Repeatability
Medium
The structural task — synthesize inputs, estimate TAM, segment customers, write exec summary — is repeatable. But each engagement involves different industries, client contexts, and strategic stakes, requiring fresh judgment about which data to weight and how to frame conclusions.
Ambiguity Tolerance
Medium
The deliverables are reasonably well-defined (TAM, segmentation, 1-page exec summary), but 'investment-grade' and 'C-suite suitable' are subjective quality bars. The agent cannot reliably know when the output clears that bar without human validation.
Data & Tool Availability
Medium
The user has gathered the inputs, but the agent needs those files actually ingested — PDFs, spreadsheets, and presentations all in accessible formats. Gartner/Forrester reports are often paywalled or DRM-protected, and the agent cannot independently verify or supplement the data with live market sources.
Error Cost
High
A flawed TAM estimate or poorly framed exec summary going into a Fortune 500 pitch deck can damage the firm's credibility and cost a significant engagement. Errors here are not easily reversible once the deck is in front of the client.
Human Judgment Required
High
Choosing the right TAM methodology (top-down vs. bottom-up), deciding which vendor comps are actually comparable, and calibrating the strategic narrative for a specific client relationship all require judgment that AI routinely gets wrong in subtle ways. The relationship context — what this particular Fortune 500 client cares about — is invisible to the agent.