Pipeline 05
A score is useful only when it changes the next action.
Qualification converts verified buyer and account evidence into a declared treatment. Scoring can summarize that judgment. Routing completes it by assigning one owner, one next action, a deadline, and a return path.
Its boundary: Lead Generation captures identity, permission, and action. Qualification decides what treatment the evidence supports. Routing assigns responsibility. Sales confirms whether a real buying motion exists.
Eligibility Gates
Point systems fail when enough weak positives can outweigh one decisive problem. Run identity, serviceability, permission, and record-purpose checks before any score can trigger sales work.
Gate rule: a high score cannot repair invalid identity, impossible fit, missing permission, or a misclassified record.
The Decision Matrix
Fit asks whether the account could become a successful customer. Current evidence asks whether something has changed enough to justify attention now. Combining them too early sends active poor-fit records to sales and ignores strong accounts that need more useful exposure.
Keep the experience useful. Wait for a verified change before funding direct sales work.
Inspect the evidence, account context, and buying group. Route only when the response standard is met.
Deliver what was promised, preserve consent, and avoid creating expensive work from weak evidence.
Activity can expose a partner, customer, support, or edge-case need. It does not erase poor fit.
Identity rule: preserve person and account evidence separately. Anonymous account activity never becomes named person-level intent.
Scoring Logic
A useful model keeps positive, negative, and time-sensitive evidence visible. It produces a small number of treatments that people can explain, test, and reverse. A one-point difference should never create a different commercial reality by itself.
Threshold rule: each band states the evidence required to enter, the permitted action, the service level, and the evidence required to leave.
MQL, SAL + SQL
Marketing-qualified lead (MQL), sales-accepted lead (SAL), and sales-qualified lead (SQL) can vary by company. The discipline should not. Each stage needs observable entry evidence, one owner, and an honest exit or return condition.
Opportunity rule: create an opportunity when a real buying motion justifies forecasting and continued investment, not when a score crosses a threshold.
The Routing Contract
A useful route carries the decision, evidence, next action, deadline, and rejection path. The receiving team should not need to reopen five systems or schedule a meeting to understand why the record arrived.
Calibration
A scoring system earns trust when higher-priority treatments consistently produce better business evidence and when sales can explain why records were accepted, returned, or disqualified. Review results by segment and rule version because blended averages can hide broken pockets.
History rule: preserve the score, rule version, decision, override, treatment, and outcome. Otherwise each model change rewrites the past.
The AI-Native System
AI can clean records, retrieve approved evidence, evaluate declared rules, detect changes, prepare routing packets, and monitor service levels. Human owners still define the treatments, exceptions, outreach boundaries, and consequences.
Measurement
A fast route can still be wrong. A high acceptance rate can reflect weak standards. A strong opportunity rate can hide customers that later fail. Measurement should reveal whether evidence, treatment, execution, and customer outcomes remain aligned.
Reporting rule: preserve the difference between observed, declared, enriched, inferred, scored, overridden, estimated, and confirmed evidence.
The Qualification Operating Record
One record should preserve identity, evidence, rules, treatment, route, exception, and outcome. That history lets the team explain a decision, resolve a dispute, recalibrate the model, and compare old rules with new ones.
Current Tools
The CRM should remain the operating record. Add a specialist model or routing layer only when the existing system cannot express the evidence, account relationships, treatment logic, or response obligations the team has already chosen.
Tools and links reviewed Q3 2026. Verify fit, data, privacy, AI terms, and pricing before use.
Examples Worth Studying
Both examples are vendor-published customer stories. They describe each company's reported operating method, not independently audited performance or causal proof.
Tractable described replacing spreadsheet-based qualification with CRM workflows. The team defined a lead scoring process, adjusted scores from activity, nurtured records through the buying cycle, and passed only records that met its standard to sales.
Lesson: scoring works when it governs treatment and handoff. The number should change the path, not decorate the record.
Explore the Tractable storySwagUp described moving routing logic out of custom Salesforce code. The team documented its rules, assigned teams, added timing where source fields needed to settle, and gave operators a direct way to test and revise the flow.
Lesson: routing becomes maintainable when the ownership logic is explicit, testable, and controlled by the people responsible for the process.
Explore the SwagUp storyPipeline 05
Verify the record. Run the gates. Keep fit separate from current evidence. Assign a treatment, owner, action, and deadline. Preserve the return reason and outcome so the model improves with every decision.