You’re joining QuillAI, a Series C B2B SaaS company that sells an AI writing assistant to mid-market and enterprise customers. QuillAI has 3.2M monthly active users across web and Chrome extension, with ~420K weekly active users. Revenue is primarily seat-based ($18–$35/user/month) with enterprise contracts. QuillAI competes with Grammarly Business, Microsoft Copilot, and Notion AI.
QuillAI’s leadership believes the company is hitting a “local maximum”: teams keep shipping features (tone controls, templates, meeting notes) but core user outcomes aren’t improving. The CEO wants QuillAI to become “the system of record for workplace writing,” but the product org lacks a consistent way to ensure research creates durable knowledge that compounds over time.
QuillAI serves three primary personas:
| Persona | Share of WAU | Primary Job-to-be-Done | Current Behavior | Key Friction |
|---|---|---|---|---|
| Sales reps | 38% | Draft outreach that gets replies | Heavy template use; copy/paste into CRM | Output feels generic; compliance concerns |
| Customer support agents | 34% | Resolve tickets faster with correct tone | Use rewrite + summarization | Hallucinations; inconsistent brand voice |
| Product/engineering managers | 28% | Write specs and updates clearly | Use long-form drafting | Hard to cite sources; trust issues |
Competitive pressure is rising: Copilot is bundled into existing Microsoft contracts, and Grammarly is pushing “AI governance” features. Enterprise buyers increasingly ask for auditability, consistency, and measurable productivity gains.
Over the last 2 quarters:
Your research team runs many studies, but stakeholders complain:
The VP of Product asks you to propose a plan that ensures research contributes to QuillAI’s internal knowledge base and drives better product decisions—while still moving fast.
In this interview, assume you are the PM owning Activation & Early Retention. Walk through how you would:
You do not need to write PRDs. Focus on how you’d ensure research produces compounding learning and better prioritization under real constraints.