The AI Agent That Assumes: Why Automation Needs to Ask, Not Guess
AI agents are good at filling in gaps quietly, which is exactly the problem. Here's why business automation needs to be built to ask rather than assume, with real examples.
Build vs buy calculator
Custom software is powerful, but it is not always the right call. Answer a few honest questions about your situation and get a clear recommendation, including when off-the-shelf is the sensible choice.
Recommendation
Hybrid approach likely
Add an indicative custom build budget to compare against your subscription figure above. This frames the decision. It is not a quotation.
Not sure which approach is right?
This gives you a steer, not a verdict. Describe your current systems and how the work actually flows, and I'll give you an honest view, including if the answer is to keep what you have.
Talk it throughThe honest default is to buy. Off-the-shelf software is cheaper up front, maintained by someone else and quick to adopt. If a product genuinely fits how you work, custom software is hard to justify.
Custom starts to earn its place when your workflow is unusual, when you are stitching several systems together by hand, when off-the-shelf leaves a pile of manual work behind, or when the software is a real competitive edge rather than a commodity. Very often the sweet spot is ahybrid: keep the tools that work and build a focusedintegration orautomation layer around them.
Whatever the score says, the real decision comes from looking at your actual process. If you want a second opinion,tell me how things work today.
Insights
AI agents are good at filling in gaps quietly, which is exactly the problem. Here's why business automation needs to be built to ask rather than assume, with real examples.
When you inherit software with no documentation and no developer to ask, the temptation is to rewrite it fast. Here's a safer way to work out what it actually does first.
AI coding agents don't carry memory of past decisions between sessions the way a human developer does. That gap shows up as inconsistency in your codebase, and it matters more than most people realise.
Each answer feeds a weighted score out of 100 for how well-suited your situation is to custom software. Signals like a specialised workflow, many integrations, unusual rules and manual work left over push the score up. Signals like a standard workflow and good off-the-shelf options pull it down. Under 40 suggests off-the-shelf, 40 to 63 suggests a hybrid, and above that suggests custom is worth considering.
No, and that is the point. If off-the-shelf software fits your situation, the tool will say so. Custom software is a serious investment and it is not the right answer for most standard requirements.
Keep the off-the-shelf tools that already do a good job, and build a smaller custom layer around them: integrations so they share data, automation for the manual steps, or a bespoke operational front end. You get most of the benefit of custom software without replacing everything.
No. The optional cost picture uses your own subscription figure and an estimated build budget with a simple assumption for ongoing running costs. It is there to frame the decision, not to price a project. A real figure needs the workflow scoped properly.
Custom software is not a one-off cost. Budget for hosting, support, security updates and ongoing improvements. A common rough planning figure is somewhere around 15 to 20 percent of the build cost per year, though it varies with complexity.
Talk through the right approach
Tell me about your current systems and where the friction is. You'll get a straight answer on buy, build or hybrid, from the person who'd do the work.
Or email me at josh@thackr.co.uk