Second Axis

PMs in AI: Automation, Data, and the Future

Date: September 25, 2025 • Author: Second Axis • Category: Event Recap

Landscape event photo
Small event photo 1
Small event photo 2
Small event photo 3

AI is rapidly reshaping how product managers work — not in some distant future, but right now. At our recent PMs in AI session in Boston, two founders, Geoffrey Morsotti (MIT Orbitz, now Hayark) and Agrim Mitha (Founder of Undergraduation.com), shared candid insights on the tools they use, the bottlenecks they face, and how they see the PM role evolving.

What emerged was clear: PMs aren't looking for AI to replace their judgment. They want AI to handle the grind — the repetitive, time-consuming work of ticket prep, reporting, and data collection — while they focus on strategy, storytelling, and decisions that still demand a human touch.

Automating Sprint Rituals

Sprint prep is one of the most time-consuming rituals in product management. Agrim described how his team used to manually create every task and check-in daily. With Asana, Make.com, and AI-powered automations, subsequent tasks now generate automatically based on dependencies, and daily check-ins are summarized for him.

"Earlier, we created every task manually. Now, AI automates subsequent tasks—like onboarding flows—and it saves us huge amounts of time." — Agrim Mitha

Geoffrey added that protocols like Model Context Protocol (MCP) and A2A can unify multiple tools — GitHub, Notion, Asana — into a single context-aware system. This allows PMs to cut down on repetitive syncs and instead focus on decision-making.

"MCP and A2A protocols let us connect repos, tasks, and docs in a way that cuts out countless meetings." — Geoffrey Morsotti

The message: sprint prep is no longer just about writing tickets. It's about building a connected workflow where AI drafts and organizes, and PMs review and approve.

Smarter Storytelling with Data

Collecting data is easier than ever — AI can automate daily usage metrics and retention reports. But when it comes to presenting those insights, PMs emphasized that humans still matter.

Agrim illustrated the importance of framing:

"Instead of saying retention is 20%, I'll frame it as one out of five people stayed. That's the human role — making the data simple and meaningful."

Meanwhile, Geoffrey highlighted the leaps in analysis tools like Tableau, which now allow natural language queries:

"You can now talk to your data. Throw it into Tableau, ask a question, and get trends in seconds."

The consensus: AI can crunch numbers, but PMs still own the narrative — tailoring the story for executives, investors, or engineers.

Staying Ahead of Faster Development

With dev tools accelerating output, teams can now test ideas in days — building MVPs with Cursor, prototyping with Base44, or launching quick ad tests on Reddit. But this speed creates a new challenge: feedback loops often lag behind.

"Speed without visibility just moves problems forward in time." — Geoffrey Morsotti

Both panelists stressed the importance of explainable alerts. PMs don't want vague red flags; they want signals that specify which cohort, which release, and why metrics shifted. Without that context, proactive systems risk becoming noise.

Will AI Replace PMs?

The audience pushed on the big question: what happens to PM jobs as AI takes over more tasks?

Both Geoffrey and Agrim agreed: PMs won't disappear. Instead, they'll manage AI agents. Geoffrey cited MIT research showing that the best outcomes come from humans and AI working together. Agrim stressed that presentation, stakeholder alignment, and trust will always need a human touch.

"Data collection can be automated, but tailoring the story for your audience is still human." — Agrim Mitha

New Skills for PMs

As AI seeps deeper into workflows, PMs need to skill up. The advice? Learn to:

  • Set up MCP servers to unify tools.
  • Use no-code automation tools like Make.com or Zapier.
  • Develop prompt literacy to guide AI effectively.

As Geoffrey noted, today's PMs don't need to be deep coders, but they do need to understand how to wield AI tools with context.

Key Takeaways

  • Automation is real: Sprint tasks, check-ins, and data collection can now be delegated to AI.
  • MCPs unlock workflows: Protocols that connect tools can eliminate repetitive syncs.
  • Storytelling is still human: PMs remain the translators of data into decisions.
  • Proactivity matters: Alerts must be explainable, not just noise.
  • The role is shifting: PMs will increasingly manage AI agents, not just projects.

The future of product management isn't AI vs. PMs — it's AI with PMs. And those who embrace the shift will spend less time chasing tickets and slides, and more time leading strategy.