Buying Discovery Before Building | Software Delivery as a Service

“I need online booking” contains an understandable intention and several open decisions. How long is each appointment? Who can book? How is unavailable time recorded? Can two people request the same slot? The answers change the solution and its cost.

Someone must produce this knowledge. When the work is treated only as free preparation for a future sale, it tends to be limited to what the provider can finance before a commitment exists. When it is deferred until construction, the buyer assumes obligations before learning some of their conditions.

Purchased discovery gives this work an object of its own. It can produce rules, a prototype, alternatives, and a technical finding that support a clearer decision. The purchase is justified by the decision it helps the buyer make. The proposed change is its economic reach: if AI reduces the effort of producing and evaluating this knowledge, smaller questions can become worth serving without a commitment to a much larger build.

What Makes Discovery Useful

A large collection of documents can be less useful than a short experiment. If the question is whether an existing calendar prevents concurrent bookings, repeating the feature list does not answer it. That capability must be examined and its consequences for the solution explained.

The commitment should define questions, outputs, and acceptance conditions. It should also acknowledge what will remain open. Discovery does not offer complete certainty; it produces enough information for a specific decision or explains why that decision cannot yet be made.

When each investigation costs less, it may become practical to clarify the next choice and deepen only the path selected. The buyer can compare booking alone with booking plus reminders and see the corresponding offers. Reducing complexity may reduce price; omitting information needed to assess the work may increase risk. Extensive discovery remains appropriate when dependencies or consequences justify it.

The customer must receive something they can use. Usage rights, access to artifacts, dependencies, and continuation terms matter from the first purchase. A result that makes sense only in the provider’s memory can create dependence where it should create clarity. The ability to take the artifacts to another provider depends on the agreement and should not be promised without those conditions.

Completion Can Mean Choosing Not to Build

Suppose the investigation shows that configuring an existing tool meets the barbershop’s need. Discovery has helped avoid an unnecessary implementation. Measuring its success only by conversion into development would classify that result as a failure and create a poor incentive: always discover that something must be built.

The provider must be able to receive payment for a useful answer. The buyer must be able to recognize its value even when it contradicts the initial idea. This is one reason to purchase concrete questions and examinable criteria instead of a generic promise of innovation.

Discovery may also fail to answer the question it was contracted to address. In that case, saying that all learning has value is insufficient. Exploratory work remains subject to the agreed conditions, including the uncertainties the parties accepted investigating.

Where AI Fits

Preparing alternatives, testing an interface, and examining an integration may require less effort with AI. That could make smaller discoveries viable for intentions that previously did not justify dedicated attention. But generating more options does not mean producing more understanding. Selection, evaluation, and explanation are part of the cost.

The desired result is a purchase proportionate to the question. The buyer need not finance an entire solution in advance to discover whether it deserves to exist. The provider need not hide all investigative work inside a future sale. Both reach the next conversation with something they lacked before: a better basis for accepting or declining the next commitment.

The transition from discovery to an offer is explored in From Learning to the Next Offer.