We are a fast-growing thirty-person company ready to implement our first operational AI pilot, but we want to avoid the common trap of pilot purgatory where projects stall out. How do we define the scope and select a low-risk, high-impact bottleneck to ensure our first AI project succeeds and builds momentum?

Question

Tyler Smith · Accepted Answer

To prevent pilot purgatory, a thirty-person company must avoid complex, multi-department AI initiatives. You must identify a single, narrow operational bottleneck with a highly specific use case. Sriram Subramanian recommends defining a clear purpose and scope for any initial AI application to manage expectations and ensure success.

The best place to start is with internal information retrieval. A thirty-person company often struggles with training and knowledge sharing, as critical operational knowledge is trapped in the heads of a few key employees. Building a simple, internal-only AI search agent trained exclusively on your finalized standard operating procedures is an ideal first step.

To implement this, assign a single Rock to a member of your leadership team to lead the pilot. Their goal is to build an internal knowledge bot that allows junior employees to query your operational handbook. This has a low risk of failure because the tool is entirely internal and does not touch your clients.

Define the scope strictly. The AI must only retrieve information from your approved documents, and it must cite the exact page or file name. This simple pilot proves the tangible value of AI to your team, improves time management, and builds the confidence needed for more advanced customer-facing automations in future quarters.