When Can You Trust AI to Act?
Your team sees what AI can do but is unsure how much responsibility to give it. Explore how to set clear boundaries, spot mistakes, and keep people in control.
A way to discuss what to trust—and when to step in.Speaker brief · 2026
Help people make sense of changing work.
Your audience sees what AI can do. The harder questions are what to trust, who stays responsible and how work changes. I bring engineering, product and civic perspectives to that conversation.
Salesforce product leadership · Previously Microsoft
San Francisco · Conferences, panels and team sessions
Selected speaking
As event VP and panel moderator, I brought Robert Nishihara and Nik Spirin into a discussion of AI infrastructure.
Panel record ↗
2026 · Two-session breakout
Dreamforce — Secure By Design
Official session record ↗
2026 · Main stage
Product Operations Summit San Francisco
Official agenda ↗
2026 · On-demand webinar
What's Next for MuleSoft: Mid-Year Innovations & Roadmap Update
Official page ↗
2024 · Practitioner session
Dreamforce
Event reference ↗
Joint IBM / Red Hat sessionsSpeaking topics
Explore these topics for a conference, panel or team session. We can discuss the focus that fits your event.
From Copilots to Coworkers
AI is entering everyday work. What needs to change in people’s roles, expectations, and review habits so the team can make it useful?
A starting point for bringing AI into your team.Freedom to Build. Confidence to Grow.
Your teams want to move quickly, but their tools and responsibilities keep multiplying. Explore how shared safeguards can support progress across different teams.
A clearer way to balance independence and oversight.Session formats
Who this can help
Teams making sense of AI, leaders navigating change, product and technology communities, and founders working out how to earn customer trust.
Formats
Keynotes, team sessions, workshops, and panels. We can discuss the format and length that fit your event.
Planning together
We agree on the topic, attendees and what the session should help them understand.
Organizer resources
Short bio
Sparsh Agarwal is a product leader at Salesforce and MuleSoft, previously at Microsoft Azure. He writes and speaks about making AI useful and trustworthy, drawing on experience building products and helping teams navigate change. He holds an MBA from UC Berkeley Haas and studied computer science at UW–Madison.
Long bio
Sparsh Agarwal is a Director of Product Management at Salesforce and MuleSoft, where he works on helping organizations use AI with confidence. His focus is the gap between what technology can do and what people can trust it to do. He writes and speaks about the practical decisions behind adopting AI, keeping people in control, and changing how teams work. Previously, he built edge AI products at Microsoft Azure. Based in San Francisco, Sparsh holds an MBA from UC Berkeley Haas and studied computer science at the University of Wisconsin–Madison.
He also invests in technology for the physical world in his personal capacity. His public conversations include the Beyond the Backlog podcast and moderating the 2024 Haas AI Summit’s infrastructure panel with Robert Nishihara and Nik Spirin.
Beyond work, his experience includes South Asian dance, a Berkeley Board Fellows project with Performing Arts Workshop, volunteering with Food Lifeline and Girls on the Run, alumni chapter board involvement with Bay Area Badgers, AI trust and public policy, informed by resident interviews in Amsterdam and a prepared briefing on identity protection and employment in Washington, DC. His team developed Mokum, an AI-trust proposal with a scorecard and trust stamp.
Invite Sparsh to your event.
Include the event, date, audience, format, and the decision or behavior the session should change.
meetsparsh@outlook.com