Marketing fills the funnel; sales can't tell hot leads from cold ones. SDRs waste time on low-fit leads while gold leads sit cold for days.
The AI approach
Score each inbound lead on firmographic + behavioural signals; route top tier to AEs immediately, mid to SDRs, low to nurture.
The outcome
~+38% conversion
Conversion lift on top-tier leads ~38% (faster, fitter touch). SDR time goes to mid-tier where coaching matters.
Try itInput → Process → Output
Input — inbound lead form
contact@maxx-distribution.co.thfilled "Request demo" form · 2 min ago
companyMaxx Distribution Co.
industryFMCG · 4 plants
employees180
revenue band฿800M-1.2B
titleHead of Operations
last visitedpricing · case study × 2
enriched: Clearbit · 7d behaviour from MAP
Process — AI pipeline
1Enrich lead (firmographic)ReadSymbolic
2Score fit + intentReadPredictive
3Apply routing playbookRulesSymbolic
4Create CRM record + first taskWriteGenerative
Output — tier + routing decision
Click Run demo to enrich, score, and route the lead.
Lead score · routing decision SLA: 5 min
Tier 1 · gold
fit 0.84 · intent 0.71
87
OwnerP'Ploy (AE · West region)round-robin
First taskPersonalised demo email drafted@5 min
Follow-upCalendar invite (next 2 business days)auto-set
Drafted email opener: "Hi — saw you spent some time on the FMCG operations case study. Maxx running 4 plants is exactly the shape of customer where we typically save 18-22% on warehouse OT. Worth 20 min next week?"
Three AI types in this use case
SymbolicRouting playbook (territories, segments, queue caps); SLA timers; CRM record schema; round-robin within tier.
PredictiveLead-scoring model (XGBoost on conversion history) + intent score from web/email behaviour; uplift model.
GenerativeLLM drafts the personalised first-touch email when the lead lands in an AE queue.
The stack
MAP · HubSpot / Marketo
Enrichment · Clearbit / ZoomInfo
Model · XGBoost / LightGBM
CRM · Salesforce + LLM
When this works
≥ 6 months of conversion history to train
Enrichment data is reliable for your TAM
AE/SDR capacity matches the tier sizing
When it fails
Bias from history — model favours past customer types, blocks new segments