An investment firm: Attio company-classification, enrichment and outbound tooling

In one line. An investment firm runs recurring office hours with 80x to build an LLM-driven company-classification system in Attio (~95% accuracy), decision-maker email enrichment, and a repeatable outbound process, with the work extending to an Affinity migration script.

Client & context

The client is an investment firm working out of Attio. Rather than a single fixed project, the relationship runs as ongoing office hours: 80x picks up whatever is in front of the team, classification, enrichment, outbound, and most recently a migration off Affinity.

The problem

The team wanted every company in Attio tagged by industry and sub-industry accurately enough to trust, decision-maker emails attached to those records, and a repeatable way to run outbound. Sending was manual through Outlook, with room to add automation and domain setup for scale. Separately, a planned Affinity migration needed to preserve AI note-taker summaries, content the v1 Affinity API does not expose.

What we built

• An LLM classification and tagging system in Attio, with sub-industry mapping and company record pages that carry a founder-contact section.
• An email-enrichment workflow using AnyMail Finder (with Clay evaluated as an alternative), targeting CEO/founder emails first and CTO/co-founders as secondary contacts.
• An outbound process designed to batch 100+ priority companies, with Apollo evaluated for sequencing.
• An Affinity migration script, plus a scoped upgrade to Affinity API v2 to extract AI note-taker text.

How we did it

Classification was tuned in the office-hours loop: renaming categories to match real company types (the B2C bucket became "Other B2C Technology") and refining the model to distinguish vertical from horizontal software, on top of infrastructure-software and B2B-fintech categories that were already landing well. On enrichment, AnyMail Finder hit 8 of 10 decision-maker emails and always returned at least one address per company, with Clay lined up as a benchmark. For the migration, we identified Affinity API v2 (which returns HTML-formatted AI notes) as the upgrade needed to carry note-taker content, and planned to validate extraction against the client's API key in Postman before setting a start date.

Outcome

• Classification system running at ~95% accuracy.
• AnyMail Finder returning decision-maker emails at an 8/10 hit rate, with at least one email per company.
• Outbound running at 20-30 messages/day during active prospecting.

Takeaway: A standing office-hours engagement that put a ~95%-accurate classification layer and decision-maker enrichment into Attio, with an Affinity migration scoped on an API-v2 upgrade.