Odoo Lead Enrichment
I built a lead pipeline
inside Odoo.
A sales manager was assembling lead lists across multiple tools. I built a working prototype that discovers companies, enriches and scores them, and sends results directly into the CRM.
A sales manager was manually assembling leads for nine reps.
The target user described a workflow spanning six or more tools: company discovery, contact data, enrichment, spreadsheets, a dashboard and CRM imports. The manager was doing the coordination work for a nine-person sales team.
The brief became specific: generate a prioritized list of relevant companies and decision-makers, with results delivered where the team already works.
Project context
I created this work sample for an HG Insights application process. It drew on real user interviews and a feedback demo. It was not employment at HG Insights or a commissioned Odoo deployment.
Three interviews shaped the brief, and data tests narrowed the scope.
I interviewed three sales profiles for roughly an hour each. I chose the outbound sales manager’s workflow because the operational burden, end user and intended output were clear.
I then used AI-assisted tool exploration and a coverage test across approximately 800 companies to assess what Phoenix could support. The results changed the product scope.
Included in the prototype
- Direct-sales prospecting
- Enterprise accounts with 1,000+ employees
- US, Canada, Mexico and Brazil
- Company data, contact enrichment and scoring
Excluded after testing
- Implementation partners: software used did not establish software resold
- Smaller businesses: thin technographic coverage
- Unverified assumptions about renewal dates
The user launches a search in Odoo and receives enriched, ranked leads.
Set the search criteria in Odoo.
The user specifies geography, company size, technology criteria and a target count. A webhook starts the backend workflow.
Find and enrich companies.
Phoenix provides company and technology data. FullEnrich supplies contact enrichment when the available Phoenix access does not include contact-provider credits.
Calculate an explainable priority score.
The custom scoring model combines timing, account fit, potential Odoo need and contact reachability. Each lead includes a rationale.
Deliver results progressively into the CRM.
Supabase stores each company’s processing state. Leads are pushed to Odoo as they become ready, while asynchronous contact enrichment can finish later.
This kept the user’s launch controls and final results inside the existing CRM, reducing the need for another interface.
I changed the pipeline when larger batches stopped completing.
Process and deliver smaller units of work.
A design that waited for a whole batch ran into execution limits. The revised workflow records progress per company and pushes ready leads incrementally. The appendix reports first results in roughly one minute.
Make recovery and duplicate claims explicit.
Persisted status, atomic work claims and periodic recovery checks make interrupted work easier to resume. The design addresses concurrent processing without assuming that every external write can succeed exactly once.
Revise scoring after user feedback.
Demo feedback led to a launch button and a scoring revision: timing 45%, fit 25%, Odoo-related need 20% and reachability 10%, with penalties and decay. This is a prioritization heuristic, not a validated conversion prediction.
The demo validated the workflow; commercial impact still needs testing.
The work sample documents a functioning Odoo workflow, a user demo and follow-up changes. It also reports a large reduction in enrichment-credit usage during feasibility testing after inefficient searches were corrected.
It does not establish sustained production adoption, measured sales uplift or readiness for thousands of leads per run. Signal coverage, score quality, larger-batch reliability and cost controls remain validation work.
A production pilot would measure adoption, cost and sales usefulness.
The next evaluation would track completed runs, credits per accepted lead, user ratings and eventual conversion by score tier. That would test whether the prioritization helps the sales team, beyond demonstrating that the integration works.
Let’s discuss your team’s deployment needs.
I’m seeking an AI Deployment Strategist role in San Francisco, where I can combine strategy consulting, customer discovery and practical AI implementation.