AVA · Numbers before the conversation

Work out the payback of automation before you buy it

AVA helps you estimate the payback of automation or AI adoption on your own figures: what the manual process costs today, what cash effect automation would give and what budget would pay back. It runs in your browser, with no form and no price from us. If the effect does not even cover running costs, it says so.

For anyone deciding what to tackle first, or holding a quote and wanting to check it against arithmetic.

The calculation

What are we costing

An engineer opens a drawing or project file and hand-builds a bill of materials, a take-off or a quote from it

Not sure which one?

Your numbers
Currency
Volume is stated

With a fixed team volume, more people share the same work: process cost stays the same and team utilisation changes. If everyone performs the stated volume, choose “Per person”.

ops
people
person-min
person-min
$ / mo
% on top of salary

The example rate is only an illustration, not tax advice. Enter what the employer actually pays on top of salary. Currency does not set the country or tax regime, and changing currency keeps your rate.

Two bounds: all operations normal and all operations heavy. These are workload scenarios, not an average forecast or confidence interval. Enter the sum of active minutes of every participant per operation; exclude waiting and queue time.

h / mo
%

Account for working schedules, leave and other duties. For different salaries, use a salary average weighted by time spent on this process.

Optional: a known share produces a weighted result for normal and heavy operations. Leave blank for two extreme scenarios.

Fine-tune the assumptions · optional

This is where we do not decide for you. Change what you know better and press Calculate.

What a mistake costs

It has to be unpicked by hand and part of the work is redone

What state the data is in

The usual case: part in the ERP, part in email and spreadsheets

Where data is allowed to go

Cheaper to build and to run, but data leaves for the provider

How cautiously to count

The cautious scenario combines less automation with higher running costs; the optimistic scenario does the reverse. Your own coefficients replace scenario assumptions.

%
%

Leave blank to keep the scenario assumption. The review share is how much of the automated work’s time stays with a person: labour, not the share of results someone opens. Your values replace the matching scenario coefficients.

$ / mo
$

In the selected currency, including licences, infrastructure, support and supervision beyond the manual review already counted. Blank uses the scenario estimate; zero means no such expense.

Automation removes: 76.5 % of the operationA person still checks: 15 % of what was automated

Your own data: budget, errors, margin · optional

These lines appear in the result only once you enter your own figures. We do not invent them for you.

$
%
%
$

Provide all three values: errors before, after, and external loss per error. Exclude labour already counted in operation time. An increase in errors reduces the effect.

Where the freed hours go

Default: a vacancy left unopened, a contract cut

%

Savings appear only when a specific payment goes away: a contractor, overtime or a planned hire. 100% means the full value of freed hours becomes savings, which is the upper bound. If payroll and hiring stay the same, use 0%: hours are freed, but no cash is saved on labour.

Press it and see the result right away.

Result

Example

This is worked out on typical figures. Put in your own and press Calculate.

Bounds: all operations normal / all operations heavy. The average depends on the actual mix of operations.

Positive operating potential

At full effect, income or avoided cost covers running costs in both scenarios. Implementation payback depends on its budget, ramp-up and verified assumptions.

Net cash effect per month

$270 – $1 200

after the running cost of the system ($180 a month), at full effect

Total team operations per month: 20. Share of freed labour that becomes cash savings: 100 %. Savings appear only when a specific payment goes away: a contractor, overtime or a planned hire. 100% means the full value of freed hours becomes savings, which is the upper bound. If payroll and hiring stay the same, use 0%: hours are freed, but no cash is saved on labour.

Spent today
80 h – 240 h / mo$8 400 – $25 000 per year
Freed up
52 h – 155 h / mo$5 400 – $16 000 per year · Value of freed labour, before realisation
That is about
0.3–1 FTEThe hours return to the team. What they become is your call.

Budget that pays back

  • in 6 months$470 – $3 600
  • in 12 months$2 100 – $10 000
  • in 24 months$5 300 – $24 000

Maximum initial investment covered by accumulated benefit over the period under these assumptions. A zero bound means no positive budget is supported in that scenario. Displayed budgets are rounded down. Undiscounted; not a contractor quote.

Looks like a solution

Document AI

Turning files and documents into structured data: extraction, classification, completeness checks. A person confirms the result before anything is written.

How this is computed

Freed hours = 20 × 240 ÷ 60 × 76.5 % × (100 − 15) %

Net cash effect per month = 52 h × $8.7195121951 × 100 % + $0 − $181.5372776794

Hourly cost: salary $1,430 × (1 + employer charges 0 %) ÷ 164 available hours per month = $8.7195121951.

Running costs: $178.6784229128 + 20 × $0.1429427383 = $181.5372776794 per month.

Scenario costs are priced in roubles. Currency conversion uses the saved Bank of Russia rate dated 15.10.2026: 1 $ = 83.95 ₽, not a live quote. Your costs use the selected currency.

Year one is counted with a ramp to full effect: 40% · 75% · 100% · 100% by quarter. Running cost is paid from month one regardless.

What automation actually takes over

  1. Parsing the project file and extracting its elements
  2. Classifying elements by size group and type
  3. Counting quantities, areas and lengths
  4. Assembling the spec in the working format

What the assumption rests on

Nearest delivered project: sandwich-panel specs from CAD drawings. On control objects the app matched the manual spec line by line: 784 of 784 panels and 189 of 189 wall panels, area within 0.001 m². Each object took 10–30 seconds instead of 2–15 hours by hand. For your process the share stays an assumption until a measurement confirms it.

What this calculation does not know

We do not put our own price into it

We don’t publish a price range, so the calculator doesn’t plug in our price. Instead it works out a budget ceiling: how much the investment may cost and still pay back within your horizon. You can hold any contractor’s quote up against that figure, ours included.

The automation share is an assumption, not a measurement

Automation share, manual review and running costs are assumptions that move with the scenario, the cost of a mistake, the state of the data and the perimeter. The ramp-up is the same everywhere: 40% for the first three months, 75% for the next three, full effect after that. A linked project does not validate these coefficients for your process. Choose a scenario or enter your own values under “Fine-tune the assumptions”. An implementation decision needs measured effort and verified costs.

Freed hours are not yet saved money

Hours turn into money only when a specific payment goes away: a contractor, overtime or a planned hire. Set the share of freed labour that actually becomes savings; the initial 100% is the upper bound. If the person moves to work that earns, choose margin mode: it counts margin per freed hour and does not add salary savings, so the same hours are never counted twice.

Why you see a range, not one figure

Without a heavy-operation share, the range shows two extremes: every operation normal and every operation heavy. It is not an expected result or a confidence interval. If you know the mix, enter the share of heavy operations for a weighted estimate. Count the active labour of everyone involved in an operation, not waiting time.

It is not a financial model of the project

The calculator works out the operating effect and a simple payback period. It does not model profit tax, VAT, discounting, loans, the cost of capital, the implementation schedule, seasonality or demand growth. Currency does not set the country or tax regime: you set employer charges yourself, and exchange rates are dated.

What next

Let us measure it on your data

The result depends on your inputs and assumptions about automation, review, costs and realising the effect. Send 30 anonymised operations: we will measure labour and automation share and return the calculation, method and remaining assumptions. Free.

Current calculation:
Process: Specs and take-offs from project files
Currency: USD (1 $ = 83.95 ₽, 15.10.2026)
Operations per month: 20
Volume is stated: For the whole team
Total team operations per month: 20
People on this process: 3
Normal operation: work by all participants: 240
Heavy operation: work by all participants: 720
Available hours per person per month: 164
Gross salary per person: 1430
Employer charges: 0 %
Automation removes: 76.5 %
A person still checks: 15 %
Your fixed system cost per month: 178.67842291276799
Your system cost per operation: 0.1429427383302144
Where the freed hours go: Into cost saved
Share of freed labour that becomes cash savings: 100 %
Spent today: 80 h – 240 h per month
Freed up: 52 h – 155 h per month
Net cash effect per month: $270 – $1 200
Budget that pays back in 12 months: $2 100 – $10 000
Result: Positive operating potential
That is about: 0.3–1 FTE
Model: Bounds: all operations normal / all operations heavy. The average depends on the actual mix of operations.

or email us directly: hello@xteam.pro

Questions

Why does headcount sometimes leave the result unchanged?

With whole-team volume, people share a fixed amount of work: its labour and cost stay the same, while team utilisation changes. In per-person mode the total volume is multiplied by headcount, changing both benefits and per-operation costs. If several people contribute to one operation, add their active minutes in its labour-time field.

Why do you not show what a project costs?

Because we do not have an honest figure for this page. The same sentence describes very different amounts of work: “automate request handling” can be a ready-made scenario on clean data or a project across five systems. We give a range after a short conversation and confirm it after a diagnostic. The budget ceiling, meanwhile, you can compute here without us.

Where do the figures I type go?

Nowhere. The whole calculation runs in your browser, and no entered value is sent to the server until you press Send in the form at the bottom. We do not write them into the address bar either, which is why the URL never changes as you change the figures. With your consent to analytics, only anonymised interface events reach the counters: which process is selected and which verdict came out. Salary, volume, minutes and budget never do.

Why does the running cost not depend on salaries?

System costs combine a per-operation amount and a fixed monthly amount. Model calls scale with volume; licences, support and infrastructure are separate expenses. Initial values come from the scenario: higher in the cautious one, lower in the optimistic one. Replace them with actual costs in your selected currency.

What does “on the edge” mean?

That under the same assumptions the answer depends on the mix of operations: if every operation were normal, the cash benefit would not cover running costs, and if every one were heavy, it would. That cannot be settled from memory. Enter the share of heavy operations if you know it, or measure the process first and discuss automation second.

How is “Positive operating potential” different from payback?

The verdict answers one question: does the cash benefit cover running costs once the system reaches full effect. Payback appears only when you enter a budget or a quoted amount. The calculator then accrues the effect month by month with the ramp-up and shows the month in which the accumulated benefit covers the investment, or says it does not within 24 months. With a zero budget, first-year ROI is undefined.

What do you do with 30 operations if I send them?

We measure on them what was an assumption here: how long an operation really takes and what share can actually be removed. You get a file with the arithmetic and the method, so it can be checked without us. It is free and commits you to nothing. If it does not pay off, we write that; it happens.

Is this the same AVA that lived on its own domain?

Same problem, code rewritten from scratch. The old version demanded a name, email, phone, company and job title first, then opened a chat with a model that produced prose without a single figure. Now it is the other way round: numbers immediately and without a form, conversation afterwards and only if you want one.

Similar projects by process type

The calculator shows the case and the service for the process you pick. Here they are for every type at once.