Why you’re really here
Let’s stop pretending. You didn’t pick this up because you love AI. You picked it up because someone at your firm — your PM, the head of credit, a partner in a hallway — said “we need to be using more AI,” and you realized that vague mandate is going to land on you whether you want it or not. That mandate is also an opportunity almost nobody in your peer group will recognize. This is about that.
The honest picture at your firm
Your firm has been told by its LPs, its consultants, and every conference panel since 2024 that it needs an “AI strategy.” Most firms responded by buying Copilot licenses, sending a memo, hosting a lunch-and-learn nobody listened to, and letting everyone figure it out alone. In plain terms:
The tool is already paid for
It’s on your laptop right now.
Almost nobody uses it well
Most opened it once, asked a generic question, got a generic answer, and quit.
Senior people want a “that person”
Someone they can point to and say: “X actually figured out how to use this.”
First mover gets the visibility
The visibility that normally takes three to five years of grinding to earn.
What “visibility” actually means
In a private credit shop, the real coin is being trusted by the people above you — not just to do good work (everyone does good work), but trusted enough to be the person they pull aside when there’s a hard call. That trust converts into bigger deal allocations, more interesting sectors, faster promotion, and a bigger number on your bonus letter.
That path is normally slow because the things that earn it — sharp judgment on credits, defended views in IC, instincts about sponsors — take years to develop and are hard to show from the outside. AI competence is a shortcut. Not because being good at AI substitutes for being good at credit — it doesn’t — but because being visibly faster, more thorough, and more available than your peers is a signal that registers immediately, and the people above you have specifically been told to look for it. You’re graded on a curve, on a skill nobody at your firm has bothered to develop.
The four plays
The most common mistake when you get good at something new is telling everyone constantly. That backfires fast in finance. The move is the opposite: don’t talk about AI. Let your output and your speed do the talking, and when someone asks, tell them in two sentences without making it weird.
Be the person who always has the question pre-answered
In every IC meeting there’s a moment where the most senior person asks something nobody prepped for. Most people mumble. You’re now the person who can prep for it. The night before a deal review, spend ten minutes in your Copilot deal notebook: “What are the five questions a senior PM is most likely to ask about this deal that we haven’t already addressed?” Walk in with the answers in your head. You don’t say you used AI — you just have the answer. Three or four IC meetings of that and your reputation shifts.
Document your wins in numbers, not adjectives
Keep a running log — same page is fine — of moments AI saved you serious time or caught something you’d have missed. Not stories. Numbers and outcomes.
- “March 14: CIM read for Project Aurora in 3 hours instead of a day. Found three EBITDA adjustments the deck buried.”
- “April 2: Stress-tested the Carmine model with Excel Copilot. Caught a circular reference in working capital that would’ve flowed to the wrong leverage figure on the memo.”
- “April 8: Drafted the Riverstone quarterly update in 35 minutes. Last quarter took 4 hours.”
At bonus time you don’t argue with adjectives — you point to a list. “I cut my CIM read time 60% over six months and here are the deals where it mattered” does not get ignored.
Become quiet help to people one level above you
When a VP or principal mentions they’re swamped, or that they don’t have time to read the full data room, that’s an opening. Don’t push your AI knowledge — just say: “I can put together a synthesis of the data room in an hour if that helps.” Then do in 60 minutes what would’ve taken them six. You give them the synthesis. You don’t ask for credit. Three or four of those and you have a senior person actively rooting for your career. That’s worth more than any single deal.
Be the person junior analysts come to
New analysts get told to “use AI” by the same partners who told you. They’ll bounce around, get frustrated, and start asking for help. Be the person they ask. Spend twenty minutes walking them through your setup. Point them at this site. Be visibly generous. Senior people notice who the juniors go to — the person informally mentoring the new class is, by definition, seen as further along.
The two things not to do
No “AI Champion” in your signature. Don’t volunteer to “lead the firm’s AI initiative” until your reputation is established enough to say no later. Early titles are anchors, not boosters — they make you look like the person doing AI instead of the actual work.
It costs social capital and gains nothing. People become curious on their own timeline once they see your output improving. Until they ask, focus on your own work. The moment they ask, be helpful and specific. That’s the whole protocol.
Six months from now
If you do the drills, work the workflows, and run these plays for a few months:
- You produce memos sharper than your peers’, in half the time.
- You walk into IC with answers prepped for questions nobody else saw coming.
- A VP and a principal rely on you for synthesis work, even though it’s not formally your job.
- New associates ask you for help instead of each other.
- Your PM mentions, unprompted, that you’ve been on a tear.
- At the bonus conversation, you have a list of specific wins with specific numbers.
None of that requires being a genius. It requires the boring practice in the drills, showing up the way you already do, and letting compounded reps open a gap between you and the people still copy-pasting prompts off LinkedIn.
Custom agents — once you’ve earned them
You’ll notice there’s no “download these six AI agents” button here. That’s on purpose.
Generic agents produce generic output. An agent built to write IC memos for “a private credit fund” produces a 2021-flavored memo, not your firm’s. And handing you pre-built agents on day one short-circuits the learning — it’s the answer key before the exam. You never build the instinct, and the day the template breaks you’re stuck.
So you build the instinct first, through the drills and workflows. Once it’s there, agents stop being crutches and become force multipliers. That’s the v2 — real custom agents built for your specific firm, delivered as ongoing support: an IC Memo Drafter on your firm’s format, a Deal Screener on your actual box, a Portfolio Monitor that ingests your compliance certs, an industry-research agent for your core sector, and an LP-communications agent in your voice. Each built for you, each updated as your process and the tools evolve.
For now, the rehearsal version lives in Day 12 of the drills — a context block that turns a fresh chat into a custom specialist. When the real agents arrive, that block is your spec.
Start your wins log right now (Play 2). Open a fresh note titled “AI Wins,” and write your first line the moment AI saves you real time this week — with the number. In six months that list is your bonus conversation.