You run five discovery agents. Every week they hand you five reports: support signals, NPS drivers, interview themes, journey maps, segment updates. Then what? The problem is synthesis. You sit with all five and try to find the connections by hand. Does the interview theme match the journey friction? Is the NPS driver the same as the support pain point? How do these rank against each other, and which are dependencies? That takes hours. The Opportunity Prioritization agent does it automatically every Friday at 11 AM.
Three layers of synthesis
The agent ingests all five DISCOVER outputs and applies three passes.
Cross-dataset validation: if three reports point to the same issue, support signals plus interview themes plus journey friction, it is a high-confidence opportunity. If only one report mentions it, confidence is lower. The strongest signals surface first.
Impact estimation: using segment size, willingness to pay, and engagement data, the agent estimates how many customers an opportunity affects, how much it would reduce churn, and whether it enables new revenue. Rough numbers, but better than guessing.
Dependency mapping: does fixing opportunity A unlock opportunity B? Does solving slow onboarding depend on first fixing confusing data mapping? The agent maps these so you can prioritize the sequence.
What each opportunity looks like
Every item in the output is OST-ready: a 1 to 2 sentence problem statement, the evidence (which reports and which data points validate it), the affected segments, an estimated impact, and a rough effort size. Impact is concrete, not "customers want X" but "40 customers want X, and it would reduce churn by 3 percent." The agent ranks the top 10 by (impact times segment size times willingness to fix) divided by effort, and explains the rationale for the top 3. It also flags surprising findings, patterns the individual reports missed.
What it changes
Instead of five separate reports you reconcile by hand, you get a single prioritized list with confidence scoring, impact visibility, and implementation sequencing. You catch consensus opportunities early, because multiple reports pointing at the same thing means real pain. You stop building things nobody asked for. Roadmap planning goes from a data-gathering slog to a decision meeting: you walk into Monday with opportunities already validated by multiple signals, impact estimates attached, and the sequence mapped.
The prerequisite is that all five DISCOVER agents are running, since this agent reads their weekly outputs. If any are missing, the synthesis has gaps. Get those live first, then schedule the Prioritization agent for Friday and let it produce your first stack before Monday planning.