Three Weeks After Choosing Attio, I Started Replacing It.

In September I wrote that I'd set up a CRM before I had a pipeline, and that Attio was the right one. Both things were true. On October 15 I started building my own anyway, inside the admin side of clarentconsulting.com, and I'm running the two side by side until I cancel Attio.

The reason isn't Attio. It's that my business grew two other systems the CRM couldn't see.

What Attio got right

Everything I said in Part 5 held. Email and calendar synced in. Granola call notes landed on the right company. Claude could write to it, so I never typed into a form. And the objects-versus-lists lesson was the right lesson; the new CRM is built on it.

Picking the wrong tool isn't the failure. Never writing down why you picked it is. The Part 5 reasoning made this an easy decision to revisit.

What didn't fit

My pipeline turned out to hold four different kinds of work: consulting engagements, advisory roles, the occasional angel check, and community relationships with veteran and alumni organizations that may or may not turn into business. Attio gave me one pipeline with one set of stages. An angel check does not go through "Proposal Sent."

A deal could also have only one company. Real deals look like "startup, plus the fund co-investing, plus the person who introduced us." I was burying all of that in deal names: "Advisory," "Pro Bono," "possible angel check."

And the two systems I'd built since Part 5, the client portal and the per-client Google Drive folders, were invisible to it. A deal reached Active in Attio and nothing happened anywhere else. I was the integration, and I'm an unreliable one.

What I built instead

It all lives in the same database as the portal, so a deal and the client's portal space are one record seen from two sides. The parts worth copying:

• Four pipelines, each with its own stages. Consulting, Advisory, Angel, Community. Every stage is marked open, won or lost, so one "all deals" view works across all four.
• Any mix of people and companies on a deal, with roles. Decision maker, champion, referrer, co-investor. The company's role is separate from the person's.
• An angel check is an option on a deal, not a separate deal. One advisory relationship that might become an investment is one record.
• Drive and the portal are first-class. Internal documents link from the deal to Drive. A won consulting deal becomes a portal client in one click, invites included.
• A timeline per person and per deal. Emails, meetings, notes and texts, in one place.

The Attio data came over in one script, seven deals, twelve people and eight companies, each row keeping its Attio ID so the import can be re-run safely.

The sync problem, solved three ways

Attio's best feature was free email and calendar sync. Replacing it took three pieces, because my mail lives in three places.

• The Clarent mailbox syncs itself. An internal Google Cloud app, read-only, pulls Gmail, Calendar and Drive into the CRM every hour. Only mail with people who are already in the CRM is stored. Internal apps skip Google's verification process, which is the part that would have cost real money.
• My personal Gmail is a skill I trigger. Google won't let the internal app touch a personal account, so when a relationship starts in my personal inbox, I say "log my emails with so-and-so" and Claude reads the thread, shows me what it will log, and writes it to the deal.
• Texts are a skill too. There's no cloud API for iMessage. When a deal moves over text, Claude reads that one thread on my Mac, drafts a summary of decisions and open items, and saves it to the deal after I approve it.

Granola meetings land on the timeline through the same hourly job. And a resync button pulls over anything that still ends up in Attio while both are running.

On-demand beats always-on for anything personal. A skill I trigger can't log mail I didn't mean to share.

What I gave up

• Polish. Attio is a product built by a team. Mine is tables and screens built in evenings, following the same visual rules as the portal.
• The easy answer for clients. I still recommend Attio to most small firms. Building your own only makes sense when the CRM has to talk to systems you also built.

What I'd tell someone considering this

• Start on a real product. You learn your data model by living in someone else's.
• Move when you're the integration. Copying the same fact between tools every week is the signal.
• Keep the old IDs. An import you can re-run is an import you can trust.
• Run both for a while. Then cancel the old one on a date you set in advance.

Questions I get about this

Isn't building your own CRM the classic founder mistake? Usually. The mistake is building one to avoid a $30 subscription. Mine exists because the CRM, the portal and the client folders describe the same engagement, and keeping three copies in sync by hand was already failing.

What about the Granola notes? The hourly sync puts every meeting with a CRM contact on their timeline, with the Granola note attached.

Would you set this up for a client? Not first. I'd put them on Attio or something like it and revisit in six months, once we know where their engagements actually live.

If you've outgrown your CRM, or think you have, ask me how I'd decide. And if you made this move and regretted it, I'd really like that story.