Design Philosophy
Two principles drive every product decision at Merchant Protocol. Everything else — features, workflows, UI, integrations — is downstream of these.
Principle 1: The CRM should stay current on its own
"A tradesman's job is to do the work. Not to update software."
Most CRMs fail in the trades because they require a separate data entry step that never happens consistently. A field tech finishing a job at 5pm isn't going to open a laptop and fill in a form. And if they do, it'll be incomplete, delayed, and resented.
The correct model is to capture data as a byproduct of the actions people are already taking to do their jobs — not as an additional administrative burden on top of them.
What this looks like in practice
| The action | What gets captured |
|---|---|
| Tech marks themselves en route | Job status → "In Transit", ETA recorded |
| Tech takes a job photo | Photo logged to project, timestamped, geotagged |
| Customer signs off on completion | Job marked complete, signature attached to project |
| Customer pays invoice | Project status closes, revenue recorded |
| Lead fills out website form | Contact created, project opened, intake notes attached |
| Tech sends a customer a text | Message logged to the customer's conversation thread |
| Tech clocks in/out | Time recorded against the job automatically |
| Estimate is approved | Project moves to active, schedule opens automatically |
The CRM isn't a place you go to update things. It's the record of what already happened.
Why this matters for data quality
Data entered after the fact is always incomplete. Data captured in the moment is complete by default — because the action itself provides the context. You don't need to ask a tech what time they arrived if the app tracked it when they tapped "En Route." You don't need to ask what the customer said if the text thread is already in the record.
This also means managers and office staff always have current information without having to chase anyone. Dispatch can see job status in real time. Owners can see job completion without waiting for end-of-day reports. Customers can see their own project status without calling in.
Principle 2: AI should multiply a tradesman's relationship capacity
"The best tradesmen win on relationships. AI gives one person the capacity of five."
Every trade business runs on repeat customers and referrals. The trades that win long-term aren't the cheapest — they're the ones that follow up, remember details, check in at the right time, and make customers feel like they matter.
The problem is capacity. One person can maintain maybe 50–100 active customer relationships at a high-touch level before things start falling through. The business grows, the customer list grows, and the personal touch that made the business successful starts to erode.
AI solves this by handling the relationship maintenance work that doesn't require human judgment — so the tradesman can focus on the moments that do.
What AI handles
Follow-up at the right moment After a job closes, AI sends a check-in at day 3, a review request at day 7, and a satisfaction follow-up at day 30. No one on the team needs to remember to do this. No customer falls through the cracks because someone forgot.
Drafting estimates and scopes AI reads the job notes, photos, and conversation history and drafts an estimate. The tech reviews and approves in two minutes instead of writing from scratch in thirty. Same quality. A fraction of the time.
Proactive outreach AI identifies customers who are overdue for a maintenance visit, whose warranty is expiring, or whose last service was a year ago and the season is right. It initiates the conversation — not with a generic blast, but with a message that references their specific history. The tradesman just gets notified when a customer responds.
Continuous relationship memory Every interaction — calls, texts, visits, notes, photos — is summarized and surfaced. Anyone on the team can pick up a customer conversation without asking the customer to repeat their history. AI extracts the key facts (what was done, what was said, what was promised) and keeps them accessible.
What AI doesn't replace
AI handles the logistics and the administrative follow-through. It does not replace the actual relationship moments — the call where a customer is frustrated, the conversation where a tech explains what's wrong and what it'll cost, the handshake at the end of a job. Those require a human, and the platform is designed to get humans to those moments faster by clearing the administrative overhead out of the way.
The goal: one tradesman with the relationship capacity of a five-person team. More jobs closed. Higher retention. More referrals. Without hiring three more people to manage the follow-up.
How the two principles reinforce each other
Zero-friction data capture and AI relationship capacity compound.
When the CRM is always current — because data flows in automatically from the work itself — the AI has accurate, real-time context to work with. It can send a follow-up that references the actual photos taken at the job. It can draft an estimate based on the actual notes from the site visit. It can surface the right customer at the right time because it knows exactly what was done and when.
A CRM that requires manual updates is a CRM with incomplete data. And AI working from incomplete data produces generic, low-value output. The data quality principle isn't just about reducing admin burden — it's what makes the AI effective.