de felsen
Understand. Build.

Agents that
do the work.

We build agentic AI, RAG systems, chatbots, workflow automations and live dashboards for businesses.

00The idea

An agent reads, decides and acts across the tools you already use. We design the workflow around your problem, build the agent, and give you a dashboard that shows what it did.

01What we build

Agents and automation,
built around your work.

Agents do the thinking. RAG keeps their answers grounded in your own data. Chatbots talk to your customers. Workflows move the work between systems, and the dashboard shows all of it.

01Thinks and acts
Reads the requirement
Gathers the context
Decides what to do and what not to do
Performs the task
Does a QA
Reports back

Agentic AI

Agents that read, reasons and takes action on a task, the way a trained team member would. It works through each step and checks itself before anything goes out.

02Answers from your documents
pricing.pdfpolicies.docxhandbook.mdcrm export
Refunds are allowed within 30 days for annual plans.source: policies.docx, p.4

RAG systems

Agents that search your own files, policies and data before they answer, and show where each answer came from. No made up replies.

03Talks to your customers

Chatbots

Website and WhatsApp chatbots that answer questions, qualify visitors, take bookings and hand over to a person when it matters.

04Built for scale
SalesFinanceSupportHROpsLegalApprovalsAudit logAccess

Enterprise agents

Agents that work across teams and systems, with approvals, access controls and a full record of every action they take.

05Runs on its own

Workflow automation

The repetitive handoffs between email, forms, CRM, spreadsheets and accounts, joined into one flow that runs on its own, day and night.

06Shows the results

Dashboards

A live view of what your agents did, what they are doing now and where they need a person. Hours saved, tasks run and exceptions, in one place.

02Watch it run

One task,
start to finish.

This is what an agent does when work arrives. Each line is a step it takes on its own. Pick a task to see a different run.

elapsed 0.0s
Result
Working
Example run. Names and figures are made up to show how it works.
03Workflow automation

Work moves
by itself.

Every workflow follows the same shape. Something happens, the agent understands it, makes a decision, acts in your systems and records what it did. You set the rules. It runs them every time.

Trigger

Something happens

An email, a form, a new row, a payment.

Received0
Understand

Agent reads it

Pulls out who, what and how urgent.

Read0
Decide

Your rules apply

Plus judgement where the rules stop.

Decided0
Act

Systems update

Replies, records, invoices, tasks.

Done0
Report

You see it

Logged to the dashboard. Odd cases come to you.

Logged0
Now passing through: Order #4821 from a returning customerExample flow
04Dashboards

See everything
your agents did.

Each build comes with a dashboard. You see the work done, the time saved and the few cases that need a person, updated as it happens.

Operations · TodayLive · example data
Tasks completed1,284+12% vs last week
Hours saved this week63.5about 8 working days
Average handling time41swas 14 min by hand
Waiting for a person3review needed

Tasks per hour

Activitylatest first

    Dummy numbers, used to show the dashboard only.
    05How we work

    Understand the problem.
    Build what it needs.

    We don't start with technology. We start with the part of your business that is slow, repetitive or keeps going wrong.

    1. 01

      Understand

      A call to find the real problem in your business, how the work is done today, and what the agent needs to do.

    2. 02

      Build

      We build the agent and the automations, connect your tools and test on real cases. It goes live with a dashboard, and we keep improving it.

    06FAQ

    Got questions?
    Here are answers.

    What is an AI agent?

    An AI agent is software that can read information, decide what to do and take action, one step after another, until a task is finished. Unlike a chatbot, it does not wait for you to ask the next question. It carries the work through and tells you what it did.

    How is that different from normal automation?

    Normal automation follows fixed rules: if this, then that. It breaks when something unexpected arrives. An agent can handle messy inputs like free text emails, PDFs and unusual requests, and use judgement inside the limits you set. We use both. Rules where rules work, agents where judgement is needed.

    What is RAG?

    RAG stands for retrieval augmented generation. Before the agent answers, it searches your own documents and data for the relevant facts and uses those. Answers stay accurate to your business and can point to the source they came from.

    Do you build chatbots too?

    Yes. We build chatbots for websites and messaging apps that answer questions from your own content, qualify visitors, collect details and book calls. The one on this page is an example: open it from the top right.

    Which tools can it work with?

    Most tools a business already uses: email, CRMs, spreadsheets, accounting software, help desks, calendars, chat apps and your own databases. If a tool has a way in, we can usually connect it. We confirm this on the first call.

    Is my data safe?

    Agents only get access to what the task needs. Sensitive steps can require a person to approve before anything is sent or changed, and every action is logged so you can see exactly what happened.

    How long does a build take?

    It depends on the problem and how many systems are involved. After the first call we agree a clear scope and timeline before any build starts.

    What happens after it goes live?

    You get the dashboard from day one. We watch the results with you, handle the cases the agent flags, and improve the workflow as your business changes.