Home / About
About
Not as a product idea on a whiteboard. A consumer lender was receiving thousands of WhatsApp messages a month, in four languages, from people who needed answers at ten at night. We built an agent to handle it, it worked, and businesses in other industries started asking for the same thing.
How it started
The first version answered the six questions that made up most of the volume. It was useful and it was limited. The second version understood what people were actually asking, routed conversations into the right department, collected documents, and knew when to stop and call a person.
Between December 2025 and August 2026 it handled more than 44,000 customer conversations and over 303,000 messages for that one client, Sheshisa Loans. It has been running every day since, including the days when it got something wrong and we fixed it. That is the difference between a product that has been tested and one that has been used.
Everything on this site is built on what we learned in those nine months, including the parts we are careful about. We do not publish a resolution rate, because the agent was never connected to that client's core system and a number we cannot defend is worth less than no number at all.
Martin Schultz
Founder, FloServ
The founder
I have spent years working in sales, marketing and lead generation, and one thing kept standing out to me: most businesses do not have a lead problem. They have a follow-up, communication and process problem.
Leads come in and nobody responds quickly enough. Old customers sit forgotten in databases. Calls get missed. Websites collect traffic without starting conversations. Teams spend hours doing work that could have happened automatically.
That is what pulled me deeper into automation and AI.
I am not interested in adding AI to a business simply because AI is the latest thing. I want to understand where opportunities are being lost, where people are wasting time and where a better system could create a measurable result.
That thinking became FloServ.
We use automation, AI and smart communication systems to help businesses respond faster, reactivate opportunities, improve customer experiences and remove repetitive work from their teams.
For me, the technology is never the end product. The outcome is.
What we believe
01
A customer on WhatsApp starts losing interest after about five minutes and assumes you are closed after an hour. Everything else we build is downstream of answering fast.
02
An agent that invents an answer is worse than no agent. Ours says it does not know and hands over, and we would rather it escalate ten times than fabricate once.
03
Eleven official languages, detected per message, on the channel this continent actually uses. Not translation bolted onto an English bot as an afterthought.
04
No discovery calls before anybody will name a price, no proposal decks, no pipeline nurture. You message us, we scope it in writing, you get a number.
What it has actually done
These figures come from Sheshisa Loans, a live deployment, not a pilot and not a demo.
Reported by Sheshisa Loans after nine months live, December 2025 to August 2026.
Being straight about it
Knowing where something stops is more useful than another page of claims about where it starts.
We do not put people on your phones. The agent handles volume and your own team handles the exceptions, which means you still need people, just fewer of them answering the same question.
We build the guardrails you specify and hold them consistently. What your sector requires of you is your call and your advisors' call, and we will say so rather than pretend otherwise.
If you get thirty enquiries a month, this will not pay for itself and we will tell you that in the first conversation rather than sell you something you do not need.
Every conversation. One system.
You will be talking to the same kind of agent we would build for you, with a person behind it. If it is not good, you will know in about four minutes.