Tilo Hammer
ENDE
Book a call

Legal pages are provided in German as required by German/EU law.

For owners and CEOs of mid-sized companies

Your partner for Claude, ChatGPT & Codex

You already pay for AI. I make sure it pays off.

I work directly with owners and CEOs to turn Claude, ChatGPT and Codex into concrete, measurable process improvements — from the right priorities to production use.

Free intro call (30 min.)Download the guide
Tilo Hammer, your partner for Claude, ChatGPT & Codex

Tilo Hammer

Partner for Claude, ChatGPT & Codex

Runs his own audited AI operating system. Around 90% model cost savings in live operation.

The math

What does this actually get you?

13

full working days per month

104 hours per month · 1,248 hours per year

A model calculation based on your inputs, not a promise of results.

Identify your first processGo to the detailed potential check

Transformation

You already have AI. But no AI-powered processes yet.

Claude and ChatGPT are available across the company. Workflows still run manually, knowledge stays unstructured and measurable time savings never show up.

  1. Today: People use AI for single texts and emails.

    With AI-powered processes: AI works directly inside your business processes.

  2. Today: It feels helpful, but it saves hardly any measurable working time.

    With AI-powered processes: Time saved and process quality are measured concretely.

  3. Today: Simple prompts work. More complex tasks fall apart.

    With AI-powered processes: Guided workflows and assistants deliver the result reliably.

  4. Today: Knowledge sits scattered across emails, documents and people's heads.

    With AI-powered processes: Company knowledge becomes structured, findable and usable.

  5. Today: Plenty of ideas, no clear priority, nobody accountable.

    With AI-powered processes: The processes with the strongest business case go first.

From AI licenses to production use.

I work with you personally, from setting priorities to running it in production.

Identify your first process

Overview

Explained in 90 seconds

What your AI partner does, how we work together and why you pay for results instead of tokens.

Why your AI licences do not pay off yet

Start the sequence · 90 seconds

8 scenes, with subtitles. Starts on click only.

The sequence in eight scenes

  1. Scene 1 of 8 · The contradiction · 0:00–0:09

    Used. But nothing changed.

    Invoice
    Quote
    Support
    Reporting

    Still done by hand

    Subtitles: Your people already use AI. Your processes do not. That is exactly the gap that costs you money.

  2. Scene 2 of 8 · The cause · 0:09–0:21

    A licence is not a process

    Licence

    Open chat, ask, copy

    Process

    Workflow unchanged

    Subtitles: A licence is not a process yet. Buying a licence changes not a single step in your operation. The return only appears once a concrete workflow changes.

  3. Scene 3 of 8 · Prioritisation · 0:21–0:34

    Process check: biggest lever

    Incoming invoicesTimeErrorsCapacity
    QuotesTimeErrorsCapacity
    SupportTimeErrorsCapacity
    ReportingTimeErrorsCapacity
    Master dataTimeErrorsCapacity
    OrdersTimeErrorsCapacity

    Biggest lever

    Subtitles: That is why we do not start with the technology but with your processes. Together we look at where most time is lost and pick the first step exactly there.

  4. Scene 4 of 8 · Invoice example · 0:34–0:51

    Result first, then build

    1. Email
    2. Data read
    3. Checked
    4. Cost centre
    5. Exception?
    6. Posted

    Standard case: automatic

    Exception: your team decides

    Subtitles: Result first. Then we build. An example: an invoice arrives by email. The data is read and checked, then assigned to a cost centre. Only in an exception does the system ask a member of your team. Otherwise the invoice is posted automatically.

  5. Scene 5 of 8 · Integration · 0:51–1:03

    Your systems. Your people.

    Your process
    Email
    ERP
    Files
    Approval
    Reporting

    Your team decides

    Subtitles: This runs inside your existing systems, not next to them. At every important point a person still decides, not the machine. Responsibility stays with your team, the technology only takes over the repetition.

  6. Scene 6 of 8 · Accountability and measurement · 1:03–1:15

    Results count, not usage

    Token usage
    Usage minutes
    Process result

    ~90 % lower model cost in our own operation

    Subtitles: What is measured and paid for is the agreed process result, not token usage. You can see at any time what was actually completed.

  7. Scene 7 of 8 · Proof · 1:15–1:23

    108 % ROI · 5.8 months

    108 %

    ROI

    5.8 months

    Payback

    Reference mandate 2026 · calculation basis on this site

    Subtitles: In the reference mandate the ROI was 108 percent and the payback period 5.8 months.

  8. Scene 8 of 8 · Process check · 1:23–1:30

    Process check · 30 min.

    One process. One result. Then the next.

    Review your first process

    30 minutes · personally with Tilo Hammer

    Subtitles: In 30 minutes we identify the biggest realistic lever and check whether an implementation pays off for your company.

Read the transcript

The full narration of all eight scenes.

  1. Scene 1 of 8 · The contradiction (0:00–0:09): Your people already use AI. Your processes do not. That is exactly the gap that costs you money.
  2. Scene 2 of 8 · The cause (0:09–0:21): A licence is not a process yet. Buying a licence changes not a single step in your operation. The return only appears once a concrete workflow changes.
  3. Scene 3 of 8 · Prioritisation (0:21–0:34): That is why we do not start with the technology but with your processes. Together we look at where most time is lost and pick the first step exactly there.
  4. Scene 4 of 8 · Invoice example (0:34–0:51): Result first. Then we build. An example: an invoice arrives by email. The data is read and checked, then assigned to a cost centre. Only in an exception does the system ask a member of your team. Otherwise the invoice is posted automatically.
  5. Scene 5 of 8 · Integration (0:51–1:03): This runs inside your existing systems, not next to them. At every important point a person still decides, not the machine. Responsibility stays with your team, the technology only takes over the repetition.
  6. Scene 6 of 8 · Accountability and measurement (1:03–1:15): What is measured and paid for is the agreed process result, not token usage. You can see at any time what was actually completed.
  7. Scene 7 of 8 · Proof (1:15–1:23): In the reference mandate the ROI was 108 percent and the payback period 5.8 months.
  8. Scene 8 of 8 · Process check (1:23–1:30): In 30 minutes we identify the biggest realistic lever and check whether an implementation pays off for your company.

Animated version. The personal video follows.

Where you stand

Does any of this sound familiar?

Most mid-sized companies have bought AI but changed no process. That is where the return fails, not on the technology.

  • You pay for licenses, but nobody can put a number on the return.

    Source: Only 45% of CFOs can quantify AI ROI. BCG, 09/2025

  • AI runs as a chatbot on the side, not inside your workflows.

    Source: Only 5% use agentic AI end to end. Deloitte, 03/2026

  • The pilot ran, and none of it shows up in the P&L.

    Source: 95% of GenAI pilots show no P&L impact. MIT NANDA, 2025

  • AI turns out more expensive than planned.

    Source: 33% say it costs more than expected. Bitkom 2026, n=604

  • Success is never measured, so it is never repeated.

    Source: Only 6% are high performers. McKinsey State of AI 2025

Three checkmarks or more? Then we should talk.

Book an intro call

Perspective

AI is no longer an experiment. It is an execution challenge.

The leading voices in technology agree on one point: the advantage does not come from access to AI, it comes from putting it to work consistently inside business processes.

Photos: Mark Cuban: Gage Skidmore, Wikimedia Commons, CC BY-SA 3.0 · Bill Gates: Lula Oficial, Wikimedia Commons, CC BY-SA 2.0 · Satya Nadella: Briansmale, Wikimedia Commons, CC BY-SA 4.0 · Jensen Huang: Peter Dasilva, European Union, Wikimedia Commons, CC BY 4.0

Independent statements on the economic relevance of AI. Not an endorsement of this offering. Translations are editorial.

The technology is here. The bottleneck is putting it to profitable use.

Identify your first automation process

What I do

How we work together as partners

I stay with you personally as your partner until Claude, ChatGPT and Codex do productive work inside your processes. Three steps, clearly separated and verifiable at any point.

  1. 01

    Claude Reality Check

    30 minutes, free

    Where is AI burning money today, and where is the fastest measurable process gain? You get a straight assessment, not a sales pitch.

  2. 02

    Roadmap with a defined result

    Before anything gets built

    For every process we define a verifiable work result. For example “invoice fully processed” or “ticket resolved”. Only once that is agreed do we build.

  3. 03

    Ongoing partnership

    In operation, not in a project report

    I work with you directly as the CEO: implementation with your team, operations, quality reports, further development.

Pricing

You buy results, not tokens

You never get a token-based bill. You pay an agreed fixed price per verified process result plus a predictable monthly retainer.

You never get a token-based bill. Model choice, token consumption, repeat runs and quality assurance are my problem. You pay a predictable monthly retainer plus an agreed fixed price per verified process result. A case is only billed once it has been checked and accepted.

An AI result that nobody can verify is not a result. It is a risk in nice formatting.

Verified, not claimed

Every result is checked. Only then is it billed.

Predictable

A fixed monthly retainer and a fixed price per result. No surprises on the invoice.

Fair

You do not pay for failed attempts. The risk of the model choice sits with me.

We agree the specific terms in the intro call, once it is clear which processes go first.

Evidence

Numbers instead of promises

Every number comes from documented work in my own company or in a client mandate. What I show you runs in my own day-to-day operation.

~90%

model cost savings through role-based routingSource: Own live operation

108%

ROI at 5.8 months payback and around EUR 201,400 business value per yearSource: AI officer reference mandate, 2026

610 KB → 95 characters

measured context reductionSource: Audited

In use every day

an own, versioned and audited AI operating systemSource: Own live operation

All numbers from documented own or client operation. No bought logos, no invented testimonials.

Tilo Hammer, your partner for Claude, ChatGPT & Codex

About me

Tilo Hammer

I am an entrepreneur and have worked with AI systems in my own business for years. What I recommend, I have built and run myself first.

That includes my own versioned and audited AI operating system, which runs every day and whose cost and quality I measure. The numbers on this page come out of that operation.

Today I work with owners and CEOs of owner-led companies to put Claude, ChatGPT and Codex into productive use. No consultant vocabulary, straight at your processes.

Talk to me directly

Free guide

Getting started with Claude, for CEOs

The compact guide: the 10 processes where Claude earns its money first in a mid-sized company. The guide is available for download right after you submit the form. (Guide in German.)

Start the potential checkRequest the guide

No newsletter signup. You get the guide and nothing else.

Frequently asked

What CEOs want to know up front

The six questions that come up in almost every intro call. Here are the answers in advance.

What does partner mean here?

I work with you as the CEO on an ongoing, personal basis to put Claude, ChatGPT and Codex into productive use. You do not get a project report, you get processes that are built and operated. I work with your team until the process runs in day-to-day business. After that I stay on board for operations and further development. Partner describes the way I work, not an official partner program of Anthropic or OpenAI.

What does it cost?

There is no token subscription. You pay an agreed fixed price per verified process result plus a predictable monthly retainer. We set the specific terms in the intro call, once it is clear which processes go first.

How fast do we see results?

The first process typically goes into production within a few weeks. We deliberately start with a workflow that is quick to measure. In the reference mandate the payback was 5.8 months.

Is this GDPR compliant?

We start with a governance baseline and document every data flow. Where it is needed we use European processing options. GDPR is part of the roadmap, not an afterthought at the end.

We already have Claude licenses, isn't that enough?

A license is a tool, not a process. As long as nobody builds, measures and operates the workflows, Claude and ChatGPT stay a chatbot next to the actual work. That is exactly the gap the partnership closes.

Contact

Claude Reality Check: 30 minutes, free

You outline your situation briefly, I tell you openly where your fastest measurable process gain is. If I am not the right person, I will say so.

  • No sales pitch, no slide deck.
  • Reply within one business day.
  • No tracking, no advertising cookies. Only one functional cookie for your language choice.

EU hosting available. GDPR is part of the roadmap, not an afterthought.

Or write directly: info@dirigentos.com