Auto-routing 2 500+ leads per month between regional directors
Lead routing system for an education company handling 2 500+ inquiries per month. Full process audit, automatic assignment to regional directors weighted by workload and performance. Full observability and a fair split.
- ClientB2C education company (NDA)
- Project length6 weeks
- ServicesAutomations · Sales process audit · CRM integrations
- IndustryEducation
What we achieved, in numbers.
Metrics we agreed on before launch — and that the client sees in the dashboard every week.
- 2 500+Leads per month routed automatically
- < 60 sTime from lead arrival to assignment
- 100%Region coverage - zero unowned leads
- -22%Shorter sales cycle (lead → meeting)
- +34%Lift in lead → meeting conversion
- 0Internal disputes about lead ownership since go-live
Project goal
Build a fully automated lead routing system distributing inbound leads across 18 regional directors in a way that is:
- fast - a lead lands with the right person in under a minute,
- fair - workload and premium leads are spread evenly, with no subjective decisions,
- measurable - every assignment is auditable, head of sales sees real-time state,
- light for sales ops - two people stop clicking in a spreadsheet and shift focus to process quality.
The challenges we tackled
A popular B2C education company operating across several Polish regions. A team of 18 regional directors handles 2 500+ leads per month - from web forms, Meta Ads, Google Ads, partner campaigns, events and referrals.
Before our engagement, leads were routed manually by two sales ops people from a Google Sheet. Recurring issues:
- Response time averaged 2-4 hours, peaking at 8 hours - prospects went to competitors;
- Unfair split - "premium" leads (known employer, larger budget) circled between the same 3-4 reps;
- No visibility - nobody knew current load per director, closed/lost ratios;
- Internal disputes about lead ownership reached the head of sales 1-2 times per week;
- Lead leakage - 30-50 leads per month silently bounced back to the sheet and never got picked up.
Process audit
We started with a one-week audit - interviews with 6 directors across regions, head of sales, sales ops, marketing and CTO. We mapped:
- 11 lead sources (HubSpot, landing forms, Meta Lead Forms, Google Lead Forms, ManyChat, Calendly, mail, phone, events),
- 4 product segments needing different sales competencies,
- 17 geographic areas with varying inquiry volumes,
- current capacity per rep, average handling time, historic conversion per person.
The output was a routing map with business rules and metrics to be embedded in the automation - a document the head of sales could sign off in 30 minutes.
Architecture
A full pipeline on self-hosted n8n integrated with HubSpot (CRM), Slack, calendars and a Postgres database. Every lead - regardless of source - goes through the same four-step process.
- Normalization - any source maps into a unified schema (source, product, region, contact data, marketing score).
- Scoring - source campaign weight, conversion probability (ML model on historical data), business priority.
- Match - best candidate from the pool of directors based on five criteria:
- region - hard filter, lead only goes to a director serving that region,
- product competency - reps tagged per segment,
- workload - active leads in pipeline over the last 7 days,
- performance - historic conversion per rep in that segment,
- availability - vacation, working hours, calendar status.
- Assignment + notification - lead lands in HubSpot with an owner, the rep gets a Slack and email notification within 60 seconds max, with a ready brief (product, region, journey, score).
Every assignment is logged to a dedicated audit table - who, when, why and with what score got each lead.
Fair-play rules
To prevent drift - where top reps get every premium lead - we added three mechanisms:
- Caps - no one can own more than X% of weekly regional leads,
- Floors - every active rep gets a minimum weekly volume (when capacity allows),
- Random rotation for leads of equal score - removes any subjective sense of "favoritism".
All thresholds and weights are editable by the head of sales from a simple panel - no need to call us for every strategy tweak.
Observability and alerts
A Grafana dashboard backed by Postgres shows everything head of sales and sales ops need - in real time:
- leads received daily / weekly, by source,
- time from arrival to assignment (P50, P95, P99),
- live load per rep,
- funnel conversion per person, region, product,
- Slack alerts on anomalies: unassigned lead > 5 minutes, director offline > 24h, sudden conversion drop.
Results
Three months post go-live (Q1 2026 vs Q4 2025):
- 2 500+ leads per month routed automatically, zero sales ops intervention,
- Assignment time - P95 = 48 seconds (down from 2-4 hours),
- Lead leakage = 0 - every lead has an owner within a minute,
- Lead → meeting conversion +34% - likely thanks to faster first contact and better product-rep fit,
- Sales cycle shortened by 22%,
- Internal lead disputes dropped to zero - decisions are auditable, everyone sees the rationale,
- Sales ops recovered 1.5 FTE worth of work, redirected from operational clicking to process improvement.
What's next
After the first quarter we added:
- automatic re-routing for leads not responded to in 4 hours,
- weekly digest for regional directors benchmarking their results,
- AI integration that drafts the first opening email based on lead context.
Q2 plan: automated nurturing for leads not ready for immediate sales contact - so warm opportunities reach the next rep.
Let's deploy AI
with real business value.
30 minutes with a partner - no slides, with a concrete list of recommendations tailored to your organization.
or write directly: kontakt@futurefirst.pl