All Rules

After a high-profile failure, there's a tendency for policymakers to overcorrect, creating broad rules based on worst-case scenarios. The discussion highlighted how this can stifle innovation and punish responsible actors, whether in fintech, healthtech, or any sector where new models challenge old frameworks.

Key Takeaways:

  • Be prepared for regulatory swings after industry failures—sometimes the new rules will be blunt instruments.

  • Advocate for nuance: not all business models or vendor relationships carry the same risk.

  • Document and demonstrate your own best practices to avoid being swept up in over-broad regulation.

🧾 Transparency and User Trust

One of the most practical lessons: users (and even founders) often have no way to verify if their assets or data are truly safe when intermediaries are involved. In Synapse's case, even diligent consumers couldn't confirm if their funds were insured. This "black box" problem is common in any industry with layered vendors or platforms.

Key Takeaways:

  • Try and build transparency into your product—make it easy for users to understand where their assets or data are, and what protections exist.

  • Don't rely on "trust us" messaging; provide evidence or third-party validation where possible.

  • If you're using critical vendors, audit them regularly and be ready to explain your safeguards to users and stakeholders.

🏗️ Vendor and Platform Risk: Not Just a Fintech Problem

The explosion of middleware and platform providers isn't unique to banking. Whether you're in e-commerce, logistics, healthcare, or AI, the proliferation of third-party vendors can outpace your ability (and regulators' ability) to supervise them. The panel's advice: liability and responsibility should ultimately rest with the company serving the end user.

Key Takeaways:

  • Vendor risk management is a core product and business function, not just a compliance checkbox.

  • If your product's reliability or compliance depends on a vendor, treat their risk as your own.

  • Build redundancy and contingency plans for critical dependencies—don't wait for a crisis to test your resilience.

📚 Lessons for Founders and Product Managers Across Sectors

The Synapse story is a wake-up call for anyone building in complex, regulated, or high-stakes environments. The main message: design your business and product for the world as it is, not as you wish it were. That means understanding your dependencies, staying ahead of regulatory change, and building trust through transparency and operational excellence.

Key Takeaways:

  • Don't assume someone else (regulator, vendor, or partner) will protect your users or your business.

  • Regularly review your risk exposure—especially where you rely on third parties for core functionality.

  • The absence of disaster is the real benchmark for operational success, not just growth or innovation.

 🎙️ Listen to the full episode here

That’s a wrap.

As always, the journey doesn't end here!

Please share and let us know what you would like to see more or less of so we can continue to improve your Product Tapas. 🚀👋

Alastair 🍽️.

Reply

or to participate.