Course resource
AI Win Log
A record of what your AI work actually produced. The single most useful career document in this course.
Why it matters
Six months from now you will be in one of these conversations:
- A performance review where you need evidence, not impressions
- A salary or rate negotiation
- An interview asking what you have actually done
- A client asking why they should pay you more than someone cheaper
In all four, the difference between "I use AI a lot" and "I cut our reporting cycle from three days to four hours — here is how" is the whole conversation.
You will not remember these. Wins feel significant for about a week and then become normal. Log them the week they happen.
The log
| Date | What I did | Before | After | Saving | Who benefited | Evidence |
|---|---|---|---|---|---|---|
Evidence is the column people skip and later need: a link, a screenshot, a chat message where someone thanked you, the sheet with the timings.
What counts as a win
Broader than you think:
- Time saved. The obvious one. Record the before, not just the after.
- Quality improved. Fewer errors, fewer revisions, fewer complaints.
- Speed. Same work, delivered sooner. Often worth more than time saved.
- Capacity. Work you could take on that you previously turned down.
- Something now possible. The most valuable and least recorded category — a thing you simply could not do before.
- Someone else enabled. You taught it, or built it, and others use it. This is leverage and it is what gets people promoted.
- A cost removed. A subscription, a contractor, an outsourced task.
Measuring honestly
Weak evidence is worse than none — it will be challenged and you will lose the argument.
- Measure the before. Once, properly, with a timer. Not from memory.
- Include your own time. Prompting, reviewing and fixing count.
- Estimate low. A conservative number you can defend beats an impressive one you cannot.
- Note what did not improve. It makes everything else credible.
- Prefer a range. "Roughly 4–6 hours a month" is more honest and more believable than "5.3 hours".
The monthly roll-up
Here is my win log for [MONTH]: [PASTE]
Produce:
1. A three-bullet summary for my manager, leading with business impact,
not tools
2. The single strongest item, written as a two-sentence case study
3. Total time saved, stated conservatively
4. Anything here that I could turn into something reusable for the team
Use only what is in the log. Do not inflate the numbers.
Turning it into a conversation
With your manager:
Over the last quarter I automated [TASKS], saving roughly [HOURS] a month.
[SPECIFIC EXAMPLE WITH A NUMBER.]
I would like to use some of that on [HIGHER-VALUE WORK].
Can we agree what that should be?
In a review or negotiation: lead with the aggregate, then one concrete example. Not a list — one example, told properly, is more persuasive than six mentioned.
With a client: the same structure, framed as what it meant for them.
The quarterly summary
| Quarter | Hours saved/month | Biggest win | Who else benefited | Where I used this |
|---|---|---|---|---|
That last column tracks whether the log is doing its job. A log you never use in a conversation is a diary.
The discipline
Log it the week it happens. Five minutes, Friday, in the same slot as your weekly review.
Log the small ones. Twenty minutes a week is seventeen hours a year. Individually forgettable, collectively the strongest number you have.
Log the failures too, in a separate column or a second sheet. "I tried automating X and it was not worth it" demonstrates judgement, and judgement is what senior roles are actually hiring for.
What not to do
- Do not inflate. One challenged number destroys the credibility of the whole log.
- Do not claim team wins as yours. Name who else was involved; it costs you nothing and protects you entirely.
- Do not include anything confidential in a log you might show externally. Keep a redacted version.
- Do not wait for a big win. The aggregate of small ones is the argument.