How we work
We think with you,
build it and stay.
Your AI department, alongside the people you have. Your people know the business. We know AI.
What we bring: We build AI for businesses and run our own business on it, with our own AI agents. What works in one place, we bring to you. You own your data and the setup we build specifically for you.
A few lines are enough. If you would rather talk, pick a time for 20 minutes of free advice with Gustav.
The systems you choose to connect
Your data. Your knowledge. Your rules.
What AI may do on its own, and what you approve.
Customer cases
A small selection of what we have built.
Systems we have built for businesses. We will build the next one together with you.
1 / 9
Meetings & agreements
Meetings that become tasks
After the meeting, it says what you promised and what the customer promised.
The film: from the meeting to tasks that each have an owner. The meeting is in Danish, so the transcript is too. Names and figures are examples.
- What we built
- A meeting recorder for Teams, Meet and Zoom. The meeting is transcribed on the computer itself, and AI finds the promises in the minutes and splits them into yours and the customer’s.
- What it gives you
- Nobody has to write minutes. You choose which promises become tasks, and each task is listed with customer and owner.
- The meeting is recorded
- Transcribed locally
- Promises become tasks
- One minute of audio transcribed in 6.4 seconds
- The transcript is made on the computer itself
In operation today
CRM & customer overview
Customer CRM that updates itself
The company’s internal CRM with a page per customer, where emails, meetings and tasks land by themselves in date order.
The film: the customer’s page in the CRM updates itself, and the board brings all customers together. Names and figures are examples.
- What we built
- An internal page per customer with contacts, the agreement, the projects and a timeline where emails, meetings, email drafts and tasks land by themselves. Above the pages sits the board with all customers, owner and latest contact, and the system suggests next steps based on fixed rules.
- What it gives you
- Employees manage the customers from here: they write to the contact, create tasks and move the customer on. The customer does not see it. Nothing happens until an employee has approved the suggestion.
- Email, meeting, task
- Lands on the customer’s page
- You take the next step
- Each employee sees only what they are allowed to, controlled in the database itself
- On the timeline, emails appear only with date and direction, never what they said
In operation today
Email & prioritisation
The inbox, sorted
The morning’s emails are in three piles: needs a reply, for information and filtered out.
The film: the morning’s emails in three piles with the reason next to each. Names and figures are examples.
- What we built
- An inbox where AI sorts the emails, writes the reason next to each email and can draft a reply based on the customer’s case.
- What it gives you
- You start with the emails that need you, and you do not write replies from scratch. Nothing is deleted, and you send it yourself.
- Email in
- AI sorts
- You reply
- For sorting, AI gets only the sender and subject, never the email itself
- The same email cannot go out twice
In operation today
Recruitment
Referral platform
Do you know a skilled salesperson looking for a new job? Employees and networks recommend candidates via a personal link. The reward is paid out when the candidate is hired.
The film: from a shared link to hire and reward. Names and figures are examples.
- What we built
- A page where each referrer gets their own link, a page where the candidate starts their application, and an overview for the company of referrals, hires and rewards.
- What it gives you
- The company can follow each recommendation from link to hire, and nobody has to keep track of the rewards by hand.
- The link is shared
- The candidate applies
- Hired and rewarded
- Status is fetched from the recruitment system every 15 minutes
- Approved rewards are collected in a file for the payroll system
Running at a company today
Content & visibility
The Content Machine
A system that plans, writes and publishes articles and afterwards measures where they rank in Google.
The film: each article goes through the five steps (shown in English; the articles are written in Danish). The searches and the figures are our own, measured.
- What we built
- A publishing plan with dates. AI writes each article with sources, another AI reads it through, and it is published on the date. Afterwards the system follows the rankings.
- What it gives you
- More people find you when they search. The system starts with the pages that are almost on Google’s first page.
- Topic and sources
- Written and checked
- Published on the date and measured
- 1,462 searches in Google’s top 10 on the four sites we run, 344 of them in the top 3
- 23 times as many impressions in Google in September as in March, on the four sites we run
Built for a company
Sales & follow-up
Sales cockpit with AI
When the salesperson hangs up, the call is transcribed and a draft of the email is ready.
The film: the call transcribed, the outcome suggested and the email ready for review. The call is in Danish, and so is the email. Names are examples.
- What we built
- A call list for salespeople. The call is transcribed in Danish, AI suggests the outcome (e.g. “Call again”) and drafts the email based on what the customer said.
- What it gives you
- The salesperson spends the time on the next call instead of on notes. Nothing goes out until the salesperson has read and sent it.
- The salesperson calls
- AI writes transcript and email
- The salesperson sends
- Nine checks before every call, including the do-not-call list and calling hours
In operation today
Surveys
Participant system for a large annual survey
The system that invites the participants, reminds them to respond and lets the customer follow participation along the way.
The film: from personal link to the customer’s overview of participation. Names and figures are examples.
- What we built
- Personal invitations and reminders, a page where the participant starts the survey with one tap, a page to opt out and a status page for the customer.
- What it gives you
- The customer could follow for themselves how many took part, and no email went out until it had been checked against the texts the customer had approved.
- Personal link
- A human taps
- Live status
- 21 checks test every mailing before it can be sent
- Email programs that open links by themselves do not count as participants
Running at a company today
Customer experience
Customer portals
Each participant in our mentoring programme gets their own private page with meetings, booklets and tasks.
The film: the participant’s own private page with meetings and booklets. Names are examples.
- What we built
- A personal page per customer with the three meetings, the booklets as a page, PDF and read-aloud in Danish, and the small tasks waiting before the next meeting.
- What it gives you
- The customer has all the material in one place between the meetings and can read or listen when it suits them.
- The customer opens their page
- Meetings and booklets together
- Tasks before the next meeting
- The page is private and opens with a code or a personal link
- The booklets come as a page, PDF and audio
Customer portals for personal mentoring programmes
A product for consumers
ForbrugerAgenten
An app where an AI agent writes to the electricity, mobile and internet companies and gathers the offers in one place.
The film: the agent writes to the companies, and the replies land in the app. The companies are real, the messages are examples.
- What we built
- The whole product: the app with MitID login, the agent that writes to the companies from its own address, and the inbox where the replies land.
- What it gives you
- The user does not call anyone, and their private email does not fill up. The user says yes in the app themselves.
- Log in with MitID
- The agent writes
- You say yes in the app
- From 465 impressions in Google in March to 38,645 in September
- In the App Store and Google Play
Our own product
Customers, customer figures and processes in the illustrations are made-up examples. The Content Machine’s searches and figures are real.
The next one is yours.
What should your AI department be able to do?
- What can AI do for us?
- Our people are already working with AI
- We want a plan
- We have a task
- We want everything in one place
Would you rather learn to build it yourself? See the mentoring programme →
More examples on the way, also as films
Enter your email and we will let you know when they are ready.
01 / Your process · step by step
How we build.
We start where you are: with an overview and a plan, with a task, or by bringing together what you have. We build the first thing as soon as the access is in place.
The task, success criteria and price are agreed in advance. We set the timeline once we have seen your systems.
- 01
Find the right place to start
We look at your work and the AI you use today, and point out what is worth starting with.
What you end up withA shared direction and a quote before you commit.What we clarify together
We go through your workflows and systems and agree how we will assess the result. The task, success criteria and price are part of the quote.
- 02
Build on what you already have
We build the first solution around your data and workflow.
What you end up withA pilot with agreed boundaries for data and actions.Systems, access and boundaries
We contact your current software suppliers and help with access to your data and systems. You are in copy the whole way. Together we agree which data and actions the solution gets access to.
- 03
Try it in your everyday work
We try out the solution with your people. They learn to use it along the way.
What you end up withA concrete basis for choosing the next step.Testing and training in practice
We test the pilot, go through the measurements with you and adapt the solution to your terminology. You assess the result together with us before we choose the next step.
- 04
Put it into operation with an overview
We put the solution into operation and agree who does what afterwards.
What you end up withA solution your employees can keep working with.Handover and next steps
We hand over dashboards and checks and agree responsibility, operation and support with you. We look after the solution and keep building when you are ready.
02 / Find your starting point
What needs to work
better together for you?
It starts with a task you know. Here are some of the workflows we can build around.
Tasks between your systems
From order to an invoice ready to send. From email to an updated CRM. We bring together the steps you repeat by hand today.
See automation in practiceAI agents that work together
One finds the relevant knowledge. Another prepares a draft. You agree what the agents may do and where you must approve.
See a concrete workflowKnowledge and overview in one place
Your catalogues, documents and systems as the starting point. We build the connections and views the task requires.
See your knowledge in useEnquiries and telephony
Danish-speaking agents that answer the phone and help with enquiries. The voice agents are built by JesperAI.
See the phone example03 / Built for you
Your data.
Your workflows.
Your decisions.
We build together with you. You should be able to understand the solution, use it and decide how it develops.
You own your data and your own setup
Your data and the workflows we build specifically for you are yours. You can take them with you.
You decide what the AI may do
The AI finds and gathers. It prepares drafts. It only acts on what you have said yes to. You approve customer contact, money and changes to agreements. We send no cold emails.
Your people are involved all the way
We try out the solution with your people and agree responsibility, operation and support with you.
04 / The practical part
Questions before
we get started.
What does it cost to have an AI solution built?
The price depends on the task and your systems. Once we have seen the task, you get a quote before you commit.
Can your AI agents speak fluent Danish?
The voices are Danish and neural. The agent says that it is an AI. The voice is provided by JesperAI.
How long does an implementation take?
We set the timeline together with you once we have seen your systems. We do not give a figure before then.
The first step is a conversation
Where do you stand
with AI today?
Tell us briefly about it. We reply with what is possible and whether it is worth building.
Start with a message.
Or choose a time straight away.
Should the person who leads be able to build and assess what we build themselves? See the personal AI mentoring programme