Club AI
Talk to the intelligence layer of The Quantum Club as an assistant that knows your search, cites its sources, and can take action when you confirm it.
What Club AI is
Club AI is the intelligence layer of The Quantum Club. You meet it everywhere — match scores, offer analysis, cover letters — but you can also talk to it directly as an assistant. Unlike a generic chatbot, it answers from your own context inside the platform: your profile, your applications and their stages, your tasks, and your meetings. It cites the in-app facts it used, and when you ask it to do something, it can act on your behalf after you confirm.
The four modes
Choosing the right mode is most of the skill. A toggle in the prompt box switches between them, and the placeholder text changes to tell you which one is active.
Normal is the default. The assistant answers directly from your in-app context, cites the facts it used, and recommends a next action. Use it for quick questions where you already trust the context — "what stage is my application at Acme?" or "what should I work on next?"
Search is for questions whose answer depends on facts the platform does not hold — companies, markets, current context. It uses your in-app context first, then flags what needs outside verification rather than inventing it. Toggle it on when freshness matters.
Deep Think trades speed for rigor. The assistant reasons longer, stays structured, and is explicit about its assumptions before it concludes. Use it for judgment-heavy questions: comparing two offers with different risk profiles, planning a pivot, deciding how to handle a negotiation. Expect a slower, substantially better answer.
Canvas produces work as a concrete artifact you can see and iterate on, instead of a reply in the chat. Use it when the output is a thing rather than an answer: a plan, a document draft, something with shape.
A practical rule: facts you have → Normal, facts you need → Search, decisions → Deep Think, artifacts → Canvas.
What it knows about you
Club AI is scoped to you. It reads only your own candidate context — never another member's, and never internal platform data:
- Your profile — title, current company, location, availability, and how complete your profile is.
- Your applications — each one's role, company, stage, and match score, most recent first.
- Your tasks — what is open, its priority, and what is due.
- Your meetings — upcoming sessions and their summaries.
Because the answer is built from this, a thin or out-of-date profile produces a thinner answer. Keeping your profile current is the single biggest lever on answer quality.
Citations you can check
Club AI does not ask you to take its word. When it uses an in-app fact, it cites the source, and the assistant message shows which categories of your data fed the answer — Profile, Applications, Open roles, Tasks, Conversation memory, Documents, and Meetings among them. If the context does not contain what you asked about, it says so plainly and asks for the missing source rather than guessing. Treat a cited answer as checkable and an uncited claim as something to verify.
Memory across the conversation
Each conversation is saved and named, so you can leave and come back to it. Within a thread, the assistant carries the history of what you have discussed, so you can refine — "now make it more direct", "you ignored the relocation factor" — instead of restarting. When your live situation and an earlier answer disagree, the current context wins: if an application has moved stage since the assistant last mentioned it, the fresh fact overrides the stale one. Start a new conversation when you switch topics, so old context does not bleed into a fresh question.
Actions it can take
Club AI is more than an answer engine. When it helps, it can act through the platform — but the experience is built around your confirmation, not its autonomy. What it can do for you as a member includes:
- Move you around the product — open the right page when that is the fastest path, after telling you why.
- Work on your search — find roles, analyze how a job fits you, research a company, draft a cover letter, or prepare interview questions and a briefing.
- Manage your tasks — create a task, create several at once, reschedule them, or suggest what to do next.
- Handle scheduling — propose meeting times, find open slots, check for conflicts, and book.
- Draft outreach — write a message or schedule a follow-up for you to review.
Actions that change things ask first
Anything that writes — applying to a role, sending a message, booking or cancelling a meeting — pauses for your explicit confirmation before it happens. You see what the assistant is about to do and approve it; it does not act behind your back. Reading and drafting are free; committing is yours.
Writing good prompts
The quality gap between a weak and strong prompt is enormous. Three habits close it:
- Give context, not keywords. "Review this offer: base X, equity Y, my current comp Z, my priority is learning rate" beats "is this offer good".
- State the output you want. "Give me three options with trade-offs" or "draft this in under 200 words" shapes the answer before it is written.
- Iterate instead of restarting. The second message — "now make it more direct" or "you ignored the relocation factor" — is where the real value appears.
Good member prompts in practice: "Deep Think: I have offers A and B [details]. I optimize for career growth over cash. Argue both sides, then recommend." Or with Search on: "What should I know about [company] before a final-round interview next week?" You can also attach a file — paste in a job description or a document and ask the assistant to work against it directly.
Limits to respect
- It advises; you decide. Club AI never makes decisions about your applications or offers — humans do, including you. Treat its output as a strong second opinion, and remember that every action it takes is one you confirmed.
- It works from what it is given. Out-of-date profile data or a vague prompt produce confident-sounding answers built on sand. If the context is missing, the assistant will say so — believe it.
- Verify what matters. For high-stakes facts — compensation norms, contract terms — confirm independently or ask your Talent Strategist. The strategist knows things no model does.

