Clicks whisper, prompts yell: the future of SaaS UX | Mickey Alon (Foldspace AI)

This episode's guest is Mickey Alon, co-founder and CEO of Foldspace AI. Mickey has spent 20+ years building products people actually use: he founded Insightera (acquired by Marketo), created Gainsight PX and co-authored the first book on product-led growth. Now he's betting that AI agents will completely change how we interact with software - and he has real usage data to back it up. If you're building a SaaS product and wondering what AI means for your UX, analytics and roadmap, this one is full of real numbers and hard-won lessons.🧠 Chapters:0:00 - Mickey's story: two exits and why he founded Foldspace AI6:02 - Why AI agents should amplify customer success instead of replacing it8:32 - Clicks whisper, prompts yell: tapping into user intent for the first time11:14 - Why agentic interfaces make product experiments dramatically easier11:50 - What's broken with product analytics and how intent data fixes it14:35 - Implicit vs explicit usage and why adoption is the best cure for churn17:52 - The Salesforce lesson: why UI revamps end up lipstick on a pig20:42 - How agents rewrite the PLG playbook by eliminating friction completely22:58 - Why headless SaaS is a myth and visual responses beat text29:07 - Why "generate with AI" buttons are designed to fail31:55 - MCP is a new name for API and your UI is context you get for free36:44 - The new AI-native stack: deterministic backend, agent loop, agentic interface42:01 - Building user trust with human-in-the-loop before risky changes45:08 - The 70/30 finding: users ask the agent to act, not to answer48:51 - Speed of outcome: why your first prompt should deliver 80% of the result52:45 - Reviewing AI-written code and testing agents with LLM as a judge1:00:00 - Mickey's favorite tools: Claude Code and the agentic CRM his team built1:07:51 - Prompts are the new clicks and the 3-second rule for AI speed💡 Actionable takeaways from Mickey- Mine your agent's chat logs for intent, not just clicks. Users literally tell the agent what they're trying to achieve, so you finally get demand, sentiment and problem descriptions in their own words - insight click tracking can never give you.- Think twice before your next UI revamp. Mickey watched Salesforce spend over a year on one: new users said it looked nicer, existing users hated relearning everything, and the complexity stayed. Ask first whether an agentic layer solves the actual problem faster.- Never ship a bare "generate with AI" button. Without context it's like hitting "I'm feeling lucky" on Google. Embed the agent inside the product where it knows the page, the data and the user's settings for free - that context is what makes it succeed.- Show visual responses, not walls of text. In Foldspace's measurements, visual answers get dramatically more engagement than text because we read a KPI trend in milliseconds. Worth it because engagement is what keeps users coming back to the agent.- Add human-in-the-loop confirmation before any create, update or delete. Showing "here's what I'm gonna do" before applying protects users from missing context, and trust is the whole game: lose it once or twice and users are gone.- Keep agent responses under 3 seconds. Foldspace sees users disengage beyond that, so speed is a UX feature, not an infra detail.- Track prompts-to-outcome as a KPI. Mickey's bet: prompts will replace clicks as the core UX metric, so start measuring it now.- Test your agents with scores and LLM-as-a-judge, not pass/fail. AI is non-deterministic, so measure whether behavior is getting better or worse in percentages - it's how Mickey's team ships fast without losing control.