Advice alone will not move your company forward.
We build your scalable system.

What runs by hand, the AI takes on for around 80 percent. The remaining 20 percent stay with people: check, sign off, decide. The firm scales with the system, not with extra headcount.

Request an on-site appointment
Book a strategy call

We build tailored systems that take the manual work off your people and automate the processes. Secure, and under your control. Across more than fifty scenarios. We think with you instead of working a list: early results first, measurable progress within weeks.

This is how your system is built.

How a process becomes a system that can grow. Map, build, tighten. Not a tutorial, the work in your firm.

  1. 01

    Map

    Not the tool. The process. What lives in heads, mail and slips comes onto the table, in an order someone in the firm already recognises.

    • Where time is lost, and which AI makes sense there.
    • A report and an order of work, in two to three weeks.
    • No slide deck. The process your people already know.

  2. 02

    Build

    The first process, until someone in the firm uses it. In your stack, documented, no lock-in.

    • One area first, where the bottleneck costs the most.
    • Your people stay the ones who can do it. The routine goes.
    • Not a one-shot that withers after handover.

  3. 03

    Tighten

    Measure, sharpen, next area. What runs, stays. What snags, gets tighter.

    • The same people, less shouting across the room.
    • The next process sits on the one that already holds.
    • We stay until it holds in the day-to-day.

Who we are

We built it first, then offered it.

No whiteboard. A Mittelstand firm, five departments, dozens of automations we ran ourselves. What breaks, we know from the day-to-day, not from a seminar. Only once it held did it become the offer.

Lucas put AI into a running Mittelstand firm, across marketing, sales, logistics and purchasing. Can scaled his own brands from a standing start and laid funnels so they are not rebuilt every month.

We know the processes firms get stuck on because we ran them ourselves. Mapped, time taken out of the day-to-day, what holds built on. Not a demo pilot. Years of live operations, dozens of automations, a system that can grow.

Lucas Koch and Can Tillmann

Founders and managing directors.

Two operators. More than ten years in operations, building companies, and AI.

Lucas Koch and Can Tillmann, founders and managing directors of DropD.

  • Tax firm

    Münsterland, 11 employees

    What came of it

    The morning stack is pre-captured before anyone books in DATEV.

    Incoming receipts arrived by mail, scan and paper. Pre-capture hung on two people, the DATEV booking only after that. The receipt is read, a booking suggested and anything unclear flagged; nothing is booked until someone in the receipts circle approves.

    Volume
    around 17 receipts a day
    Time back
    4.5 hours a week
    Live after
    5 weeks

    Service

    Invoices and receipts

    What was built

    Pre-code the receipts, a human still approves.

    Time

    Pre-capture in minutes instead of the morning

    Who worked on it

    2 people in the receipts circle

  • Trades

    Westphalia, 17 employees

    What came of it

    The draft is there before the master clocks off.

    Every enquiry was picked up again in the evening. The catalogue lived in Excel, the master typed the quote himself. The enquiry now sits against their own catalogue of works, the draft comes with line items, and the master cuts and approves before anything goes out.

    Volume
    9 enquiries a week
    Time back
    6 hours a week
    Live after
    3 weeks

    Service

    Process automation

    What was built

    A quote from the enquiry and the catalogue of works.

    Time

    Draft in minutes, not in the evening

    Who worked on it

    2 masters, 1 office

  • Estate agent

    Rhineland, 8 employees

    What came of it

    What belongs to the property sits in one file, not three folders.

    Papers sat in mail, a folder and on the desk. Something was always missing before the notary appointment. Incoming pages land in one file per property, contract and annex are read, and missing pages are on the list before anyone drives to the notary.

    Volume
    around 6 property files a month
    Time back
    7 hours a week
    Live after
    4 weeks

    Service

    Documents and contracts

    What was built

    A property file from exposé, contract and annex.

    Time

    The file together in minutes, not across the afternoon

    Who worked on it

    3 agents

  • Property management

    Ruhr area, 23 employees

    What came of it

    Recurring questions no longer need a new letter.

    The same questions came every week. The answer sat in one head, in minutes from 2019, or in house rules nobody could find. House rules, minutes and stock files are searchable now; the managers get a reply draft and send it themselves.

    Volume
    21 owner questions a week
    Time back
    9 hours a week
    Live after
    6 weeks

    Service

    Knowledge management

    What was built

    Owner questions against house rules, minutes and the stock.

    Time

    A standard reply in minutes, not rewritten

    Who worked on it

    4 managers

  • HVAC firm

    East Westphalia, 14 employees

    What came of it

    Monday morning goes to the conversations, not to every file.

    Files arrived by mail and post, and every one was read, including those that did not fit. Incoming files are sorted against the job profile: a short profile for those who fit, a rejection draft for the rest. Nothing goes out until management has decided.

    Volume
    4 applications a week
    Time back
    5 hours a week
    Live after
    4 weeks

    Service

    Hiring process

    What was built

    Pre-sort applications, a human still decides.

    Time

    Pre-sorting instead of every file by hand

    Who worked on it

    Managing director plus 1 office

What runs by hand does not have to stay by hand.

Almost every process in the firm can be eased, automated and scaled.

Not with a chat seat for everyone.

Not with another licence.

With the process itself.

Work out what one process costs.

Hours, people, rate. What comes out is what you pay every year.

2 hours
4 people
22 €

4 people, 2 hours a day.

€38,720 a year, 4 people, 2 hours a day

a year, for this one process.

What can actually run in a company like yours.

No industry drawers. The things that can stand in a firm: agents, workflows, automation, phone, knowledge, training on site.

AI agents

Roles that work in your house and hand off to a person.

  • Chat
  • Phone
  • Bookkeeping

Workflows

Runs that know the order. A person decides at the points that count.

  • Quotes
  • Follow-up
  • Approvals

Automation

What is retyped today runs without a hand. Classic automation where AI would only get in the way.

  • Receipts
  • Invoices
  • CRM

Phone AI

The phone picks up. After 5pm too. A handoff when it gets serious.

  • Pickup
  • Appointments
  • Follow-up

Knowledge in the house

What only lives in heads becomes queryable. Before it leaves with retirement.

  • Runs
  • Rules
  • Onboarding

AI workshop

Training on site, in your rooms or in Ratingen. Not a remote course, no exam.

  • Foundations
  • Specialist modules
  • Workshop day

Get found

In Google, and in the answers models give. Visibility is a run, not a one-off text.

  • Google
  • ChatGPT
  • Answers

What we are often told.

  1. I don't have time.

    What comes after that costs your people time, and we quantify it beforehand: which person, which week, how many hours.

  2. We've already tried this once.

    Then you know what to ask this time: for the acceptance criterion, before you ask for the price. We write down before the build how you will recognise that it is done.

  3. I'll do this myself with ChatGPT.

    If your morning with ChatGPT produces the same result, our calculation has not added up. We build what still holds on the second run, and on the third as well.

  4. Then you'll replace our people.

    That is exactly what we do not do. We automate tasks, not people: the same people, less chasing, more time for the work only they can do.

  5. My data is not going into the cloud.

    Then it does not. A dedicated server in Germany or at your site, you decide. No US cloud. No model trained on your data.

How you measure us.

  1. 01

    The same result on every run.

    Not one good pass. The next one and the one after that have to do the same. If ChatGPT gets you the same result in a morning, you do not need us.

  2. 02

    Done is defined before anyone builds.

    Acceptance is on paper before the first line. No almost-there, no slide deck. You check yourself whether what was agreed was delivered.

  3. 03

    Your team runs it without us.

    After three months the process is yours. If we keep operating it, that is your choice, not our condition. If after handover you can only reach your system through us, you bought the wrong thing.

  4. 04

    Nothing we do not use ourselves.

    No licence we earn a cut on. No catalogue tool that does not run here. We sell the process we run ourselves, not the one that pays a commission.

The team that never clocks off.

Eight AI agents run dropd. The portraits are personas, not staff. Lena talks to you. The others take the work that would otherwise sit. Two people remain liable.

Lena

Chief eXternal Officer

24/7Live

  • Writes in this chat.
  • Books a slot when you want one.
  • Answers what the page actually says. Does not invent prices.
  • Around the clock. No pause, no holiday.

Knowledge

What we write down before anyone asks.

AI news, GDPR, practice. The radar and the guides, without going through the hub page first.

Everything under Knowledge

Questions, answered honestly.

All questions