A VC Firm Swapped Eight Excel Sheets for One Base Run by Airtable, Zapier and Claude
Draper Associates says it went from eight Excel sheets to one auto-updated base covering 400+ portfolio companies, 15,000+ contacts and 1,900+ meeting notes, searchable in plain English from a phone.
Who this is about: Early-stage venture capital firm (Draper Associates)
What it takes
- Cost
- Cost not statedThe post names Airtable, Zapier and Claude but gives no costs.
- Setup time
- A few weeksOur estimate from the source, for one business.
- Skill needed
- Some tech skillsComfortable installing an app, pasting an API key or running a command.
Explain it simply
An investment firm moved all its company, contact and meeting records into one shared database. Automations keep it updated, and an AI helps score new leads and answer questions in everyday language. When the system isn't sure about something, it stops and asks a person.
What was done
The firm decided where the data lives first: one Airtable base, one record per company, with everything else feeding it or building on it.
It built only on tools it already had (Airtable, Zapier, Claude), so fixing something means editing a prompt rather than shipping code.
Lead scoring is customised to the firm's investment thesis and past passes.
When the system isn't sure, it stops and asks a human. Every tool is documented so the next person can pick it up quickly.
How it went
The firm reports one auto-updated base with 400+ portfolio companies, 15,000+ contacts and 1,900+ meeting notes, searchable in plain English from a phone. No time or cost savings were quantified.
- The full build log is in a blog post linked from the replies, which wasn't reviewed for this draft.
- Scale figures only; the post says the payoff is hours returned to founder meetings, without numbers.
Prompt
You keep our [Airtable base] up to date.
When a new [meeting note / email / form entry] arrives, find the matching company record by name and domain. Add the note, update [fields], and score the company 1 to 10 against our thesis: [thesis notes]. Explain the score in two sentences.
If you can't match the company with certainty, or a field would be overwritten, stop and ask [person] instead of guessing.Setup notes
- Pick one base and one record per company before building any automation.
- Write your own scoring criteria instead of using an off-the-shelf scorer.
Business fit
Who this helps, and how:
Related
- Evidence grade CA Solo Lawyer Has Claude Log Each Request in Clio and Draft the First Pass via Zapier MCPVendor or secondhand, using Claude, Zapier and Clio
- Evidence grade BA Design Studio's Ops Lead Built Its Own Agency Hub on Lovable for $788.85 in CreditsStrong evidence, using Claude, Lovable and Supabase
- Evidence grade BA Dev Consultancy's Agent Cut Weekly Admin From About 9 Hours to 20 MinutesStrong evidence, using OpenClaw, Claude and Telegram
- AI use cases for business operationsDepartment hub: AI for operations, 57 examples
Similar tools
Other tools for this job, and how they differ. Chosen by our editors, never paid for.
- Make
Visual builder that's reliable for fixed steps with clear inputs and outputs, an alternative to Zapier for the feeds.
- n8n
Workflow automation with AI steps that you can self-host, for teams that want to keep the pipeline in-house.
- Grok Bot
Bots can share one source of truth in Notion or Google Drive and update it on routines, asking before anything irreversible.
Stack: Grok Bot + Notion or Google Drive: one source of truth for every Bot
“Light evidence” (grade D) means: A brief, early or in-progress claim. Treat it as an idea to test, not proof. How we grade. Spotted a mistake? Suggest a correction.