Measure the ROI of an employee-built tool by comparing the workflow before and after: hours saved times a loaded hourly cost, plus errors and delays avoided, minus what the tool really costs (seats, builds, the staff time spent building and training). Take a baseline before the tool goes live, measure the same way 30 and 90 days after, and count only savings the team would notice if the tool disappeared. Most good first tools pay back within the first month or two.
"Is it worth it?" is the first question every owner asks about AI, and the hardest one to answer honestly. Vendors quote big percentages. Enthusiastic employees say it changed everything. Skeptical managers say nothing changed. None of that tells you what a specific tool, built by your own people, is doing for your business.
The good news is that measuring the return on an internal tool is not complicated. You need a baseline, a consistent way to measure, an honest list of costs, and the discipline to count only what is real. This guide gives you the method, a worked example with the arithmetic shown, a scorecard you can reuse for every tool, and the mistakes that make ROI numbers meaningless.
Why internal tools are easier to measure than you think
Most internal tools replace a specific, repeated workflow: scheduling jobs, approving invoices, logging inspections, tracking requests. That makes them easier to measure than broad software like email or chat, because there is a clear "before" and "after" for one piece of work.
The time at stake is large. Zapier's survey of businesses with up to 250 people found repetitive tasks take an average of 17.3 hours a week. Asana's Anatomy of Work research reports that workers spend 60% of their time on "work about work," the chasing, copying and status-checking around the actual job. You do not need those numbers to be true for your team. You need to measure your own. But they explain why a single good tool can show a return quickly: it removes the copying and chasing from one workflow, every day.
The ROI formula, in plain words
For each tool:
Monthly value = hours saved per month x loaded hourly cost + errors avoided per month x cost per error + any other measurable gain
Monthly cost = platform cost for the tool + usage charges + staff time spent maintaining it
One-time cost = staff time spent describing, building, testing and training
Payback period = one-time cost divided by (monthly value minus monthly cost)
A "loaded hourly cost" is what an hour of someone's time really costs you: pay plus payroll taxes, benefits and overhead. Your accountant can give you a figure. If you do not have one, use a consistent estimate and write it down.
Step 1: Take a baseline before the tool goes live
This is the step people skip, and it is the one that makes everything else possible. Before the switch, capture for one normal week:
- Volume: how many items the workflow handles (jobs, requests, invoices, inspections).
- Time per item: how long each person spends on their part, estimated by the people who do it.
- Waiting time: how long items sit before someone acts. Count items older than one day each morning.
- Errors: wrong details, missed handoffs, double entry, items handled twice.
- Who touches it: every person whose time goes into the workflow.
Write it down in one place with the date. Rough is fine. Consistent is essential.
Step 2: Measure the same way after the switch
Thirty days after the switch, and again at ninety days, measure the same five things the same way. Do not change the method because the tool now offers better data. Compare like with like. You can add the tool's own timestamps as a second view.
Step 3: Count the real costs
ROI numbers are usually inflated by forgetting costs. Include all of these:
- Platform cost: seats or plan fees for the people who build.
- App user cost: some platforms charge for the people who only use the apps. According to the vendors' pages as listed in October 2026, Retool Business charges $15 per internal user per month billed annually, Power Apps says the people who run apps need a license too ($20.00 per user per month paid yearly on Premium), AppSheet requires a license for every app user including guests, and Softr caps app users per plan and bills extra app users monthly on Pro and above. On a team where twenty people use the tool and two build it, this line can be larger than the builder seats.
- Usage charges: credits, tokens or overages on top of the plan.
- Staff time to build: the hours the manager and team spent describing, testing and refining the tool.
- Staff time to train and switch: short, but real.
- Fix time: the hours spent correcting the tool in the first weeks. Stack Overflow's 2025 survey found 45% of developers say debugging AI-generated code takes more time. If your platform leaves fixing to your people, count those hours.
Step 4: Calculate, then sanity-check
Run the numbers, then ask one question: would the team notice if the tool disappeared tomorrow? If the answer is "not really," your hours-saved number is too high, whatever the arithmetic says.
Worked example
Example: a 15-person electrical contractor builds a job scheduling and completion tool to replace a whiteboard, a shared spreadsheet and texts to crews. All figures below are illustrative, to show the method. Use your own.
Baseline week (before):
- 60 jobs a week.
- The office manager spends about 6 minutes per job entering, re-entering and confirming details: 6 hours a week.
- Two crew leads spend about 3 minutes per job texting updates and answering "where are you" calls: 3 hours a week each, 6 hours total.
- 4 jobs a week go wrong because of a missed detail or a double booking, each costing about 1.5 hours to fix.
After 30 days:
- The office manager spends about 2 minutes per job: 2 hours a week, saving 4 hours.
- Crew leads update status from their phones in about 1 minute per job: 1 hour a week each, saving 4 hours total.
- 1 job a week goes wrong, saving 3 jobs x 1.5 hours = 4.5 hours.
Monthly value: 12.5 hours saved a week x 4.33 weeks = about 54 hours a month. At an assumed loaded cost of $40 an hour, that is about $2,165 a month.
Monthly cost: two builder seats on FORGE at $49 each = $98 a month. The crews and the rest of the office use the tool as app users at no charge.
One-time cost: about 12 hours of staff time to map the workflow, build, test and train, at $40 an hour = $480.
Net monthly value: $2,165 minus $98 = about $2,067.
Payback period: $480 divided by $2,067 = about one week of use.
Notice what drives the result: a workflow that runs 60 times a week, time saved by several people, and no charge for the people who only use the tool. Change any of those and the result changes, which is why measuring your own numbers matters.
The ROI scorecard
Use one row per tool and review it each quarter.
| Tool | Items per month | Hours saved per month | Errors avoided per month | Monthly value | Monthly cost | One-time cost | Payback | Daily use |
|---|---|---|---|---|---|---|---|---|
| Job scheduling (example above) | 260 | 54 | 13 | $2,165 | $98 | $480 | About 1 week | Whole office and crews |
| Your next tool |
The "daily use" column is the reality check. A tool with a strong number and weak daily use is a tool whose number will not last.
Value that does not show up in hours
Some of the most important returns are hard to put in dollars. Track them anyway, in plain words:
- Faster response to customers. Quotes out the same day, jobs confirmed before the customer calls back.
- Fewer things falling through the cracks. No more "I thought you had it."
- Visibility for managers. Status without phone calls.
- Less dependence on one person. A process that lives in a tool, not in one coordinator's head.
- Capacity. More jobs per day with the same people.
If you can count any of these, such as jobs per day or quote turnaround time, add them to the scorecard.
Comparing against the alternative
ROI is relative. The question is not only "did this tool pay back?" but "what would we have done instead?" The usual alternatives are doing nothing, buying off-the-shelf software, or hiring a developer. For scale, the U.S. Bureau of Labor Statistics reports the median annual wage for software developers was $135,980 in May 2025, before benefits and overhead. That figure is not a reason to avoid hiring when you need an engineer. It is the right benchmark for a few internal workflow tools, and a reminder to compare total cost, not just license price.
When the number comes back low
A low result is information, not a verdict. Before you write a tool off, check which part of the formula is weak.
- Low volume. The workflow runs a few times a month. The tool may be fine; the choice of workflow was not. Point the next build at something that runs daily.
- Low use. The tool works, but half the team still uses the old channel. That is an adoption problem: set a switch date and close the old way.
- High cost per user. If your platform charges for every person who opens the app, a tool used by a large crew can cost more than it saves. That is a pricing problem, and it is worth fixing at the platform level.
- High fix time. If your people spend hours each week correcting the tool, the platform is pushing quality work onto your team.
Each of these has a different fix. Find the weak part first.
Mistakes that make ROI numbers meaningless
- No baseline. Without "before," every "after" is a guess.
- Counting hours nobody can find. If saved time did not go anywhere you can name, do not count all of it.
- Ignoring the people who only use the apps. On many platforms they are the biggest cost line.
- Forgetting build and fix time. Staff hours are a real cost even when no invoice arrives.
- Measuring once. A 30-day number with no 90-day follow-up hides tools that quietly fell out of use.
- Measuring the AI instead of the work. Number of prompts or apps built is activity, not return.
ROI checklist
- Pick one tool and one workflow.
- Record a baseline week: volume, time per item, waiting time, errors, people involved.
- Agree a loaded hourly cost with your accountant and write it down.
- List every cost: seats, app users, usage, build time, training time, fix time.
- Switch over on a set date.
- Re-measure at 30 days the same way.
- Calculate monthly value, monthly cost and payback.
- Ask whether the team would notice if the tool disappeared.
- Re-measure at 90 days and record daily use.
- Add the tool to the scorecard and review it each quarter.
How FORGE handles it
FORGE is priced so the return is easy to see. It is $49 a seat a month with 3 builds included, for the people who build. The people who only use the apps are free, so a tool used by your whole crew costs the same as a tool used by two people. Changes cost a small fraction of a build and only count when something changes, packs of 10 builds are $150, and a build FORGE fails is free. Caps and alerts at 80 and 100 percent mean nothing bills by itself, so your cost line is the number you planned.
The fix-time line stays small because a panel of top AI models plans each tool, a dedicated AI crew builds it, and every piece is checked for quality and security before it is saved. Most teams have their first real tool the same day, which shortens payback from the start. A complete record of who did what helps you see how each tool is used and changed over time.
Our team onboards every company with a workflow workshop that starts with the work costing you the most time, which is where the strongest returns are. See pricing, read how it works, or talk to us.
Questions people ask
What is a good ROI for an internal tool?
A good first tool pays back its full cost, including the staff time spent building it, within a month or two, and keeps saving after that. If a tool has not clearly paid back within a quarter, look at whether people are actually using it before deciding it was the wrong tool.
How do we measure time saved without a stopwatch?
Ask the people who do the work to estimate time per item for one normal week before the switch, and again after. Back it up with timestamps where you have them, such as when a request came in and when it was done. Consistent rough numbers beat precise numbers you only take once.
Should we count time saved if nobody was laid off?
Yes, as long as you know where the time went. Hours that move to customer calls, more jobs per day, or work that used to be skipped are real value. Write down where the time went, so the number is not just a theory.
What costs do people forget to include?
Staff time spent describing, testing and learning the tool, time spent fixing it in the first weeks, usage charges on top of the plan, and the cost of the people who only use the apps if your platform charges for them.
How often should we re-measure?
At 30 days and 90 days after the switch, then once or twice a year. The 30-day number shows whether the tool works. The 90-day number shows whether it stuck.
How do we compare ROI across several tools?
Use the same scorecard for each: hours saved per month, errors avoided per month, cost per month and payback period. Rank tools by net monthly value, and look closely at any tool with low daily use.
See what FORGE builds for your team
We onboard every team ourselves, starting with the work that costs you the most time.
Talk to us
