Your agentic employee drafts, typesets, researches and follows up while you do something else. You get finished work, the choices it made, and a short list of what needs you.
Nora, please turn the attached notes into an offer in our template. Two-day workshop, twelve people, price as in the notes. Send it to me only. I will forward it to the client myself.
DOCXworkshop-notes.docx
NoraAgentic employee
Tue 09:26
Re: Offer for the spring workshop
To: Laura Bennett
Laura, the offer is attached as a PDF in the company template.
Two choices I made, so you can overrule them: the price is stated without VAT, as in your notes, and the offer is valid for 30 days.
One thing needs you: the notes give two dates for day two, 14 and 15 May. I used 14 May.
Nothing has gone to the client.
PDFOffer-spring-workshop-v1.pdf3 pages
02 Ask in chat
16:44
NoraEncrypted chat
Example
Today
TranscriptI am in the car. What did the venue say in the email I forwarded, and is the offer ready?
16:42
Nora
Short version. The venue confirmed Thursday the 14th: room for fourteen, projector included. The offer is ready as version 2.
Open point: they need the final headcount by Tuesday noon.
16:43
Remind me on Tuesday at 9 about the headcount.
16:44
Nora
Done. Reminder set for Tuesday 9:00, here in chat.
16:44
Message
Illustrative example. All names and details are invented. Chat is an optional module.
Every step logged
Made to measure
Cloud, Azure or on-premise
The agentic employee
An AI colleague with a role, permissions and a manager.
It works through real tasks by email and chat, remembers every project, and hands people only what a person should decide.
Work in your template
Offers, reports and slide decks come back typeset in your template, checked before they reach you.
Memory that lasts
It picks up a project a week later and still knows what is open, what went out and who owes what.
Nothing sent without you
It never writes to anyone on its own. A third party gets a message only on your explicit instruction.
Limits enforced in code
Who may give it work, what it may read and whom it may answer is checked by code, not left to good manners.
A record of every step
Every message and every result is kept in full, so you can always see what it did and why.
An agentic employee is set up for a role, with your templates and your way of working. Three of the ten roles we describe:
Executive assistant
Forward the thread. Get back what they want from you.
For a director, founder or partner with more mail and meetings than hours. A forwarded thread comes back as a short account of what is wanted from you, or as finished outputs with a draft reply in your voice.
Briefing notes for tomorrow’s meeting, typeset in the company template
Follow-ups kept: “come back to me if they have not answered in a week”
Left to you: what goes to people outside, every decision and every signature
InboxExample
LB
Laura Bennett
Tue 13:58
Fwd: Room booking, spring workshop
To: Nora
What do they want from me?
Forwarded thread: Room booking, spring workshop
NoraAgentic employee
Tue 14:05
Re: Fwd: Room booking, spring workshop
To: Laura Bennett
Laura, one thing is wanted from you: the final headcount by Tuesday noon.
The rest is settled. The venue confirmed Thursday the 14th: room for fourteen, projector included.
Draft reply in your voice Not sent
Thank you for confirming Thursday the 14th. You will have our final headcount by Tuesday noon.
Sales support and offers
From notes to an offer in your template.
For an owner or sales lead who writes offers between meetings. Notes or a forwarded inquiry come back as an offer in the company template, with prices marked as with or without VAT.
Version 2 after your comments. Version 1 stays in the project.
The client deck in the same visual style
Left to you: price and terms, negotiation, and sending to the client
Offer-spring-workshop-v2.pdfExample
LarkfieldOffer · page 1 of 3
Offer: spring workshop
Prepared by Laura Bennett, Operations Director
Item
Price
Two-day workshop for twelve people
Day two: 15 May
Total
Prices without VATValid for 30 days
Larkfield · Offer, version 21 / 3
Project coordinator
“Where are we with X?” Answered a week later.
For a project lead and a small team running several projects side by side. It keeps a file on every project: status, journal, files, and the full history of messages from email and chat.
A meeting recording becomes a transcript, then minutes with action points
Promised deadlines watched, with a draft chaser ready the day after one passes
Left to you: priorities, changes of deadlines, and the decision to chase people outside the team
10:12
NoraEncrypted chat
Example
Today
Where are we with the spring workshop?
10:12
Nora
The offer is done as version 2. You have had it since Tuesday. The venue confirmed Thursday the 14th.
10:12
I am watching two deadlines. On Tuesday at 9:00 I remind you of the final headcount. The caterer promised the menu by Friday, so I come back to you on Saturday morning.
10:12
One thing needs you: the speaker’s slides were due on Monday. A chaser is drafted. I have not sent it.
The agentic employee is our main product. The same team also designs and builds software to order, and helps your people get AI working.
Custom software
Tools, integrations and dashboards built around a process that deserves its own. Including deployment, security, documentation and support once it is live.
Web applications
Internal systems, portals and dashboards that run in the browser and connect to the systems you already use.
Websites
Company and product websites with a clear structure, fast pages and search engines in mind.
Mobile applications
Apps for phones and tablets, for your customers or for your own people on the move.
AI systems and applications
Software with AI inside: automations, assistants and agents. Leading cloud models by default, Microsoft Azure or fully on-premise when your data calls for it.
AI audit and strategy
We map how the work actually flows, find the tasks where AI pays for itself and price them honestly. You leave with a ranked list, a budget and a business case.
Training and workshops
Hands-on sessions built around your team’s real tasks, from first steps with AI to building your own assistants. For leaders, specialists and whole departments.
Digital twins
Digital twins of patients, processes and products. Part of our research and design work in the fields below.
Where we work
Fields we know well enough to go deep.
Our services work in any organisation that runs on documents and decisions. In the fields below we also bring domain knowledge of our own: the research, the regulation and the vocabulary your people use.
Healthcare and biomedicine
Hospitals and research teams run on documents, and most of those documents never leave the building. We work with models that can stay on your own hardware, and with teams who need the output checked, not just generated.
Clinical and administrative documents
On-premise deployment
Digital twins of patients and care processes
Pharma, biotech, chemistry and toxicology
Literature, safety data and regulatory text at a volume no team can read. We build the search, the summaries and the checks that show you where the model got it wrong.
Literature and evidence screening
Safety and regulatory documents
Support for discovery and lab work
Medical devices, regulation and market access
Work that a regulator will read later. We prepare, structure and check the documentation, and we keep the record that an audit asks for.
MDR and IVDR documentation
HTA and reimbursement dossiers
AI Act compliance for your own systems
Defence, CBRN and public safety
Training, exercise records and decision support in a field where AI is still rare and the rules are strict. We say clearly what a model can support and what stays with a person.
Training for civil protection
Exercise and incident reports
A human in the loop by design
Public administration and regulated industries
Your data may not go to a public cloud, and every decision has to be explainable. Both are design questions we answer before the first prototype.
On-premise or Azure
Decisions you can explain
Records that survive an audit
Universities and research organisations
Teaching, research and the administration around them. We have run this from the inside, so we know which parts are worth automating and which are not.
AI in teaching
Research workflows
Grant and project administration
Not on the list? Ask us and we will tell you honestly if the fit is not there.
We do not build systems intended to design, optimise or acquire harmful chemical, biological or nuclear capability.
How we work
Start with an audit. Then take the path that fits.
The same services, combined in a different order. For one client training is the result; for another it comes after the agent is live, so the team learns to work with it.
Hand work to an agent
An agentic employee takes over a defined workload.
Audit
Pilot agent
Training on the agent
Scale up
Support
Build a tool
Custom software around a process that deserves it.
Audit
Build
Integrate
Support
Upskill a team
Your people learn to use AI on their own work.
Audit
Training
Follow-up
Founders
Academic depth. Business experience.
Between us we bring experience from university and faculty leadership, from founding and running companies, from research projects at home and across Europe, and from turning research into products and services.
CEO & CTO
Dr. Dominik Palla
AI researcher and software engineer
Leads all research and development and stands behind every product and implementation. Runs the company.
Experience leading a university AI centre; training, workshops and consulting for companies and public institutions
Founded and ran a software company for six years; hands-on engineer and team lead
Applied AI research and contract projects; publishes on agentic AI systems
Computer scientist with a PhD in applied informatics focused on AI, from the University of Hradec Králové.
CEO & CTO
COO & CFO
CBDO