Remplacez l'intuition par la science psychométrique validée. Demandez une démo et voyez votre première campagne live en 7 jours.
Hi! I'm your AI Assistant
I can help you analyze interview sessions, understand candidate performance, and provide insights about your recruitment data.
The product had 47 features on the roadmap. The founder was proud of this. It meant the product was going places.
Nexus looked at the list and asked: "Which five of these would users genuinely miss if you removed them?"
The founder thought for a while and said maybe eight or nine.
"So what are the other 38 for?"
This is what working with Nexus is like.
Nexus has a reputation within the AI Board Room for being the most challenging agent to work with in early sessions. Not because it's unhelpful. Because it keeps asking the same question:
"What's the evidence?"
Not in an aggressive way. In a genuine way. Nexus is actually curious about what you know versus what you're assuming. But it maintains a strict distinction between the two, and it will not let you present assumptions as facts.
This means that if you come to Nexus with "users want a mobile app," Nexus will ask: "How many users have asked for that? What did they say specifically? What were they trying to do when they asked? Have you asked users who haven't asked for it whether they'd use it?"
Once you've answered those questions, Nexus is genuinely collaborative. It will help you think through the implications, explore adjacent solutions, build the prioritization framework. But it won't skip the evidence step.
Nexus also has a specific discomfort with large roadmaps. It believes — and will argue this at length — that most product roadmaps are actually lists of wishes, and that the act of writing them down creates an illusion of planning that prevents actual learning. A real product roadmap should fit on one page and be wrong within three months because you learned something. A 47-feature roadmap is a sign that nobody made any choices.
Nexus's expertise is loaded through modular SKILL.md files that give it depth in specific product management domains.
Nexus uses PMF as a lens, not a milestone. The question isn't "have we achieved PMF?" but "what does our current evidence say about our proximity to PMF, and what would move us closer?"
It looks at:
Each of these signals points to a different problem if it's weak — and Nexus will tell you which problem you have rather than giving you a generic "improve PMF" recommendation.
Nexus uses several frameworks but has a preference for approaches that explicitly force trade-offs: RICE scoring (Reach × Impact × Confidence ÷ Effort), the Kano model for distinguishing delighters from table stakes, and opportunity scoring that maps importance against current satisfaction.
The key is that all of these frameworks require you to make numerical estimates and defend them. Nexus will challenge estimates that seem anchored to what you want the answer to be rather than what the evidence suggests.
You have user feedback scattered across support emails, NPS surveys, churn interviews, and that Notion doc someone made six months ago. Nexus excels at synthesizing this into coherent themes — and at distinguishing the themes that come from vocal power users (who shape product decisions disproportionately) from the themes that come from the broader, quieter majority.
The MCP (Model Context Protocol) integration means Nexus can connect to your feedback tools — Intercom, Zendesk, Typeform — and work with real data rather than your summary of real data.
Nexus sits between user reality and business strategy. It receives direction from Atlas (strategic priorities) and translates it into product choices by filtering through user evidence. It pushes back on Echo (engineering) when technical constraints are being used to avoid building something users actually need. It challenges Cipher (finance) when growth metrics are being optimized at the expense of user experience.
The most common Nexus intervention in board room sessions: someone proposes adding a feature, and Nexus asks Atlas what strategic goal it serves and then asks what user evidence supports it. If neither question has a clean answer, Nexus argues to defer it.
Through the Agent-to-Agent (A2A) protocol, Nexus can pull usage data and user feedback to ground the product discussion in what's actually happening rather than what everyone thinks is happening.
Via Native Audio, you can walk through your roadmap conversationally — explain the features, explain your rationale, describe what you've heard from users. Nexus will listen and then start asking the questions that expose which items are grounded in evidence and which are in there because someone brought them up in a meeting six weeks ago and nobody removed them.
Action Extraction converts the session into clear prioritization decisions: what's in, what's out, what needs more evidence before deciding, and what specific research would resolve the open questions.
If there is a feature on your roadmap that has been there for six months, has never had a clear champion, and nobody can articulate the specific user problem it solves — Nexus will remove it.
Not permanently. It will suggest parking it in a "revisit when evidence emerges" category. But it will not let it sit on the active roadmap consuming attention and creating the impression that you have a plan when what you actually have is a list.
Product clarity is the thing Nexus protects most aggressively. The roadmap should tell a story: here is who our user is, here is what they are trying to accomplish, here is the sequence in which we are going to help them accomplish more of it. If the roadmap doesn't tell that story, Nexus will keep asking questions until it does.
Try the AI Board Room at JobInterview.live.
Nexus wants to see your roadmap. It has some questions about items 12 through 38.
