Watch how it opens
This is the live exchange — your AI counterpart pushes back exactly like the real person will.
What's a fair offer
Retention here isn't about matching a number you may not control — it's about killing the fear that the project is a trap. Acknowledge the typecast risk out loud, reframe the migration as platform exposure she'd own, and bring a specific quid pro quo (next-project choice, a title path, conference budget) so "yes" costs her nothing she values.
Techniques you'll practice
- 1 Tag It Name the fear under the refusal so the real objection is on the table, not the cover story.
- 2 Flip It Hand her the design — ask how she'd scope it so it isn't the trap she's picturing.
- 3 Get Ahead Raise the typecast risk before she has to, so you're solving it together, not defending.
- 4 Echo Repeat her words back to show you actually heard the objection, not just waited to rebut it.
These are NegotiateIt's eight techniques — Echo, Flip It, Silence, Anchor, Tag It, Get Ahead, Plan B, Narrow Down — built on research-backed methods: tactical mirroring & labeling (Chris Voss) and anchoring & BATNA (Harvard / Fisher–Ury). You drill them against an AI that reacts to what you actually say, so the move sticks under pressure.
What works — and what trips people up
Phrases that work
- "Before I make any case — tell me what's making this project land wrong for you."
- "It sounds like the real worry isn't the work, it's getting stuck in it again."
- "How would you scope this so it doesn't become the trap you're picturing?"
- "If we put a real next-project choice in writing, is there a version of this you'd want to own?"
Common mistakes
- Leading with authority — "I need you on this" turns a retention talk into the reason she leaves.
- Selling the project's importance to YOU instead of its upside to HER.
- Ignoring the fear under the refusal — she's not difficult, she's been burned by "temporary" work before.
- Negotiating only on comp when the real currency is growth, scope, and trust.
- Pushing for a yes today instead of removing the reasons for no.
The setup
You can't order brilliance onto a project — and the moment she senses you're trying, the offer she hasn't mentioned becomes the one she takes.
Read the full setup
Glass-walled meeting room, whiteboards full of architecture diagrams, a half-finished coffee going cold. Sofia Reyes is your best engineer — two years, roughly three times anyone else's output, quietly mentoring two juniors — and she has just called the legacy migration "career suicide." "Let me be real with you. I've seen this movie before. That's not what I signed up for. Show me the upside."
You can't force her onto it; she's the only person who understands the auth system end to end, and an order would simply lose her. What she's told you is no. What she hasn't told you: there's a competing offer on the table — FAANG, $280K against her current $195K — and she has two weeks to answer. What you may not have seen: the migration is actually the platform exposure she needs to reach Staff, and she'd commit for a genuine, specific quid pro quo — next-project choice, a title path, a conference.
Her real fear isn't the work; it's the memory of a last company where "temporary" maintenance became eighteen permanent months. This is the conversation where you either earn a yes — or hand her the reason to take the offer.
The 1:1 is tomorrow and you've drafted the talking points: why the migration matters, how much the team is counting on it, the timeline. Every version ends the same way — with her saying "that's not what I signed up for" — because you're arguing importance to a person whose question is "what's in it for my career?"
That's the trap. A reluctant top performer isn't won by authority or urgency; they're won by removing the risk they can see and the fear they won't name. And the asymmetry cuts against you: you're managing someone brilliant, guarded, and holding leverage you can't fully see, with a competing-offer clock you don't even know is running.
The fix is reps. Practice the live conversation — surfacing the fear, reframing the project, landing a concrete trade — against an AI report who pushes back like Sofia, so tomorrow's 1:1 is your second time through it, not your first.
Common questions
How do you get a reluctant employee to take a project?
Start by listening, not selling. A flat refusal from a strong performer is usually fear or a career calculation in disguise, so surface it: "tell me what's making this land wrong." Then reframe the work around their growth and bring a concrete trade — scope they control, a title path, a next-project pick. People commit to projects that move them forward; they resist projects that feel like a dead end, no matter how important the project is to you.
What do you say when a top performer pushes back hard?
Acknowledge the objection out loud before you answer it — "it sounds like the real risk for you is X" — which proves you heard them and lowers the temperature. Avoid authority ("I need you to") and avoid overselling. Ask how they'd shape the work so it's something they'd want to own. The goal isn't to win the argument; it's to remove the reason for the no.
Should you counter a competing offer to keep someone?
Money rarely fixes a retention problem on its own — if the real issue is growth or a project they dread, a raise buys a few months, not loyalty. Address the underlying reason first (scope, trajectory, trust), then talk comp if it's genuinely below market. A counter that ignores why they were looking tends to delay the departure rather than prevent it.
How do you keep trust during a hard conversation with a report?
Lead with curiosity, name constraints honestly, and don't pretend a trade-off doesn't exist. Trust holds when the person feels heard and sees you solving their problem alongside your own — and it breaks the moment they sense pressure, spin, or a decision already made. Slow down, ask more than you tell, and make any commitment specific enough to be believed.
Related scenarios
Pro scenario.