Since September, we’ve added 236 contacts and 44 companies to our CRM at Garwood Growth, with AI doing much of the work.

For a professional services business, where relationships play such an important role in winning work, that’s a significant improvement. It gives us a more complete picture of who we’re talking to, which businesses we’re building relationships with and where there may be opportunities to develop those relationships further.

It also means that valuable information is available across the team, rather than sitting in individual inboxes and meeting notes.

In the final episode of the first series of Professional Services Tech Talks, Adam Maze and George White explain how we’ve built an AI workflow that identifies contacts and companies from our emails and meetings, checks them against our existing CRM records and prepares additions for human approval.

George takes us through how the workflow operates, the challenges of making it reliable and the lessons we’ve learned from putting it into practice.

They get into:

  • How AI identifies new contacts and companies from Outlook, Teams and meeting transcripts
  • Connecting our existing systems to the CRM using MCP connectors
  • Why human approval remains an important part of the process, alongside data security and rules about which contacts to exclude
  • What happened when the workload proved too much for a single AI agent, and how George addressed this using multiple agents
  • The practical challenges of running the workflow reliably, including scheduled tasks and dependencies on individual computers
  • What other professional services businesses can learn from our experience, and where to start with similar automation

Adam and George also reflect on what we’ve learned across the six-part series. From selecting an AI platform and establishing governance, through to developing client-facing tools and automating internal processes, Tech Talks has followed our own experience of introducing AI into Garwood.

The series has been deliberately practical, focusing on things we’ve actually built and used rather than hypothetical applications of AI.

This final episode of the first series brings several of those themes together. It shows how AI can reduce routine administrative work while also helping a professional services business make better use of the information and relationships it already has.

Explore the series

This is the final episode of our first six-part Professional Services Tech Talks series. Catch up on earlier episodes covering AI platform selection, governance, building Claude skills, meeting transcription and creating an AI-powered lead magnet.

For more conversations about growing professional services businesses and building value, listen to Unlocking Value, hosted by John Howard.