A retailer cleans product titles, which is simple and high volume, and also gives planogram advice, which is not simple. The team will not maintain two permanent single-model apps. What deployment fits?
Select an answer to reveal the explanation.
Short Explanation
One dispatcher can send a short errand and a hard layout question to different people, which is easier than staffing two permanent desks. You use a model router for title cleanup and planogram advice. Two apps the team will not maintain, or one large model for both, miss that split.
Full Explanation
Product-title cleanup is a narrow, high-volume job, and planogram advice is a harder judgment. A Foundry model router on one deployment can send each request to a suitable model, which fits better than two permanent single-model apps the team will not maintain, and better than pinning both jobs to one large model. Two permanent single-model apps create exactly the pair of applications the team has said it will not keep. One large model for both jobs prices every title cleanup like planogram advice and ignores the difficulty difference. One small model for both jobs saves money on titles and under-serves the planogram advice, which is the job that is not simple. Exam caveat: the router still needs a check that planogram answers stay acceptable, and a single deployment does not remove the need to review a sample of each job. Operational check: submit one title cleanup and one planogram question to the router deployment and confirm they are not forced through one permanent single-model app. Confirm the team has one router deployment to maintain rather than two permanent single-model applications left running.