aaron8.
  1. Home
  2. CRM & automation

Your CRM should tell the truth.

Ten years across HubSpot, Salesforce and Zoho, and the plumbing between them: Zapier, Make, n8n, and real code when those run out. The pipeline below is the finished thing: move a deal and the follow-up happens, nobody retypes anything, and the numbers on it are derived from the cards, so they cannot lie.

Describe the mess See the finished thing See the record

Live proof

Move a card. Watch the follow-up happen.

A working sales pipeline. Add an enquiry, push it from stage to stage, and see the follow-up that fires at each step: the reply, the reminder, the handoff. Nobody retypes anything.

live demoA CRM with the automation wired in

This working build runs in your browser and needs scripts enabled. Everything on the page describes what it does; turn them on to use it.

That board and those follow-ups are what you get. Built inside HubSpot, Salesforce or Zoho, or as your own tool if none of them fit, with the automations tested against real records before anyone relies on them.

The work

Set up, cleaned up, joined up.

CRM setup and cleanup

HubSpot, Salesforce or Zoho configured around how you actually sell, or dug out from years of duplicate contacts, dead pipelines and fields nobody remembers adding. The data audit comes first; opinions come second.

Migrations

Moving CRMs without losing history is a data job wearing a software costume. The key decision is what records merge on, and getting it wrong is how organisations quietly corrupt years of history. Migrations here are reconciled against your source of truth, not declared done when the import finishes.

Automation

The intake, the follow-ups, the handoffs between website, CRM, accounting and job management, automated end to end so nobody retypes anything. Zapier, Make or n8n where they fit, custom code where they do not, and error handling either way, because an automation that fails silently is worse than none.

AI agents, wired in properly

Agents that read and write your actual systems instead of chatting beside them. Two of mine are public: one researches a company before a salesperson rings and files it in HubSpot; one answers a question a weather app cannot. Try them before you believe me.

Proof, not promises

The numbers behind the claim.

11years of systems work where wrong records have consequences
3CRMs run in anger: HubSpot, Salesforce, Zoho, no commissions taken
2rounds of client testing before any migration goes live
2AI tools live right now, source public, free to use
Doorknock, the CRM researcher Rain Check The full story

Fair questions

Before you ask.

Our CRM data is a mess. Is it fixable?
Almost always. Wrong-key merges are the classic: everyone who shared an email becomes the same person, and years of history quietly corrupt. The repair method is the same every time: trace the mechanism, rebuild the records that matter, reconcile against a source you trust. Most messes are smaller than they feel.
Zapier, Make, n8n or custom code?
The cheapest tool that will still be reliable in a year. Off-the-shelf automation is right for most jobs; custom code earns its place when volume, error handling or per-task pricing outgrows it. I build with all of them, so the recommendation is not limited by what I happen to know.
Which CRM should we pick?
The one your team will actually use, sized to how you sell, priced for what you use rather than what the sales deck showed. I have run HubSpot, Salesforce and Zoho in anger and none of them pays me a commission, so the advice is clean.
Can AI do anything useful with our CRM?
Yes, when it is wired in rather than chatting beside it: researching a company before a call and filing what it finds, drafting from your real records, keeping data tidy as it arrives. Two of my agents are live and public, so you can judge with your own eyes first.