AI in the brokerage: study by the DKM's KI Navigator, with AssCompact Take part now!
Built by brokers who know every process and every problem from the daily grind of running the business.
Up and running quickly, and more precise with every case for this specific business.
No rigid off-the-shelf solution — Modus automates your individual processes and workflows.
Data isolation, audit trail and verification before every output.
One industry. One infrastructure. Individual processes. Different ways of working. No two brokers work the same way — Modus agents adapt to your individual processes.
Industry standards & infrastructure
BiPRO · GDV standard · Regulation
Individual systems
Ameise · Professional Works · rating engines
Individual processes & SOPs
Special concepts · claims settlement
Individual ways of working
Back office · field sales · management
99.5 %
Verified accuracy across production cases
4 Sec.
Average latency per work step
+1M
Individual work steps automated with Modus agents
Every message, email, WhatsApp, call, document or customer portal entry is converted on arrival into a unified, versioned work object and automatically assigned to the customer and case.
The system matches the request against the context layer through triage and automatically pulls in customer, risk and policy data. The agent works with facts, not guesswork.
In the reasoning loop, the agent plans the necessary steps, calls tools for data retrieval and action, and validates every intermediate result against the Context Layer before the next step starts.
Every output goes through a structural verification against policy and case data and is validated against a growing benchmark of verified production cases before the team member approves it.
The result is written back into the broker management system via schema mapping and, at the same time, sent out as customer communication: a double result from the same, logged case.
Communication, rating, broker management, insurers, pools: Modus connects to the systems along the entire value chain - via API, through web portals or desktop applications. No system replacement, no new login, no duplication.
Communication
Pools
CRM & administration
Insurers & rating
Portfolio & cross-sell
Claims
Modus proposes, your team member decides: review, adjust or let it run — data-driven and with full context. Each of these decisions is more than an approval: it's the judgment Modus learns from.
Vorbereitet zur Freigabe
Every case handled, every case closed, every additional use expands the Context Layer: customer, risk and case data. This single layer feeds every agent equally. The longer Modus is in use, the more each individual agent knows.
Woche 1
Monat 3
Jahr 1
More context means more intelligence. More improvements mean better judgment. The two together: significantly higher accuracy over time.
Antwortvorschlag · Adressänderung
KorrigiertTrefferquote über Zeit
Client data, contracts, and case history stay strictly isolated per broker.
Every single action is versioned and recorded in the audit log.
Results are checked against real data before they reach anyone.
Answers come from verified data, never from model memory.
Every work step is documented in a traceable way and remains transparent and auditable at all times.
„For us it was never about simply hiring more people, but about making sure every new broker doesn't automatically mean new manual work. Today we serve more partners with the same team, and the time we gain goes into growing the network."
Johann
Co-founder & Head of Data and IT
2 hrs
less admin per day
35 %
fewer internal follow-up questions
25 %
more time for advice
Modus writes directly back into your broker management system — already live for Ameise and Professional Works, with more to follow. No double entry, no Excel list on the side. Each integration maps the data structure of the respective system individually, because 'policy' or 'case' means something different in every broker management system — which is why the first connection takes longer than the tenth.
No. That's really the whole point. Modus sits on top of your existing stack, not beside it and not in place of it. Ameise, VEMA Office, Professional Works — whatever you're using stays in use.
The Inbox shows what's pending, who it's assigned to and what the suggested next step is — team-wide, not just for the individual team member. No case disappears into a mailbox anymore because that person happens to be on vacation. That's the difference from a normal email inbox: it's built for a team, not for one person.
Every action Modus proposes can be reviewed, adjusted or simply waved through before it actually happens. At the start of any workflow, a team member reviews practically everything. Over time — once enough cases have run cleanly — that shifts, and Modus increasingly handles routine cases on its own. People stay at the point in the process where they really matter, not everywhere.
An address change or policy adjustment comes in, Modus recognizes what needs to be done, gathers any missing details directly from the customer and informs the insurer — without anyone manually switching back and forth between three systems. These are the cases with the lowest advisory share and the highest volume, which is why they're best suited to full automation.
Modus identifies the risk type and category from the request, carries out the risk assessment through to a quote-ready state and pulls in comparison calculators or insurer interfaces. What today is often half a day of manual work across several tools now runs in a structured way in the background — in the end the team member sees the compared quotes, not the path to get there.
It learns, but separately per broker — a correction for one customer doesn't automatically change how Modus works for another broker. Every approval, every adjustment feeds into the business's own context. Over time, that means fewer corrections for the same kind of case.
Customer data, policies and case history are strictly isolated per broker — no mixing between Modus customers, not even anonymized for individual cases. What flows across systems is exclusively aggregated patterns, such as which details are typically relevant in a document type — never individual customer or policy data.