We build AI systems for marketing teams, tailor them to your processes, and hand them over so your team can continue to develop them on its own.
since 2016
at the market
max. 12
Mandate simultaneously
20.000+
Employee (Reference Client)
bvik + ABG
AI Instructor
The licenses have been distributed, the training has taken place. Nevertheless, everyone prompts differently, the results cannot be reproduced, and what comes out sounds like no one. In the end, the rework takes as long as writing it yourself.
The content roadmap stands in parallel. Not because ideas are lacking, but because the team can't manage them. There are no new positions.
And last year's project? There was a presentation, a tool, maybe a pilot. None of it transitioned into regular operations.
The models are good enough. Anyone on a marketing team who isn't getting useful results from AI today rarely has a model problem. And it's not due to a lack of discipline in the team either.
What's missing is the framework: Where in the process does the AI get involved, what knowledge about the brand, terminology, and product does it receive, who checks the result, and how does the team recognize that something is good enough? Without these four answers, every prompt remains an individual achievement. With them, it becomes a system that works even when a colleague is on vacation.
That's why we don't build prompts. We build the framework and fill it with your knowledge.
For teams that regularly need large volumes of text. The system understands your formats, terminology, and quality criteria, and guides you through the entire process—from the initial briefing to final approval. We built this for the content team at an international live marketing agency to handle Gloss articles, blog posts, and press releases.
For teams that need to get into the discussion faster. Instead of one idea someone has polished overnight, several well-thought-out variations are on the table the next morning. For an automation technology company, we are currently building a system that allows the design team to develop image concepts from input by the customer journey team.
For anything that someone manually compiles today: competitive intelligence, market data, content research, evaluations. The system does the preparatory work, your team decides.
If your case is not listed, that is not an exclusion criterion. The three types describe where we most often engage, not the limit.
We'll examine how your team works today and identify the use case with the best effort-to-impact ratio. The result will be a decision paper, not a slide deck with options.
Duration: [DURATION PHASE 1]
We are building the system and testing it with real tasks from your daily work, not with demo examples. Your team is involved because knowledge of the brand and product comes from the team.
Duration: [DURATION OF PHASE 2]
Your team is learning to work with the system and adapt it. We document what we've built in a way that someone will still understand in two years.
Duration: [PHASE 3 DURATION]
The system is now running in normal operation at your site. We’ll hand it over to your team and provide support as long as you need it. This is where Inspect & Adapt begins—more on that in a moment.
Ongoing, as needed
A system that is set up once and then left to its own devices becomes obsolete faster than traditional software. Models change, their capabilities grow, and your own processes also don't stand still. What worked well in the spring can be out of sync with reality by the fall.
That’s why every system includes a regular cycle of inspection and adaptation—Inspect & Adapt. What still delivers good results, what no longer does, which new opportunities are worth pursuing, and how has your process shifted?.
New generations, new strengths and weaknesses — what was the best way yesterday might not be today.
What's possible with AI is shifting in months, not years. A system should benefit from that, not be bypassed by it.
Teams, formats, and goals change. The system must follow these changes, otherwise it will be circumvented.
Not "Set & Forget"We can take over this tempo, or we can show your team how to lead it themselves. Both are possible. "Turn on and forget" is the only option that won't work.
Both clients did not agree to be named. Therefore, we are describing the projects without names.
Starting situation: [STARTING POINT: Team size, formats, languages, what the bottleneck was]
Built: A system for text production across multiple formats, from glossary entries to press releases, with documented quality criteria for each format.
Result: [RESULT: qualitative, what has changed in daily life]
Starting situation: [BACKGROUND: how image concepts originated previously, where things got stuck]
Built: A system that develops image concepts from the input of the customer journey team, so that alignment starts with multiple variations instead of just one.
Result: [RESULT: qualitative]
A system that runs in normal operation and fits into your existing tools.
A documentation showing how it is built and how to change it.
Quality criteria for each format, so your team can determine when a result is good enough.
A team that operates and further develops the system without asking us.
Key metrics that show whether the effort was worth it.
We come from marketing, not AI development. The difference shows where it counts: We know how a release cycle works in an industrial company, why the specialist department insists on a particular wording, and what a content team really needs on a Thursday afternoon.
We take on a maximum of twelve mandates at a time. Those who work with us work with the people who also build.
And we build our solutions so that you won't need us afterward. Enablement is part of the project, not a follow-up offering.
Bernd [LAST NAME], Founder and CEO
AI strategist and enthusiast. Lecturer at bvik and ABG on AI in Marketing and agent-based systems. At webraketen, he brings these two together: the strategy and the systems that emerge from it.
This is a good fit if your marketing team produces its own content and volume is the bottleneck. If you’ve already tried AI and know that the problem lies in more than just the tool. If you want a system that belongs to your team.
This isn't the right choice if you're looking for a ready-made product that you can license and put into use. It's not the right choice if no one on the team has time to collaborate, because without your input, the system will end up being generic. And it's not the right choice if the sole goal is to make it presentable internally.
This is the normal case and not an obstacle. What the team needs to bring is knowledge about the brand, product, and process. We will bring the AI part and pass it on.
We'll work this out before development begins, together with your IT team and your data protection department. Depending on your requirements, we'll use European hosting or models that run within your existing environment. [ADJUST TO ACTUAL SETUPS]
No, and that's the point. After the handover, your team will operate and maintain the system. We can support you through the ongoing cycle of testing and adjustments, but you aren't dependent on us for that.
[TOTAL DURATION, Framework]
Yes. That's more the rule. Part of the work in Phase 1 is to make the actual process visible, not the one in the manual.
The effort involved depends on how many formats the system needs to support, the size of the team, and the depth of integration with your existing tools. During the initial consultation, you’ll get a rough estimate of the scope before you commit to anything.
We establish the key metrics before construction begins—typically, turnaround time per format, the number of revision cycles, and monthly volume, assuming the team size remains the same. After handover, you can track these metrics yourselves.
Thirty minutes during which we'll assess where an AI system can be beneficial in your marketing—and where it isn't. You'll receive our assessment, even if we don't end up working together.
[TICKET BOOKING or FORM]
3 Einträge
3 Einträge
4 Einträge
5 Einträge
4 Einträge
