A CRM that pulls its weight - and gives sales back 25 hours a week.
How we adapted an open-source CRM to the way we work, and how an AI agent has supported our team of five ever since: maintaining data, researching contacts, spotting potential, preparing emails. Deciding and sending stays with people.
2 min read
What happens when an AI agent in the CRM does not just collect data, but shows where the next piece of business is?
The briefing is ready before the conversation starts
Five people in sales, one agent in the CRM: it keeps the records complete, researches new contacts, reports where something is moving, and has the email draft ready. A person reviews it, adds to it and sends it.
The starting point
Every CRM lives on well-kept data - and that is exactly where most of them fail. Contacts stay incomplete, before every conversation the research starts from scratch, and anyone joining the team needs weeks to know where things stand. We did not want a licence package with a hundred features, but a CRM that fits our work and that we can develop further ourselves. So we took an open-source CRM as the basis.
What was built
We reduced the data model, pipelines and views to what we need: companies, contacts, opportunities, next steps. Then came the agent. It completes records, researches new contacts from public sources, spots movement among existing contacts - a change in management, a new branch office, a tender - and prepares the briefing and the email draft. One boundary stays: sending and deciding are done by people. Every action the agent takes is traceable in the CRM.
The agent writes the draft. Whether it goes out is decided by a person - and that person now has time for it.
What it gives sales
- Preparing for a conversation takes five minutes instead of half an hour.
- The agent shows where the potential is - the team calls the right contacts first.
- Email drafts are ready, quotes go out the same day.
- Records stay complete without anyone scheduling time to maintain them.
- New colleagues are up to speed in days, not weeks.
The calculation: 25 hours a week
Five people previously spent around six hours a week each on research, data maintenance and drafting emails. Today it is around one hour: review, add, send. Five hours per person, 25 hours a week - more than half a full-time position that is selling again instead of searching (experience values from our own operation).
At 35 euros an hour that is around 40,000 euros a year. The adaptation and the agent cost about 30,000 euros at market prices (assumption) - amortised after around nine months. The real effect is not in the calculation: more conversations with the right contacts.
How it went on
The set-up is the same one we offer clients: an existing system, an agent with clear boundaries, responsibility with people. And because we use it ourselves every day, we know where it gets stuck before a client notices.
- in sales, one agent in the CRM
- 5 people
- gained per week (experience value)
- 25 hours
- to prepare for a conversation instead of half an hour
- 5 minutes
- to amortisation (model calculation)
- 9 months
Where is money left on the table in your business?
The first conversation is free and non-binding. We listen before we propose.
- AddressHexa Logic UG (haftungsbeschränkt) Albin-Köbis-Str. 12 51147 Köln
- Email
- Phone