Tooling that genuinely works.

No questionnaire. A conversation. We're at home in brand and communication research, together with you we dig for the right questions. On top of that an AI that runs the conversation: moving with the occasion, probing where it matters, and bringing in your own brand knowledge. Hosting in Europe, GDPR-proof, the data stays yours. AI runs the conversation, in the customer's language both in and out.

The Schwung Brand scan is the living example: 21 questions that move with the occasion, plus deep-dives where it matters most. What used to take a quarter can now be done in a week. Read the case →

A second application in development: a regional knowledge hub for internationals, where instead of searching, visitors have a conversation with the site. It recognises the role, answers in their own language, and removes the threshold of taking that first step in a new system. Read the case →

Useful for

• intake that now takes too much time
• research where you want to speak with many people briefly
• ad-hoc introductions that lack structure
• qualifying leads, candidates or partners
• onboarding where everyone needs the same foundation

What you get

• your own branded conversation tool, on your own domain
• a structured reflection per interview
• access to the back end to review everything
• monthly adjustments based on what stands out

No loose prompts. We bring brand expertise, you bring your own voice and history. Together we build the foundation: tone, jargon, archetype, do's and don'ts. On top of that a writing tool that draws from your own brand material, a second layer that checks every text against your brand rules, and permissions per role or team. No training on your data. AI is the engine, not the voice.

The gain isn't in speed. But in recognition. Vacancies that sound like your culture. Blogs that don't feel like boilerplate. Social posts that make people say: 'that's from them'.

A case in development at an organisation in construction: AI that gathers knowledge from the people doing the work and makes sure the message lands, even without a communications department. Telling the story, reaching the audience, deepening the conversation, at the pace the schedule demands. Purpose over means. Read the case →

Useful for

• vacancy texts at scale, without them becoming identical
• blog and knowledge-base pipelines
• social content in your own tone, not the algorithm's tone
• product descriptions for catalogues with hundreds of items
• email campaigns and customer-journey communication

What you get

• a brand-foundation document, also usable outside AI
• a writing tool that always brings that foundation along
• a quality-check flow so output doesn't just go live
• training for your editors or writers

Lots of data, little time. Together we look for the patterns that matter, brand principles steer the question, not the technology. On top of that an AI that combs through your own data: groups similar answers together, recognises which feeling goes with which topic, flags names and organisations, and surfaces the recurring themes. Privacy by design, data never leaves your environment. AI lays the first layers, a human draws the conclusion.

No dashboard to look at, but conclusions to decide with. What used to take a week, in a day. And in language the organisation recognises.

Useful for

• customer research with 50+ conversations to wade through
• team scans where you want to see patterns between perspectives
• competitor analyses on public information
• website audits for brand consistency and positioning
• sales-call analysis to find winning patterns

What you get

• an analysis on your own data, not generic
• monthly or ad-hoc reporting in understandable language
• raw data stays yours, never goes to a third party
• a methodology you can also follow manually

New tools emerge from real questions. Together we dig for what the question really is, pragmatic discovery at the intersection of technology and communication. And which regulations it has to fit (healthcare, government, AI Act, sector-specific). After that we bring the building blocks: AI, your own knowledge source, smart collaboration between components, a quality check, your own interface. On your domain, on your own infrastructure. Four weeks from first conversation to working prototype.

That's why this always stays open. What will be standing here in 2027 we don't know today, and that's exactly the intention.

First step

• a one-hour conversation in which we see whether there's something to build
• no preparation needed
• after the conversation an honest assessment: yes or no, and if yes what
• no obligation, yes, really

What it can deliver

• a prototype within 2-4 weeks
• or: the insight that AI can't solve it, also valuable
• or: a referral to someone better suited for it