Explainable Prospect Qualification
A decision workflow that separates organizational fit from buyer intent and explains READY, WATCH, and MARKET EVIDENCE classifications.
The problem
Repeated activity can indicate persistent pain without proving buying intent. A prospect can fit a target profile but show no current intent; market observations may be useful without belonging in an active sales queue.
Approach
Store observable evidence, keep FIT separate from INTENT, and classify the history with a plain-language rationale. Intent advances through six evidence levels; repeated pain raises confidence in the problem but does not by itself mark a prospect READY.
Contribution
- Kept FIT as an evidence-owner input rather than inferring it from intent.
- Reserved READY for provider-seeking or buying behavior.
- Routed developing signals to WATCH and market-only histories to MARKET EVIDENCE.
- Prepared action cards only for READY records; cards require an explicit offer and proof reference and never send messages automatically.
Technical implementation
Python components qualify stored evidence histories, build a read-only daily view, and prepare human-reviewed action cards. Interaction events persist locally. Classifications include the evidence IDs and an explanation, so the reviewer can trace a decision back to observed signals.
Evidence & verification
Qualification logic and workflow checks
- Four realistic history scenarios passed: FIT and INTENT remain separate, READY requires provider-seeking or buying behavior, WATCH holds developing intent, and MARKET EVIDENCE stays outside the sales queue.
- Six local end-to-end checks passed across evidence capture, appended history, qualification, queue update, proof retrieval, action preparation, and persisted interaction state.
- The QA report for the qualification Python file passed 3/3 checks: minimum length, minimum word count, and valid Python syntax.
Results
The verified result is a locally tested, explainable classification and human-review workflow. The daily view prioritizes READY records and can filter the evidence-derived fields. No real prospect was contacted, and no client, revenue, conversion, or production outcome is claimed.
Limitations
Classification depends on the quality and completeness of human-selected source evidence and FIT input. These checks establish local behavior for the tested cases; they do not predict purchases or guarantee complete market coverage.