40/60 isn’t the result of any market research or a quote from an industry report. It’s an observation from our own product practice — from how much time and attention it actually takes to look after AI systems we’ve launched ourselves and maintain month after month across Qkwit, Brokik, and Cheaperfordrug.
Where the 40/60 split comes from
We treat this split as a starting point for a budget conversation, not a hard rule to plug into a spreadsheet. We see it consistently, from the first day a model starts reading real documents or answering real users, through the months of less visible work that follow.
Quality monitoring and data drift
A model that handles documents well on launch day doesn’t guarantee it will handle them just as well three months later. You need ongoing observation: is data extraction still accurate, is the number of cases needing manual correction going up or down. In practice that means regularly sampling outputs against what a person would have done, plus a simple alert when the correction rate starts climbing. Without that layer, a company finds out about a quality drop only when a customer reports it — the latest possible moment.
The documents a system works on don’t stand still. Invoice issuers change templates, government forms get updated, receipts from new point-of-sale systems look different from a year ago. A system that read one format perfectly can start making mistakes the moment a new variant shows up. In Qkwit, where AI reads accounting documents and a deterministic engine calculates taxes and social security contributions from that reading, this drift needs to be caught before it affects the numbers, not after.
Model updates and exception handling
The language models production AI relies on change on their own — new versions ship, and a model’s behavior on the same prompts can shift subtly. That requires regularly re-testing prompts and making sure a change under the hood doesn’t quietly break something that used to work. We wrote more about splitting responsibilities between AI and code in AI proposes, code decides — that separation also makes it safer to update the AI layer without risking the calculation layer.
Every AI system, sooner or later, runs into a case nobody planned for: a document in an unusual format, a query outside its scope, a situation calling for judgment rather than classification. Maintenance means building and maintaining an escalation path to a human for those cases, not just once at launch but as a permanent part of how the system runs — we covered this model in decision automation with human escalation.
Who you need, and what it means for the budget
Maintenance can’t be solely the vendor’s job. The company needs a process owner who regularly reviews cases escalated to a human and treats them as feedback. Without clearly assigned responsibility, quality-monitoring signals easily get lost, and a rising escalation rate only gets noticed once the damage is done.
The practical consequence: a budget that ends on launch day is a budget for 40% of the project. The other 60% — monitoring, drift, model updates, escalation paths — needs to be planned and priced from the start, not discovered as a surprise a month in. We wrote about the upfront cost itself in what AI implementation really costs in 2026 — only the two articles together give you the full budget picture.
How this looks across our own products
In Qkwit, AI reads accounting documents and a deterministic engine calculates taxes and contributions from that reading — the AI layer needs constant quality monitoring and adjustment to new document formats. In Taniej po Lek, AI recognizes basket contents from a photo or text and suggests substitutes — that feature needs similar care as packaging changes or as users phrase questions differently. In Brokik, AI adapts lease documents to a free-text instruction across several markets at once, and each market is its own surface where quality needs watching separately.
In short
A budget that ends on launch day covers only 40% of the work. The other 60% — monitoring, drift, updates, escalation — needs to be planned from the first conversation about a project. Read more about our approach on the AI page.
Frequently asked questions
Is the 40/60 split the result of market research?
No. It’s our own observation from running AI products in daily use, not a quote from a study or industry report. We treat it as a starting point for a budget conversation, not a fixed rule.
What exactly counts as AI maintenance cost?
Mainly ongoing quality monitoring, responding to drift in data and document formats, model and prompt updates, and maintaining escalation paths to a human for cases that fall outside automation.
Can you estimate maintenance cost before a project starts?
Partly. A good process audit at the start helps estimate how variable the input data will be — but the full maintenance cost picture only shows up after a few months of running the system.
Does every process need the same amount of maintenance?
No. A process built on data that rarely changes needs less attention than one where document formats or user queries change often. How variable your input data is is one of the better early indicators of future maintenance cost.
Does maintenance cost go down over time?
Usually, yes, if the process is stable — the team learns the common exceptions and tunes prompts so fewer cases need manual correction. But it never drops to zero, because the input data and the models a system runs on keep changing too.