AI & Automation
AI agents and automations that do real work.
Most businesses do not need an AI strategy. They need the repetitive work to stop landing on people: answering the same questions, copying data between tools, reading documents, chasing follow-ups. That is where we start.
We build the assistants, chatbots and workflows that take that work on, using Claude, GPT, Gemini or open models, and connect them to the tools you already use.
What we build
AI agents and assistants
Assistants that can look things up, take actions in your systems and hand off to a person when they should.
Chatbots on WhatsApp and the web
Customer-facing bots that answer from your own content, capture leads and pass tricky conversations to your team.
Workflow automation
Triggers, approvals, notifications and data syncing between your tools, built in n8n, Zapier or Make.
Knowledge bases (RAG)
Search and question-answering over your own documents, so answers come from your material and not from guesswork.
Document and data extraction
Turn invoices, forms, PDFs and messy emails into structured data your systems can use.
Private AI with Ollama
When data cannot leave your servers, we run open models locally instead of sending it to a hosted API.
AI inside your product
Search, summaries, classification and smart replies added to an app you already run.
How a project runs
- 01
Pick one job
We start with a single, well-defined task instead of trying to automate everything at once.
- 02
Prototype on your real data
A working version early, so you judge actual results instead of promises.
- 03
Add guardrails
Checks, logging and a human handoff for the cases the automation should not decide alone.
- 04
Launch and tune
We watch it run, fix what surprises us and expand from there.
Related work
Questions
Which AI model should we use?
It depends on the task, budget and privacy needs. We work with Claude, GPT and Gemini as well as open models, and we compare them on your own examples instead of picking a favourite. If your data has to stay on your own servers, open models running locally with Ollama are an option.
Should we use n8n, Zapier or Make?
Zapier is the quickest to start with and has a very wide app library. Make handles visual, branching workflows well. n8n can be self-hosted and gives the most control, which suits sensitive data or heavier logic. We pick per project, and we can write custom code when a no-code tool is the wrong fit.
Can you add AI to an app we already have?
Yes. We can add LLM features like search, summaries and smart replies to an existing product, or automate manual work around it, connected to the systems you already use.
What do you need from us to get started?
A description of the task as it works today, the tools involved and a handful of real examples, such as typical messages, documents or records. Send that through the contact form and we will tell you what is realistic.
Is our data safe with AI tools?
We only send data to a model provider when that is acceptable for your use case, and we tell you which provider sees what. For sensitive material we can keep processing on your own infrastructure with local models.