AI

Negotiate Your Raise With AI: Build Your Case Before the Performance Review

November 18, 2026 · Brian Arfi Faridhi

It is the end of the year, and you are still silent about your salary.

From entry-level employees to product managers who have sat in the same role for three years, most people are thinking the same thing: asking for a raise. But what they think about is only the intention. The script never gets written. So the story repeats itself every year. Wait for the review, hope the boss remembers everything you did, then feel disappointed when the increase is standard.

This is exactly where AI earns its keep. It is most useful for the part nobody wants to do: assembling evidence.

I speak from two chairs at once. Twenty years shipping products at Tokopedia, Hijra, and Flip, plus three startups of my own. So you get my version as an employee who negotiated often, and my version as a product leader who, almost every year-end, sat across the table from people asking for raises. The second version is the one you need.

What the person reviewing you actually sees

As the reviewer, let me share a secret that is not really a secret. Your boss almost never remembers the details of your work over the past year.

What sticks is a general impression. "They are decent." "They help out a lot." Impressions are hard to negotiate. Evidence is negotiable. And gathering that evidence is your job, not your boss's.

That is why people who walk into a review with a tidy list of accomplishments almost always walk out with a better outcome. It is not about being smooth with words. As the reviewer, it becomes easy for me to fight for that person upstairs. I need ammunition. You are the one who supplies it.

The problem is that building that list feels excessive. You are afraid of looking like a show-off. That is where AI helps most. It becomes the place where you dump everything without holding back, before shaping the final result into something that sounds reasonable.

How to do it, step by step

Step one: collect the raw material. Open your email, work chats, documents, tickets, whatever trace of your work exists from this year. Dump all of it into the AI, then give it context about your role, your targets, and what happened. Do not filter yet. Let the excessive stuff stay excessive. You will filter later.

Step two: have the AI turn every item into an impact statement. This is where most people fail. People write "handled the database migration." The reviewer wants to read "database migration completed with zero downtime, and support did not get flooded with complaints." Same event, very different effect at the review table. AI does this extremely well. Give it one or two example sentences you like, and it follows the pattern.

Step three: practice being the reviewer. This is my favorite trick. Ask the AI to play your skeptical boss. Have it question your list, poke holes, demand evidence. You will be surprised how many simple questions leave you stammering. Better to stammer in front of an AI than in front of the person who controls your salary budget.

Step four: build answer scripts for three scenarios: the best case, the average case, and rejection. If you get rejected, do not just accept it. Ask the AI to help you prepare concrete questions, such as what needs to be visible next year for the outcome to change, and when the discussion will happen again. People who ask those questions look serious. People who stay silent look like they have given up.

One warning from the reviewer's chair. Never ask AI to invent numbers. If you are not sure about a figure, keep it qualitative. One made-up number gets caught, and your entire list becomes garbage in the reviewer's eyes. Trust only flows one way. It does not come back.

Why this way of working matters more now

Making companies and teams save money is an old habit of mine, long before AI was trendy. The total is $4M+ per year, and it happened before the AI era. It came from various cost-saving initiatives I pushed as a product leader: process improvements, cost optimization, and more efficient product decisions. Automation was only one of them, never the single source.

That kind of leverage used to require seniority, an engineering team, and expensive systems. Now AI is the sharpest tool for the same habit, and it can be trained across everyone on a team. Including for something as small as building a proper raise argument. People who regularly use AI to do work that used to be heavy also simply look better at review time. The working culture is contagious. My own proof from the AI era: an 8-channel content distribution system I built and run alone, my Applied-AI Certification, and AI Circle.

So if you are one of those people, an entry-level employee or a three-year PM waiting for the right moment, the moment does not arrive on its own. You build it. Start small. Open your work chat, pick one accomplishment from this year, and ask AI to rewrite it as an impact statement. Five minutes. Today. Move to the next one tomorrow.

If you want to practice this way of working alongside people already doing it, drop by AI Circle at /ai-circle/. If you are from a company and want your whole team to learn this, training details are at /corporate/.