AI for SMEs. From practice, not from an agency.

InputDecisionActionFeedback InputDecisionActionFeedback

For businesses in the Rhine-Neckar region that want to use AI without having their own IT department for it. Manufacturing, law firms, medical practices, trades. Cloud, agents or models on your own premises: I show you the path that fits your needs and your data.

In my main job, I co-led the AI rollout at an industrial company. What I show runs daily on my own hardware.

Request a Practice Day
Nicolas Jansohn, founder of KikuchiLabs
Nicolas Jansohn, Neckargemünd

What ends up running in your business

Nine typical cases from companies like yours. I group them by where your data goes. Each case shows how the work gets done today and what runs once it is in place.

On your own premises

The model runs on hardware inside your company. The data never leaves the building.

  • Minutes of meetings and calls

    Today
    After the meeting, someone writes the minutes from notes, often when nobody remembers what was decided.
    After
    Minutes with decisions and action items, signed off by a participant.

    I run this as MitScript

  • Email drafts from the case file

    Today
    For every inquiry, someone digs out the quote, the order and the last emails and writes the reply by hand.
    After
    The reply waits as a draft. A person checks it and sends it.
  • Company knowledge you can ask

    Today
    Anyone who needs an old quote or a contract clause searches folders or asks the one colleague who knows.
    After
    You ask in your own words and get an answer with its source.

Cloud models

Where the data allows it, I use the strongest models for writing and research.

  • Keeping an eye on the market

    Today
    Someone follows competitors, tenders and industry news on the side, whenever there is time.
    After
    A regular report with sources on what has changed since the last one.
  • Quotes and tender responses

    Today
    Every quote is pieced together from old documents, and someone ticks off the tender requirements by hand.
    After
    A first draft from past quotes that your sales team reviews.
  • Sorting incoming requests

    Today
    Someone reads the shared inbox top to bottom and forwards whatever is not theirs.
    After
    Each request reaches the right person, with a suggested reply.

Agents

They work through several steps on their own. A person makes the decision.

  • Checking quotes against your terms

    Today
    Someone checks line by line whether a quote matches the price list, discounts and payment terms.
    After
    The agent flags every deviation, and you decide what applies.

    My agents run on Offtangent

  • Following up on deadlines

    Today
    Delivery dates and promises live in emails and in people's heads, so a delay often shows up late.
    After
    The agent reminds, follows up and tells you when a promise slips.
  • Keeping two systems in sync

    Today
    Products, prices or stock in your ERP and your online shop drift apart, and someone compares lists by hand.
    After
    The agent proposes changes. A person confirms each one.

By industry

Law firms
Capturing and tracking deadlines, searching case files, preparing first drafts of briefs.
Medical practices
Doctor's letters from dictation, appointment requests from patients with a prepared reply.
Trades
Quotes from a dictated site survey, site reports spoken on the spot.
Manufacturing
Checking quotes against your terms, chasing supplier delivery dates.

We work out which case fits your company on the Practice Day.

The AI Practice Day

One day on site at your company, with the people who know your processes. Fixed price €1,490 net (excl. VAT), including a results document.

At the end of the day, you know which three processes are worth it for you and what they cost.

Schedule

  1. Morning

    What really works today

    I show live on my own hardware what AI can do today: dictation, documents and an agent that works through tasks.

  2. Midday

    Where it fits in your processes

    Discovery at the whiteboard with the people involved: which tasks cost time, who decides, which data is stored where.

  3. Afternoon

    What it costs and what GDPR means

    On-premise or cloud: rough costs for both options and what each option means for your data.

Result

A two-page document: three prioritized use cases from your company and rough costs for on-premise and cloud. With it, you decide whether to continue and with what.

Price

€1,490 net (excl. VAT)

1 day on site, fixed price, results document included

Request a Practice Day

How a project works

Four steps. Each step ends with your decision. No step commits you to the next one.

  1. Step 1: 1 day, €1,490 net

    Practice Day

    Demo, whiteboard, costs. You decide whether to continue with one of the three use cases.

  2. Step 2: 1 to 2 weeks, fixed price

    Discovery

    I map one process in detail: data, interfaces, people involved. The fixed price is set based on the Practice Day. You decide on the pilot.

  3. Step 3: 4 to 6 weeks, fixed price

    Pilot

    The solution runs in real operation, with your data and your people. Scope set after Discovery, fixed price agreed before we start. You decide whether it stays.

  4. Step 4: monthly

    Operation

    Only if the pilot works in daily use. I keep the system up to date and adapt it. I set the monthly price together with you after the pilot.

What I do not do

  • No website builder sites and no server administration.
  • No cloud wrappers without a look at the data.
  • No agency with ten services.
  • No projects without a Practice Day first.

Who is behind this

I am Nicolas Jansohn. In my main job, I am Head of Sales at an industrial company. There I co-led the rollout of AI into running processes.

What I learned: the technology was never the problem. The problem was translation. What does this mean for the processes? Who decides what? What happens to the data? AI is not “open a chatbot and type something”. AI is alignment and clarification first, then a tool.

Privately, I have built and run my own systems for years. I use them every day. KikuchiLabs brings this experience to other companies.

I run KikuchiLabs alongside my position as Head of Sales. That is deliberate: I see every day where AI rollouts get stuck in practice, and I bring that straight to you. That is why I take on no more than three projects at a time and only talk about tools I use myself.

More important than the model is the question which data may go where. I answer that before the first prompt.

Everything I offer here runs on my own systems.

What I use every day

Contact

Write to me which company and which process; I answer within 24 hours on working days.

Which business, which process? 10 to 3000 characters.

Prefer email: info@kikuchilabs.de

Nicolas Jansohn, Neckargemünd, Rhine-Neckar metropolitan region