Getting started with AI: AI consulting on your own tasks
Every client starts with a free AI assessment: 60 minutes by video and a written recommendation. Many businesses bought AI long ago, and everyday work still does not change. The tool is rarely the reason. We start from your team's work, in sessions billed by the hour with a ceiling.
Book a free AI assessmentWhen nothing changes despite the licenses, there are usually three reasons: nobody has worked through their own tasks with the tool, it is unclear which company data may go into a prompt, and intermediate results are still copied by hand from one program to the next.
That is why you buy no packages from us. We first find out how your team works and uses AI today, and then take only the step that helps. We use existing tools where they are enough and build missing connections or custom software where needed.
You do not need to know in advance which tool you need or which task to start with.
How working with us starts
- 01
Free AI assessment
60 minutes by video with you as the owner or managing director, ideally joined by someone who does the work. We find out how your team works and uses AI today and look at one or two recent cases that were laborious, slow or error-prone. You do not need to prepare or document anything. You hear a first assessment at the end, and within two working days you receive a written recommendation from Sebastian Behler: which improvement solves which of your business problems and what the next step is. Where standard is enough, we say so, and no plan follows. The assessment is free and without obligation, once per company.
- 02
A short plan as the offer
If a step with us makes sense, you receive a short written plan: what we do, what comes out of it, the hours we expect and the ceiling. The plan also states the hourly rate. We exceed the ceiling only with your approval in text form.
- 03
Work in sessions as you need it
You book coaching, workshops, meetings and development work when you need them. We bill by the hour, monthly and with a time report. You can stop at any time.
What we work on in the sessions
The plan brings together what you need right now: coaching, consulting or implementation. You do not have to go through any of it in order.
Your team hands over whole tasks
On cases from its own week, your team practises handing whole tasks to AI and accepting the result: with a goal, material, boundaries and a review step. For individuals, shorter sessions are enough; for a team, there is the two-day AI enablement workshop.
You know which tool fits and what may go into it
We advise you on which AI tool fits your rules and tasks, which data may go into it and which access your team really needs. Data protection and the main points of the EU AI Act are part of this, as is a list of what may go into a task.
AI works with your files and programs
We connect the tools your team works with today to the AI: file storage, project and task management, email and documentation. For this we use the ready connectors of Claude, ChatGPT Work with its always-on dots, or Microsoft Copilot, and written work instructions; nothing is built along the way. For colleagues who write code, we set up Claude Code or Codex. Each agent gets exactly the access its tasks need. We put the documents that matter in order and record which source applies.
Software that belongs to you
We build custom software for connections and workflows that go beyond ready connectors: programs without a ready interface, such as an ERP system, multi-step workflows across several programs, fixed automations for every step that follows a rule, interfaces of your own, and agents that run without anyone starting them, if you wish on your own server with open-source agents such as Hermes Agent or OpenClaw, when data and model should stay in-house. This pays off when the rules and exceptions of your workflow are too precise for a written instruction, when the work should run without anyone triggering it, or when you want to see exactly how the AI works.
What runs keeps running and gets better
When a task or software is running, we can run and improve it for you, billed by the hour with a monthly ceiling. If it runs in software we built, we record every run: which documents and instructions it used, which model, what it cost, and whether your team accepted, changed or rejected the result. If you work with Claude, ChatGPT or Copilot, we analyse usage and costs. Corrections go back into the source or the work instruction, and we test changes on the same real examples. Your team checks the results; prices, contracts, payments and anything that goes to your customers are approved by the responsible people in your company. If your team wants to take over a task itself, we hand it over with review instructions, practice on real cases and a supervised month.
One open question, tested on real examples
Where a step is not obvious, we test the one open question in a bounded pilot with a few real examples from your company. You then decide whether to continue.
An analysis workshop when several workflows are in question
When several teams or workflows are in question and the assessment does not settle where to start, we spend a day with your team on the main workflows, and you receive a written report. We bill this by the hour too.
By the hour, with a ceiling
The assessment costs you nothing. After that we bill by the hour, with a ceiling we exceed only with your approval, and every invoice comes with a time report. You can stop at any time and pay the hours up to then. A booked meeting counts at least two hours and a workshop at least four, each including preparation and follow-up. What we build belongs to you.
What you can rely on
You normally buy your licenses yourself, directly from the provider, on your own contract. Our recommendation for a tool does not depend on who sells you the seats.
We add four things to your AI tool: people who can hand over tasks and check the results; processes that settle who approves what; knowledge, meaning work instructions, reliable sources and examples; and software where the tool alone is not enough. You keep the first three if you change tools. Your work instructions stay with you as plain text files.
We use ready connectors where they have proven themselves. When a rule must always hold, for example that an email is only ever created as a draft, we build it into the connection itself, not only into the agent's instructions. It then holds whichever agent uses the connection.
Governance is part of the workshop and not an add-on package. Data protection and the main points of the EU AI Act come first, before a team starts putting real company data into a tool.
We speak openly about tools and providers. What another company's product can do, we say as it is. What we build ourselves belongs to you.
You talk directly to the developer who is responsible for your solution. No sales team in between, and no handover to someone who does not know your business.
Book a free AI assessmentFAQ about getting started with AI
What does the AI assessment cost?
Nothing. The AI assessment is free and without obligation: 60 minutes by video and a written recommendation within two working days. Each company gets one.
What does the work cost after that?
We bill by the hour. Before you commission anything, a short plan states the hourly rate, the expected hours and a ceiling that we exceed only with your approval in text form. Every monthly invoice comes with a time report, and you can stop at any time.
Do we need to know in advance which AI tool we need?
No. That decision is exactly the one we take off your hands. You do not need to know which task is a good fit either. We go through how your team works and ask questions until the tasks that cost time today become visible. Then we work out the rules and the exceptions, and recommend the smallest approach that helps reliably in everyday work. That can be a tool you already have.
May our company data go into Claude or ChatGPT?
That depends on the plan you booked and on the settings, not on the name of the tool. In the workshop we clarify, for your own accounts, which data may go in, which may not, and how your staff can tell the difference. We claim only what is actually configured in your account.
Do we need our own licenses?
Yes. You normally buy access to Claude, ChatGPT, or Microsoft Copilot directly from the provider, on your own contract. Our recommendation does not depend on who sells you the seats. In the workshop we tell you which access your team really needs, so that you do not pay for seats nobody uses.
Book a free AI assessment
60 minutes by video: we find out how you use AI today and which work slows your business down. Within two working days you receive a written recommendation. Free and without obligation.
Give us a call:
+49 30 3505 6713Or write to us:
hello@certaince.comChat with us on WhatsApp:
https://wa.me/493035056713