This Former Google PM Is Changing How Project Teams Operate With an AI PM Tool Built on the Back of AI Agents
AI in Project Management: LLMs are enhancing project management by automating status updates and freeing managers to focus on strategy.
Efficiency Gains: AI workflows reduce response times for product teams, allowing them to spot and resolve issues more quickly.
Data-Driven Decisions: ThriveAI enables informed decision-making by continuously monitoring metrics and analyzing user feedback.
Evolving Rituals: Delivery practices are shifting as AI automates documentation and monitoring, enhancing team collaboration.
Guardrails Necessity: Effective AI integration requires understanding its limitations and building appropriate workflows for optimal results.
We caught up with him to understand how his AI-native organization handles its projects. Here's what he told us.
How LLMs Connect Teams and Tools in Project Management
Traditionally, project managers created value by connecting dots across teams and tools. LLMs are the first technology that can actually help with that at scale.
Instead of chasing status updates or manually pulling reports, AI can now take on that overhead, freeing project managers to focus on guiding strategy, building alignment, and keeping teams focused on outcomes — while also increasing their leverage to drive a bigger impact across more projects without burning out.
Unlock for Free
Create a free account to finish this piece and join a community of forward-thinking leaders unlocking tools, playbooks, and insights for thriving in the age of AI.
Already a member? Log In
It's a great time to be a PM, so be excited! It’s rare to have a moment where the whole status quo is up for grabs and new capabilities are dropping every day.
Be excited! It’s rare to have a moment where the whole status quo is up for grabs and new capabilities are dropping every day.
From Engineer to Product Manager to AI Founder
I started as an engineer, then led product and engineering at scaleups, before joining Google as a product manager.
At Google, I worked closely with program managers, and I dreaded projects that didn’t have one — not every team there gets staffed with one! And things always slipped without a program manager.
I took that frustration and built ThriveAI, an AI junior PM baked with product best practices and custom workflows to help product teams stay ahead of the chaos.
How Agentic Workflows Save Project Teams Time and Money
Thrive is built on AI agents. We started with anomaly detection, feedback clustering, funnel health, and even ad-hoc tasks like summaries and one-off analyses. These are repetitive and high volume, so perfect for AI.
Our agentic workflows mean that product teams that use our product — which includes us — now spot issues within hours instead of days, fix usability problems before they spread, and ship confidently without waiting on manual checks. Our agents have reduced blind spots across the org and given teams more time to focus on building.
Here's a good example of a successful agentic workflow. We built an AI-native workflow for a large marketplace with tens of millions of daily actives. Their team was drowning in user feedback, and that meant that early-warning signs were getting buried.
So, we built an AI agent that reads every note, surfaces patterns, and flags anomalies. And it has become a core delivery system: Our client's team relies on it to catch usability issues early, long before they spread, saving them tens of thousands of dollars each month and keeping delivery on track without adding headcount.
Using AI to De-Risk Product Bets and Improve Decision Making
Building, prototyping, and designing software is easy now. The hard part is figuring out what to build. Features are commodities. What matters is making the right bets.
I spend almost no time on status or reporting anymore. My focus is on using AI to de-risk bets and increase the odds of success.
So, I use my product to get a constant read on what's working and what's breaking. In practice, that means Thrive keeps an eye on product metrics, user feedback (email, survey, in-app, etc), and telemetry (via PostHog). It surfaces issues like onboarding friction in a user segment after release or sudden funnel drops. It also gives me a short weekly pulse in Slack that highlights trends and risks — it's like a friendly teammate dropping in with an update.
The result is that my decisions are grounded in data where it's available, and in product sense where it's not.
How AI Is Transforming Delivery Rituals in Project Management
Delivery rituals are changing for most of us. At Thrive, AI drafts the first cut of almost everything — project briefs, roadmap items, you name it. Then, humans refine it.
Here's the breakdown:
- Alignment happens through Slack nudges instead of big meetings.
- Validation comes from Thrive's AI monitoring telemetry and feedback in real time, and pulling data for ad-hoc analysis — I just touched on this above.
- On execution, AI keeps an eye on progress and raises flags when something looks off. And humans then decide what to do about it.
An AI Tech Stack: Background AI and Slack-First Workflows
Our stack has shifted away from staring at dashboards and sitting in status meetings to background AI plus Slack-first workflows. Here's our project delivery tech stack:
- Slack
- Linear
- GSuite
- PostHog
- Thrive
And here's how they interact.
- Slack is our hub for updates, alerts, and managing Linear
- Linear handles our tasks and backlog, with AI doing triage (assignees, duplicates, etc.). It's synced with Slack (employee communication software) and PostHog.
- PostHog handles telemetry and product metrics, piping it into Slack.
- Thrive then analyzes what PostHog pipes in for anomaly detection, reporting, and weekly pulses. And these are again delivered via Slack. For the most part, it works in the background, learning my preferences, and only pinging me when something truly needs attention — a broken flow, rising user sentiment, whatever it is.
- GSuite - as our digital asset management system for docs and comms.
All of these tools play an integral role. But I've gotta say I'm a big fan of Linear’s new AI updates. The auto-triage, in particular, has been impressive. It suggests assignees, labels, and even links duplicates. I can’t imagine anyone actually enjoying that kind of work. Now, it just happens in the background and my team can stay focused on the real work.
Why AI Needs Guardrails and Scaffolding in Project Workflows
This is all great, but I'll say this: LLMs can be so unpredictable! One moment, they blow you away; the next, they miss something a fifth grader would catch.
The trick is knowing where they shine and designing workflows around that. And then building scaffolding and tools for the areas where they don’t — like chunking long inputs, validating numbers, or wiring it to pull real data instead of making things up.
LLMs can be so unpredictable! One moment, they blow you away; the next, they miss something a fifth grader would catch. The trick is knowing where they shine and designing workflows around that. And then building scaffolding and tools for the areas where they don’t.
At Thrive, AI fully owns reporting, status updates, and drafting roadmap items. Humans bring domain expertise, relationships, and ruthless prioritization.
Because here's the thing: Everyone has access to the same AI, so it can’t be your edge. Your differentiator is how you set priorities, build trust, and make the tough trade-offs.
That’s where you need more world-class humans.
The Future: AI Managing Incremental Product Improvements
Before long, AI will fully own incremental product improvements. For every 1–2% optimization, AI will design experiments, run them, measure results, and iterate on its own.
Humans will only focus on the big, high-stakes bets.
Project Managers Need to Get Off the Sidelines With AI
My advice is simple: Stay curious and self-aware. Know your strengths and lean on AI to amplify them. Be honest about your weaknesses and let AI cover you — it’s that teammate who never judges and always shows up.
Most of all, don’t sit on the sidelines. This is a once-in-a-career chance to help shape the future of work.
Follow Along
You can follow along on LinkedIn as Ishwar builds Thrive and changes how project teams operate. And check out ThriveAI!
More expert interviews to come on The Digital Project Manager.
