Feature · Clay Lead Scoring Template and Integrations

Clay lead scoring template plus the Strkr integration handoff.

Clay runs the enrichment with firmographic, technographic, intent, and behavioral signals pulled across its connector catalog. Strkr runs the scoring engine with the Strkr AI model trained on the account history. The handoff is through the Strkr open API with the enriched signals landing as scoring inputs. This page walks the template workbook, the signal set, the handoff, and the audit trail.

Start free trial See pricing
The Clay plus Strkr pattern

Enrichment on one side, scoring on the other, open API in between.

The Clay plus Strkr lead scoring pattern is a two-tool pattern that respects what each tool is good at. Clay runs the enrichment with its catalog of firmographic, technographic, intent, and behavioral signals. Strkr runs the scoring with the Strkr AI model trained on the account history and the outcome data. The two tools connect through the Strkr open API with the enriched signals landing as scoring inputs. This section walks the shape of the handoff and why the two-tool split outperforms single-tool approaches for most mid-market revenue teams.

What Clay does

Multi-source enrichment with a visual workbook.

Clay sits on the enrichment side with a visual workbook that pulls signals from across its connector catalog: firmographic data from the standard providers, technographic from BuiltWith and similar sources, intent from Bombora or 6sense, behavioral from the common event sources, and web-scraping for the long tail. The workbook composes the signals with waterfall logic, pulling from one provider and falling back to the next when the first does not have the record.

What Strkr does

Scoring engine with the Strkr AI model trained on history.

Strkr sits on the scoring side with a scoring engine that reads the enriched signals from Clay and runs them through the Strkr AI model trained on the account history. The model learns from the actual outcomes (deals that closed, deals that lost, deals that no-decided) rather than from a hand-tuned rule set, which keeps the scoring responsive to the business reality rather than frozen in a configuration from last year.

The handoff

Clay posts enriched records to the Strkr API.

Clay posts the enriched record to the Strkr open API through the standard HTTP action in the Clay workbook. The post carries the enriched fields in the Strkr field names (configured through the field mapping on the Clay side). Strkr receives the post, updates the account or lead record, and fires the scoring recalculation on the fields that changed. The handoff is a single API call per record rather than a multi-step reconciliation flow.

The score lands

The scored lead appears in the right pipeline queue.

Once the scoring recalculation completes, the lead lands in the right pipeline queue based on the score threshold: high-score leads route to the SDR team for immediate outreach, mid-score leads route to the nurture campaign, low-score leads sit on the account record with the score visible for future reference. The pattern replaces the manual triage of enriched records with a scored queue that the SDR team works from directly.

The feedback loop

Outcomes feed back to the Strkr AI model.

When the scored lead moves through the pipeline (converts to a deal, closes a deal, loses a deal, no-decides), the outcome data lands on the Strkr AI model as new training data. The model updates on a weekly cadence (configurable), which keeps the scoring responsive to the business reality rather than frozen. The feedback loop is the primary reason the two-tool split outperforms single-tool approaches over time.

The template

A Clay workbook that any team can clone and adapt.

The template is a Clay workbook that handles the standard revenue-team enrichment pattern: firmographic from the primary provider with fallback to the secondary, technographic from BuiltWith, intent from the configured provider, behavioral from the configured event source, and the Strkr API post. Teams clone the workbook, adjust the field mapping to their Strkr schema, and go live within a day. The template is maintained alongside the Strkr field schema so cloning starts from a working baseline.

What makes this different

Enrichment and scoring are different problems.

Enrichment is a data problem: get the right signals from the right sources with the right waterfall logic. Scoring is a modeling problem: learn the relationships between the signals and the outcomes. The two-tool split lets each tool handle its specialty. Single-tool approaches either compromise on enrichment breadth (which limits the signal set the model learns from) or compromise on scoring sophistication (which locks the team into rule-based scoring that cannot learn from outcomes).

The maturity model

Rule-based first, Strkr AI second, full feedback loop third.

Teams typically mature through three stages. Stage one is rule-based scoring on the Clay enriched fields (firmographic match, technographic fit, intent score threshold). Stage two is Strkr AI scoring on the same enriched fields with the model trained on the first six months of outcome data. Stage three is the full feedback loop with the model updating weekly on fresh outcome data. The maturity path is predictable and takes six to twelve months end to end for most mid-market teams.

The signal set

What Clay feeds the Strkr scoring model.

The signal set that Clay feeds the Strkr scoring model determines the scoring quality. Too few signals and the model has nothing to learn from. Too many signals and the model overfits on noise. The right signal set is a function of the business pattern: SMB plays usually need a smaller signal set weighted toward firmographic fit, enterprise plays usually need a larger signal set with heavier weighting on intent and behavioral signals.

Firmographic signals

Industry, size, revenue, geography, funding.

Firmographic signals are the baseline for every scoring model: industry (NAICS or SIC code), employee count, revenue range, geographic region, funding stage, parent-subsidiary structure. Clay pulls these from the primary provider with waterfall fallback to secondary and tertiary providers. The signal density is typically high (90-plus percent coverage) for mid-market and enterprise and lower (60-70 percent coverage) for SMB.

Technographic signals

The tech stack the prospect is running today.

Technographic signals come from BuiltWith, Datanyze, or similar providers through the Clay catalog. The signals indicate which specific tools the prospect is running: CRM, marketing automation, data warehouse, support platform, payment processor. For revenue tools, the specific CRM the prospect is running is often the strongest technographic signal because it correlates with the sales motion fit.

Intent signals

What topics the prospect is researching now.

Intent signals come from Bombora, 6sense, or similar intent providers through the Clay catalog. The signals indicate which topics the prospect is researching currently (as inferred from third-party cookie activity across the open web). Intent is a leading indicator for in-market timing rather than a strong fit signal. The scoring model usually weights intent on the recency dimension (last 14 days) rather than on the overall level.

Behavioral signals

Website visits, content downloads, email engagement.

Behavioral signals come from the standard first-party sources: website visits (through the configured analytics tool), content downloads (through the marketing automation), and email engagement (through the campaign tool). Behavioral signals are the strongest predictor of in-market timing because they reflect direct engagement with the business rather than inferred activity.

Enrichment metadata

Which provider sourced which signal, when.

Clay carries enrichment metadata on every signal: which provider sourced the field, when the field was last refreshed, and the provider's confidence score on the field. The metadata lands in Strkr alongside the signal value, which lets the scoring model account for signal freshness and provider reliability. The pattern replaces the "trust the enriched field at face value" posture with a quality-aware scoring approach.

Negative signals

Signals that score down rather than up.

Scoring is not only about positive signals. Negative signals (industry out of ICP, employee count below the floor, competitive technographic fit, no intent activity across the trailing quarter) score the record down. The Strkr AI model handles negative signals naturally because it learns from outcomes rather than from a positive-signal rule set. Rule-based scoring approaches typically miss the negative-signal pattern unless the rules are explicitly written for it.

Signal weighting

The Strkr AI model learns the weights from outcomes.

The weights across the signal set come from the Strkr AI model training rather than from a hand-tuned configuration. The model learns which signals correlate with closed-won deals, which with lost deals, and which with no-decided deals, and weights the signals accordingly. The pattern replaces the "pricing executives argue about signal weights" conversation with a model that learns the weights from the business history.

Signal explanations

The score comes with the top three signals that drove it.

For every scored record, the system surfaces the top three signals that drove the score. The SDR sees "Score 87: in ICP industry (tech), running competitor CRM, three content downloads in trailing week" rather than just the number. The explanation is critical for SDR adoption because it answers the "why should I call this one" question without a separate research step. Explanations are a property of the model output, not a feature add-on.

The template workbook

What the shipped Clay template looks like and how to adapt it.

The template Clay workbook is a starting point that most teams adapt before going live. The adaptation is typically light (adjust the field mapping to the specific Strkr schema, pick the providers from the connector catalog based on the subscription the team already has, set the waterfall priority). This section walks the shipped template columns and the common adaptations teams make in the first week.

Column one

The seed record (email, domain, or company name).

The first column is the seed record: an email address, a domain, or a company name that identifies the record Clay is enriching. Clay accepts any of the three and resolves the ambiguity through the waterfall logic. For form-fill scenarios, the seed is typically the email. For inbound domain-matching scenarios, the seed is the domain. For account-list scenarios, the seed is the company name.

Column set two

Firmographic enrichment from the configured providers.

The second column set carries the firmographic enrichment: industry, employee count, revenue range, geography, funding stage. Each column calls the configured provider with the waterfall fallback to the secondary and tertiary providers. The adaptations teams make in the first week typically involve adjusting the waterfall order based on which provider gives the best coverage for their specific target segments.

Column set three

Technographic enrichment from BuiltWith or Datanyze.

The third column set carries the technographic enrichment: current CRM, marketing automation, data warehouse, support platform. The columns call BuiltWith or Datanyze depending on the subscription the team has. The adaptations typically involve adding specific tech-stack signals that matter for the team's positioning (if the pitch is strongest against a competitor, add a column for that competitor).

Column set four

Intent enrichment from Bombora or 6sense.

The fourth column set carries the intent enrichment: trailing-14-day intent score across the relevant topic bundles. The columns call Bombora or 6sense depending on the subscription. The adaptations typically involve picking the topic bundles that match the team's positioning (CRM-related topics for a CRM team, data-warehouse topics for a data team, etc.).

Column set five

Behavioral enrichment from first-party sources.

The fifth column set carries the behavioral enrichment: website visit count, content download count, email engagement rate, trailing-30-day website activity. The columns pull from the first-party sources (Google Analytics, the marketing automation, the campaign tool). The adaptations typically involve adjusting the trailing window length and adding or removing specific first-party signals based on which the team's stack produces reliably.

Column set six

Enrichment metadata (provider, freshness, confidence).

The sixth column set carries the enrichment metadata: which provider sourced each field, when the field was last refreshed, and the provider confidence score. The metadata is standard across every enrichment column and feeds Strkr for the quality-aware scoring approach. The adaptations are minimal because the metadata shape is consistent across providers.

The output column

The HTTP post to the Strkr open API.

The output column is the HTTP post to the Strkr open API with the enriched record as the payload. The field mapping is configurable on this column: Clay field names map to Strkr field names based on the tenant schema. The post uses the standard Strkr API authentication (API key or OAuth) and returns the created or updated record ID for the Clay workbook to log for reconciliation.

The audit column

The Strkr record ID logged back for traceability.

The final column logs the Strkr record ID returned from the API post. The log provides reconciliation between the Clay workbook and the Strkr record set so a manual audit can trace a enrichment back to the specific Clay run that produced it. The pattern is important for compliance reviews and for debugging enrichment issues that surface downstream.

The scoring model and the open-API posture

How the Strkr AI scoring engine runs and how to tune it.

Once the enriched signals land in Strkr, the Strkr AI scoring engine takes over. The engine runs the model on the enriched record, produces the score and the top-three signal explanation, and fires the downstream flow based on the score threshold. This section walks the model shape, the tuning levers, and the open-API primitives that let a team integrate the scoring with external systems if they need to.

The model shape

Gradient-boosted trees trained on account history.

The Strkr AI scoring model is a gradient-boosted tree model trained on the account history. The model inputs are the enriched signals from Clay plus the Strkr-side CRM signals (deal history, support ticket volume, previous engagement). The output is a 0-to-100 score with the top-three driving signals surfaced for the SDR. The model is explainable at the record level rather than only at the aggregate level, which matters for SDR adoption.

The training cadence

Weekly retraining on fresh outcome data.

The model retrains weekly on fresh outcome data: deals that closed, deals that lost, deals that no-decided, deals that advanced to specific stages. The retraining is automatic with the admin receiving a summary of the model shift week over week. If the shift exceeds a configured threshold, the admin reviews before the new model goes live. The pattern keeps the model current without a manual retraining project every quarter.

Threshold tuning

High-score threshold, mid-score threshold, low-score threshold.

The three thresholds (high, mid, low) are tunable per segment: the SMB thresholds differ from the mid-market thresholds, and the enterprise thresholds differ again. The tuning is a visual configuration in the admin console rather than a code change. Teams typically tune the thresholds in the first quarter of operation based on the conversion rate observed in each threshold band.

Score history

The score over time per record, with the signal shifts.

Each record carries a score history with the signal shifts that drove the changes. A record that scored 65 last month and scores 85 this month shows the signals that changed (new intent spike, additional website visits, new technographic fit). The history is visible to the SDR for context and to the ops lead for model-health review.

Manual override

An SDR can tag a record for override.

For edge cases where the SDR has information the model does not (a executive conversation, a competitive event, a strategic account relationship), the SDR can tag a record for override with a required reason note. The override surfaces on the admin dashboard for review and feeds back to the model as a corrective signal. The pattern respects that the model is a tool rather than an oracle.

Score decay

Stale enrichment scores decay over time.

Scores based on enrichment data decay over time based on the enrichment freshness: a score from last week on firmographic-only signals holds longer than a score from last week on intent and behavioral signals. The decay is a property of the model and is reflected in the surface rank on the SDR queue. Stale scores drop in rank unless the enrichment refreshes, which keeps the SDR queue focused on the records with current signal strength.

The open-API posture

Every scoring surface is addressable from the open API.

The scoring engine is addressable from the Strkr open API: list scored records, read a record's score and explanation, read the score history, read the model metadata, trigger a rescoring manually. The open API lets teams build custom downstream flows (dashboards, exports, downstream integrations) without being locked into the Strkr UI for every scoring interaction.

The audit evidence

Every scoring event lands on the audit trail.

Every scoring event lands on the audit trail with the model version, the input signals, the output score, the top-three signal explanation, and the downstream routing outcome. The audit evidence is queryable in the standard report builder with saved views the compliance reviewer signs off on in advance of the audit week. SOC 2 Type II certification is in progress through 2026-Q4. HIPAA with BAA is available for healthcare-adjacent lead scoring.

Clay enrichment plus Strkr AI scoring, through the open API.

Strkr open API, inbound webhook triggers, Strkr AI scoring engine, score explanations, threshold tuning, model retraining on weekly cadence, score history, manual override, and audit evidence ship on every paid tier. The Clay side runs on the Clay subscription with the shipped template workbook as the starting point.

Start free trial See pricing
Common questions

What buyers ask about this feature.

Do we need both Clay and Strkr for lead scoring?

The two-tool split is the pattern we recommend for mid-market and enterprise revenue teams who want the strongest enrichment breadth and the strongest scoring sophistication. Clay runs the enrichment with its connector catalog across firmographic, technographic, intent, and behavioral signals. Strkr runs the scoring with the Strkr AI model trained on the account history. Teams with simpler needs can run scoring on Strkr-side signals alone without Clay, but the signal set is narrower and the scoring quality typically plateaus faster. The right fit depends on the segment: SMB plays often do well on Strkr-only scoring, enterprise plays typically benefit from the Clay enrichment.

How does the Clay workbook template work?

The template is a Clay workbook with the standard columns for seed resolution, firmographic enrichment, technographic enrichment, intent enrichment, behavioral enrichment, metadata, and the HTTP post to the Strkr open API. Teams clone the workbook, adjust the field mapping to their specific Strkr schema, pick the providers from the Clay connector catalog based on their subscription, and go live within a day. The template is maintained alongside the Strkr field schema so cloning starts from a working baseline. The workbook replaces a custom enrichment pipeline that would otherwise require engineering ownership.

How does the Strkr AI scoring model learn?

The model is a gradient-boosted tree model trained on the account history. Inputs are the enriched signals from Clay plus the Strkr-side CRM signals. The model retrains weekly on fresh outcome data (deals closed, lost, no-decided, advanced to specific stages). The output is a 0-to-100 score with the top-three driving signals surfaced for the SDR. The admin reviews the weekly retraining summary and the new model goes live automatically unless the shift exceeds a configured threshold, in which case the admin reviews before the model moves to production.

What happens if the Clay enrichment fails for a record?

The Strkr record is still created with the fields that resolved, and the scoring engine runs on the available signals. The audit trail captures which fields enriched and which failed, and the enrichment metadata column records the fallback outcome. Partial enrichment is treated as a lower-confidence input to the scoring model, which weights the score accordingly. The pattern avoids the "all-or-nothing" failure mode where a single enrichment provider failure blocks the whole record from scoring.

How does the scoring explanation help SDR adoption?

The score comes with the top three signals that drove it. The SDR sees "Score 87: in ICP industry (tech), running competitor CRM, three content downloads in trailing week" rather than just the number. The explanation answers the "why should I call this one" question without requiring a separate research step, which is critical for SDR adoption. Rule-based scoring that surfaces only the number typically produces slower SDR adoption because the SDRs do not trust the number on first exposure. The explanation closes that trust gap.

Is this HIPAA compliant for healthcare lead scoring?

Yes. Strkr is HIPAA compliant with a BAA available on request. For healthcare lead scoring where the enrichment signals might include PHI (patient-count fields, condition-adjacent intent signals), PHI fields are tagged at the schema level and the scoring model treats them with the configured PHI handling policy. The audit evidence captures the PHI flow so a HIPAA reviewer can trace which records were scored on which PHI-tagged signals. SOC 2 Type II certification is in progress through 2026-Q4. GDPR and CCPA compliance are in place.

Related features

Pair this with...

AI Lead Scoring AI Lead Scoring for Healthcare AI CRM Integration Services AI CRM Business Workflow Automation AI Sales Assistant

Try it free. Bring your team next week.

No sales call, no migration consultant, no four-month implementation. Enter your card, get 14 days of the full Pro tier, cancel any time before day 14 with zero charge. Spin up a workspace, import your CSV, and have something useful before lunch.

Start free trial See pricing