Course resource
AI Tool ROI Calculator
Whether a subscription is worth it, worked out from your numbers rather than a sales page.
Step 1: what an hour of yours is worth
Pick the line that fits and fill it in.
| Situation | Calculation | Your number |
|---|---|---|
| Salaried | Annual salary ÷ 2,000 working hours | |
| Freelance | Your billable hourly rate | |
| Business owner | Monthly profit ÷ hours worked per month | |
| Student or unpaid | What you would pay someone to buy the hour back |
My hourly value: ₹________
Use a conservative number. If it flatters you, the whole calculation flatters you.
Step 2: where the time actually goes
Log one normal week. Do not estimate from memory — people are consistently wrong about this.
| Task | Times per week | Minutes each | Minutes/week | AI-assistable? |
|---|---|---|---|---|
For the last column, be strict. Mark yes only if AI does the bulk of the work and you review. Mark partial if it drafts and you substantially rewrite. Mark no if it needs judgement, relationships, or physical presence.
Step 3: realistic time saved
The honest multipliers, from people who actually measure:
| Task type | Typical saving | Notes |
|---|---|---|
| First drafts of routine writing | 50–70% | The draft is fast; editing is not |
| Summarising documents | 60–80% | Highest reliable saving there is |
| Research and synthesis | 40–60% | Verification eats into it |
| Data cleaning and reformatting | 50–70% | Very reliable when output is checked |
| Repetitive replies | 40–60% | After you build the prompt |
| Creative and strategic work | 10–30% | Mostly idea generation, not execution |
| Anything needing accuracy guarantees | 0–20% | Verification cost cancels most of the gain |
Take your minutes/week for AI-assistable tasks, apply the multiplier, and subtract 20% for the time you will spend prompting, re-prompting and verifying. That 20% is not pessimism — it is the part everyone forgets.
Minutes saved per week = (assistable minutes × multiplier) × 0.8
Hours saved per month = minutes saved per week × 4.3 ÷ 60
Value per month = hours saved per month × your hourly value
My value per month: ₹________
Step 4: the real cost
| Line | Amount |
|---|---|
| Subscription | |
| Any second tool | |
| API or usage charges | |
| Learning time, first month (hours × your rate) | |
| Ongoing prompting and verification (already in the 20% above) | 0 |
| Total first month | |
| Total steady state, monthly |
Learning time is the cost people leave out and then quietly resent. Budget four to six hours in month one.
Step 5: the verdict
Steady-state monthly value ₹________
Steady-state monthly cost ₹________
Net ₹________
- Net strongly positive → subscribe, and check again in three months.
- Net near zero → stay on the free tier. Revisit when your usage changes.
- Net negative → the problem is usually not the tool. Either you are not using it for the tasks where it wins, or those tasks are not a big enough slice of your week.
The trap
The calculation above measures time saved on work you already do. It ignores the thing that usually matters more: work you can now do at all.
If AI lets you take on a kind of project you previously turned down, or ship something weekly that you previously shipped monthly, that is worth more than the hours saved and it will not appear anywhere in the table above. Note it separately:
Work I can now do that I could not before:
Review
Redo this in three months. Two things will have changed: your prompting will have improved, and the pricing will have moved.
| Date | Monthly value | Monthly cost | Decision |
|---|---|---|---|