Free decision tool · answer in a few minutes
Build vs Buy vs AI Agent Decision Tool
For technology and product leaders choosing how a piece of work should get done. Answer about one task at a time, as it runs today, and change how much each question counts if your case calls for it.
Built by XONIK. Nothing you enter leaves this page.
In brief
This tool tells you whether a task should be built into your own systems, bought as a product, given to an agent, or left alone for now. You answer questions about how the task varies, what a wrong result costs, who owns it and what it touches. It ranks the four outcomes and says what would change the result.
Step 1 of 6: Cost of a wrong result
0 of 12 answered
- Cost of a wrong result (current step)
- Shape of the work (not yet reached)
- Ownership and reach (not yet reached)
- Change and fallback (not yet reached)
- Scrutiny and rules (not yet reached)
- Running it and buying it (not yet reached)
Cost of a wrong result
Shape of the work
Ownership and reach
Change and fallback
Scrutiny and rules
Running it and buying it
Answer the 2 questions left on this step to go on.
Your result
How they rank
Want to talk through what this means for your organization?
Talk through this decision with usOn this page
The decision this tool frames, and the ones it leaves to you
The tool takes one task and asks how it should get done: written into your own systems, bought as a product, given to an agent, or left alone for now. It answers from what you tell it, and it shows its reasoning, so the answer can be argued with in a meeting.
It does not choose a vendor, a model or a budget, and it does not tell you whether a law or regulator treats your use as high-risk. Work on one task at a time. A result for one task says nothing about the one beside it.
| The result | What it commits you to | What the trade costs |
|---|---|---|
| Build into your own systems | Rules and checks you can read, test and change, with an engineer who maintains them | Speed at the start, and the vendor's upkeep |
| Buy a product | A vendor's way of working, current without your effort, plus a named person who owns the configuration | Control over how it behaves and when it changes |
| Deploy an agent | Software that picks its own steps within a written remit, with a person reviewing what matters | Predictability, since behavior shifts when the model underneath changes |
| Do not automate this yet | Settling ownership, the right result and the rules first, while people keep doing the work | The saving you hoped for, for now |
Who answers, and how long does the sitting take?
Answer for the task as it runs today, not as the sponsor wishes it ran. The questions touch operations, risk, technology and the function that does the work, so one person rarely knows every answer. A sitting with the task owner and a technology lead takes about as long as a working meeting, and the questions on rules and on who runs it afterward are where people stop to find out.
When nobody has checked, choose the answer that says so. It is the accurate one, and the tool treats it as information. Each question can be marked as mattering less or more. Use that where you have a reason, and tell the room that you did.
How do you read the ranking, the reasons and what would change it?
The result ranks every route. The gap between first and second matters more than the name at the top. A narrow lead means the task sits near a boundary. Routes level on points are all named.
Below the ranking are the answers that gave the leading route most of its points. Challenge those first, since they carry the result. Next come the single answers that, changed alone, would hand the lead to another route. If one of them is a guess, find out before you decide. Read the row for not automating yet even when it does not lead.
Taking the result into the room
- 01
Show the ranking before the reasons
The room reacts to the result first, and the reasons then get read.
- 02
Test the answers that carried the lead
The owner of the task confirms each from what happens, not from how it is described.
- 03
Settle the answers that would flip it
Assign each uncertain one to someone, with a date, and run the tool again when they report.
- 04
Write the conditions before the route
The owner, the person who runs it, the way to stop it and the rules question are written down and signed by the sponsor.
When this tool fits, and when it does not
It fits when
- You have picked one task and need to decide how it gets done
- Someone is proposing an agent or a purchase and the case is untested
Look elsewhere when
- You are comparing named vendors, which the vendor evaluation matrix does
- You have several candidate tasks to rank; use the prioritization tool first
- The question is whether the money works; use the business case calculator
The tool trades detail for speed. It shows which route your answers favor and what to check next. Whether a particular product or build will work, only a trial on your own cases can show.
What do people ask before using the build, buy or AI agent decision tool?
Can the tool tell me which vendor or model to choose?
No. It does not choose a vendor, a model or a budget, and it does not say whether a law treats your use as high-risk, which is a question for counsel. It takes one task at a time and ranks build, buy, agent or wait from your answers.
When would the tool tell us not to automate a task yet?
When something has to be settled that every route would inherit: nobody answers for the result, nobody can say what a right result looks like, a wrong result is costly and found late, or nobody has checked which rules apply. People keep doing the work while you settle it.
The tool ranks one route first. Is that the decision?
No, it is a first position to argue with. Points add up but cannot enforce conditions, so a route can lead while no named person answers for the result or nobody is set to run it. Treat those as gates before the build starts.
Can I change how much each question counts?
Yes. You can mark any question as mattering less, normal or more. Use that where you have a reason, and tell the room that you did. The result also lists single answers that, changed alone, would hand the lead to another route, so check any of those that are guesses.
Related
Want to argue the result with someone?
Describe what you are weighing in a few lines. A senior practitioner replies with a suggested first conversation.