Why nobody quotes you on the first call
How much business automation costs is the first question owners ask, and the honest answer starts with a question back. Two quotes for the same brief can differ by a factor of four.
The difference is almost never the software. It sits in how many processes you want touched, how clean your data is, and who keeps the system running afterwards.
Scope is most of the price, so it helps to read what an automation and AI systems project covers before you compare two numbers. What follows is the arithmetic we use, written out.
Five things move the number
Give two vendors the same brief and the gap between their quotes comes from this list, in roughly this order of weight.
- How many processes. One workflow is a project. Four connected workflows across two departments is a program, and the coordination costs more than the code.
- Which tools you already run. A platform with a documented API is cheap to work with. One that fights automation adds workarounds to every single step.
- The state of your data. Duplicated contacts and free text where a field belongs push the build up before anyone writes a line of logic.
- How deep the integration goes. Reading from a system is easy. Writing back into it safely, with a rollback when something fails, is a different job.
- Who maintains it afterwards. A system your own team can change costs more to build and far less to own.
The line nobody budgets for
Data is the one that surprises people. When a vendor quotes without asking to see a real export from your CRM, the number is a guess, and the guess is always low.
We ask for twenty real records before quoting anything. Twenty rows tell you more about a build than an hour of meetings.
A price given before anyone has seen your real data is not a quote. It is an opening position.
What gets published and what does not
We do not publish one figure for a custom system, because a single number would be wrong for almost every reader. Scoping comes first, then a fixed price that does not move while we build.
Recurring services do carry public prices. The Growth Partner program starts at $2,500 a month on a six month commitment, or $3,500 a month with a three month minimum. Managed maintenance is $350 a month, month to month, and it is never a condition of the build.
Everything is quoted in dollars. Stripe converts at checkout if your card is in another currency.
Cheap builds send a second bill
The lowest quote usually wins by leaving something out. Here is what tends to be missing, in the order it hurts.
- No error handling. The workflow runs until an API returns something unexpected, then stops quietly. You hear about it from a client.
- No documentation. Six months later nobody remembers why a step exists, so nobody dares touch it.
- Built in a personal account. The freelancer owns the connections, the credentials and the run history.
- No sandbox. Changes get tested on live records, which works until the week it does not.
Rebuilding a system costs more than building it once, because you pay for the fix and for the mess it left in your data. That is the real argument against the cheapest quote.
The part of the bill your own team pays
Every automation project costs internal hours as well as money. Somebody has to explain the process, hand over credentials and sign off on the exceptions. That person is usually the one you can least spare.
Budget around one day a week from them for the length of the build. Vendors who never mention this are the ones whose projects slip, because the work stalls waiting for an answer nobody has time to give.
Naming that person before kickoff costs nothing. It also removes the most common reason a fixed price turns into a change request halfway through.
How to tell whether it pays
You can run this before you call anybody. Take one task, measure it honestly, then multiply.
Say a job takes 25 minutes and happens 30 times a week. That is 12.5 hours a week, or about 575 hours a year across 46 working weeks.
At a loaded cost of $34 an hour, one task is $19,550 a year. Automate 80 percent of it and you recover roughly $15,600 every year, without adding a person.
Now put the quote next to that figure. Divide the build cost by the annual saving and you have the payback, in years.
The rule we use
Under twelve months of payback, build it. Between twelve and twenty four, build it only when it also removes a risk you care about. Over twenty four months, park the idea and look again when volume grows.
Time is part of the price
From the first conversation to a first system running in production, four to five weeks is our normal span. Small builds ship in a week. Bigger programs run to three months, and they get cut into pieces that each go live on their own.
What should worry you in any proposal is a plan with nothing in production for six months. A long build is not a more careful build. It moves the moment of truth away from the day anyone can act on what it teaches you.
Timeline also moves the money in a direction owners like. A system live in week five starts paying back in week six, while the same budget spent on a six month plan returns nothing until next year.
Three questions that move a quote more than any discount
Vendors price uncertainty. Take the uncertainty away and the number comes down on its own.
- Which single process hurts most this quarter? A narrow brief gets quoted tighter than a wide one.
- Can you export a real sample of the data involved? Ten minutes of export saves days of assumptions.
- Who on your side owns this in month six? A named owner shortens the handover and lowers the price.
Where to start if the number still feels abstract
Most owners do not need a proposal first. They need a ranked list of what their own processes cost them, which is what an AI and operations audit puts in writing within two working days.
Prefer to do it alone? The piece on what to automate first has the same filter written out, with the arithmetic in the open.
Five things you can do tomorrow.
- Pick the task your team complains about most and time it for three days.
- Multiply minutes by passes a week, then by 46 weeks, then by your loaded hourly cost.
- Export twenty real records from the system involved and look at how clean they are.
- Ask every vendor what happens when the workflow fails at three in the morning.
- Divide each quote by the annual saving. Under twelve months of payback is easy to defend.
