GCP Generative AI Leader Practice Question 50: Build vs Buy Gen AI
Google Cloud Generative AI Leader · Test 5 · question 10 of 10
A mid-size insurer wants an AI assistant that lets adjusters search claims documentation and get answers with citations. The use case is common across the industry, the insurer has no in-house development team, and leadership wants something running within weeks. A competing insurer with a large engineering team wants a claims-triage experience tightly coupled to its own proprietary scoring models and internal workflow, and is willing to spend months building it.
Which approach best fits each company?
Show answer and explanation
Answer:
The build versus buy decision turns on what each organization actually needs and has available. The first insurer has a common, well-understood need (document search with citations), no development team, and a short timeline, which is exactly when a prebuilt offering such as out-of-the-box agents in the Agent Gallery delivers value fastest with the least engineering effort. The second insurer needs deep integration with its own proprietary models and workflow, plus it has the engineering capacity and the time budget, so building a custom solution on a developer platform is the better fit; a generic prebuilt tool would not reach that level of tight integration.
Option B is wrong because "always" ignores the first insurer's constraints; building would cost it time and staff it does not have for a need a prebuilt tool already covers well. Option C is wrong because a prebuilt offering cannot be assumed to match every proprietary workflow; the second insurer's requirement for deep integration with its own models is exactly the case a generic offering is not built for. Option D reverses the fit described in the scenario: it assigns the custom build to the company without a development team and the prebuilt tool to the company that explicitly wants tight, proprietary integration.
Reference: What is Gemini Enterprise?