In that order. Most failed systems projects skip the middle step and try to automate the chaos. That's how you end up with a very efficient machine for losing leads.
Count everything. Trace every lead path end to end and every record from entry to report. Audit what the vendors — and the CRM — report against what actually happened. No fixes yet — just evidence, documented well enough to end arguments.
Repair the plumbing: tracking that counts, systems that talk to each other, reporting that reflects reality. Built simple enough that your team runs it without me.
The step most rollouts forget. A system nobody uses is a leak with better branding. Automate what's proven, not what's hoped.
For any business whose customers now ask ChatGPT instead of Google.
ChatGPT, Perplexity and Gemini don't read your marketing — they read your data, and they answer with whatever they can reach. Your next customer may never see your website at all. At one client, AI search could see 5 of 15+ services, and the local listings AI tools read from were missing or wrong. No one had looked until we counted.
I audit how the machines describe your business, find where the data graph is broken, and hand you a sequenced fix list a developer or VA can execute without me.
For B2B sales teams of 4–8 on Salesforce, Zoho or an industry CRM.
Most teams I meet make the same confession: no one trusts the CRM. Reps keep real notes elsewhere, managers rebuild reports in spreadsheets, and forecasting is theater. The data isn't bad because people are lazy — it's bad because the systems don't talk to each other and the process fights the tools.
I climb inside, find where data enters, leaks and dies, and rebuild the flow so the CRM becomes the thing it was supposed to be: the one place where the numbers are true.
For home inspectors, trades, specialty care, owner-operators.
A flat-fee deep dive for owner-operators who suspect something's off but can't prove it. Your vendors grade their own homework; I check it. At one client this surfaced 94 invisible leads a year, an $883/month vendor invoice running with zero key events in analytics, and a scheduler bug quoting customers $1,000 too high.
You walk away with documented evidence — confirmed multiple independent ways — and a phased roadmap. What to fix now, what to fix next, what to stop paying for.
That's normal — the symptom rarely matches the disease. Describe what's bugging you and I'll tell you where I'd start looking.