The problem
UK hospitality venues were difficult to identify, qualify and connect with relevant EV installers. The system combined public discovery, contact enrichment, lead scoring, personalised draft generation and a partner dashboard.
What worked
The pipeline could collect venue records, enrich contact channels through multiple providers, suppress duplicates, generate contextual drafts and expose qualified records through a dashboard. Human review remained between automation and external communication.
What failed
The commercial assumptions were weaker than the software. The offer depended on partner responsiveness, timing, grant messaging and a multi-sided handoff. Building more workflow did not solve the absence of a reliably converting distribution channel.
What I would do differently
- Validate paid demand with manual delivery before automating the full chain.
- Measure one conversion bottleneck at a time.
- Retire invalid public claims immediately when an offer ends.
- Keep reusable enrichment, audit and dashboard components separate from the vertical.
Why show a failure?
Because production judgment includes knowing when to stop. The reusable result is not an EV business; it is demonstrated capability in data pipelines, workflow controls and honest system evaluation.