Gave a 700-person facilities company its first CRM, AI meeting notes and dedicated AI phone agents

The problem

A facilities management company with 700 employees serving apartment communities, HOAs and commercial properties had no CRM, no automation and no AI tooling. All new business came from word of mouth, and there was no system to spot or follow up on additional services that existing customers might buy.

Nine managers visited client properties every week and captured everything on a notepad. After each meeting they typed up notes, tried to identify issues and opportunities, and eventually sent a follow-up — often hours or days later. Details were missed, follow-up was inconsistent across managers, and the owner had no view of what had been discussed or promised on site. Missed calls to the owner and two senior managers went to voicemail with no structured response.

What it cost

Upsell revenue inside existing accounts was the single largest growth lever in the business and none of it was being tracked. Manager time went into transcribing notes instead of client work. Slow, inconsistent follow-up after site meetings meant service issues and expansion requests went stale, and inbound enquiries from the website and Google Business Profile had nowhere to land.

What we built

Three connected systems on one platform. A custom CRM and sales pipeline on Next.js and PostgreSQL with separate tracks for new acquisition and existing-customer expansion, automated SMS and email sequences via n8n, Cal.com booking synced to each manager's calendar, Google Review requests triggered by Jobber job-complete events, and a live owner dashboard.

An AI meeting intelligence layer: every client meeting is transcribed by Fireflies, analysed through OpenRouter to surface discussion points and upsell opportunities, and a follow-up email is drafted straight into the manager's inbox for review. An AI website chatbot qualifies visitors by property type, square footage and services and creates the CRM record. And three dedicated Retell AI phone agents — one each for the owner and two senior team members — with individual numbers, personalised missed-call handling and separate reporting.

What changed

The company now has a repeatable account-growth engine instead of reactive selling, managers no longer type up meeting notes, and no inbound call or website enquiry goes unanswered. The platform was built and handed over for the client to run independently. Outcome figures are still to be collected from the client.