Three moments
When I review my team of product managers' documents, the questions that I ask them actually help them add more valuable inputs to their work.
Not because they didn't know. Because they hadn't been prompted to think about it.
Think about the last time someone reviewed your resume. When they asked you about your achievements, didn't you suddenly see your own work differently? Your impact became real. Your accomplishments weren't vague anymore.
I believe this happens with all of us. When someone asks the right question, the answers we were struggling to articulate suddenly emerge. We weren't missing the knowledge. We were missing the prompt.
The answers were already in our heads. We just needed someone to ask.
When my manager reviewed my concept document, he asked: "Is this a correlation or a causation?"
I didn't know the difference. Not really. But that question sent me down a path where I learned it, and it changed how I framed every product hypothesis after that.
Sometimes the answer is already in your head. Sometimes the question takes you somewhere you've never been. Both only happen when someone asks.
We all, as humans, tend to miss some details or not know about some details when we are doing something. But the same us, when prompted by a friend or manager, would be able to tackle it and add those important points of view.
In 2017, I worked on an AI chatbot at DBS Bank. That's when I first saw what AI could do with the right questions. By 2020, I wanted to build an AI buddy for resume writing. I got busy with my job. The idea stayed parked. But the pattern I kept seeing never went away.
The bigger problem
As I gained more experience as a product leader, I could see this pattern at a larger scale.
Organisation structure and team topology create difficulties understanding each others' documents. There's little to no time to review peer documents. The leadership team doesn't read all the details, as their time is scarce.
And often, the author suffers alone, the sole mind behind the document, worrying if they've included all the details or not. There's a huge cognitive load to write documents that multiple stakeholders need to understand, while those stakeholders are too time-scarce to review end to end.
But I also saw something deeper. Two structural problems that no individual document review can fix.
Features don't connect to company goals
A feature might not directly communicate how it aligns with the company's KPIs or OKRs. This makes prioritisation blurry and leads to rework. Where the link is missing, the AI asks: which specific OKR does this correlate to? It brings alignment from the start, not after three rounds of stakeholder pushback.
No single source of truth for problems
Different teams focus on different customer problems, but there's no repository that brings them together. The complete picture is missing. Leaders aren't fully aware of what's happening across the org. If we maintain that single source, every PM can trace back where their solution fits in the company's big picture.
I believe AI has the potential to address these within-company alignment issues which are not usually seen or identified by companies.
Why questions, not generation
AI can help these individuals and teams lower their cognitive load and focus on ideas that really matter to customers, rather than worrying about document content and internal alignment.
Say you're writing a PRD for a feature still in discovery. You paste it in, and SPM reads what you've written:
- 1"What is the hypothesis?"Checks what's written. Finds the gap.
- 2"Is this the top riskiest assumption you really want to validate?"Hypothesis exists. Now it digs deeper.
- 3"Based on Teresa Torres, there are 3 types of hypothesis testing. Which one are you trying to validate?"Pushes toward a decision. One step at a time.
While helping to unlock potential, it also makes the users learn and evolve over a period of time. This is the opposite of Generative AI, which might make users not learn but just copy and paste. Learning based on prompts is one of the most effective ways of learning.
Who this is for
I built this for anyone involved in building products, starting with but not limited to:
- PMs and Marketers write alone, but with SPM they get an AI buddy that asks what's missing before their manager does.
- Executive Leaders digest the essence without reading every word. Assign SPM as a preliminary reviewer before they pick it up.
- AI Builders: engineers, designers, and founders are all building products now. When everyone can build, the bottleneck moves to deciding what to build and why. That's where the right question matters most.
What SPM became
That vision from September 2023 is now a live product. Super Product Manager reads your product document and evaluates it against domain-specific expectations. Every gap gets a score from 0 to 1.0. Then it asks you clarification questions about the weakest areas.
The questions escalate. If you give a vague answer, SPM comes back harder. Evidence first, then action directives, then assumptions made on your behalf. Like a senior PM who won't let you ship a half-baked spec.
Only after you've thought through the gaps does it generate improvements. And those improvements are grounded in your answers, not AI hallucination. The document sounds like you, because it came from you.
30 expert reviews across the PM lifecycle. Strategy, analysis, execution, discovery. Available as a Chrome extension for Google Docs, a web app, and an MCP server for AI coding assistants.
The clarification questions are the product.Everything else is scaffolding. The moment someone says "I hadn't thought about that" is the moment SPM delivers its value.
The outcome isn't a better document. The outcome is a better you.