AI business process automation: let the routine run itself
Diese Seite ist noch nicht in Ihrer Sprache verfügbar — die englische Version wird angezeigt.
Every business has work that follows the same steps every time: an order comes in and someone retypes it, a request waits in an inbox for approval, a report is put together by hand every Monday. That work can run on its own. We build the automations into your web app or internal systems, and teach your team to maintain and extend them with AI, so the next process you automate is one you build yourselves.
Signs a process is ready to automate
- The same data is typed into two or three systems by hand.
- Requests and approvals live in email and chats, and some get lost.
- A weekly or monthly report takes someone half a day to put together.
- Customers wait for a reply, a quote or a document that could go out straight away.
- A step happens only if one particular person remembers to do it.
What we automate
- Intake: orders, requests and applications that arrive in the right system, already structured.
- Approvals: each request goes to the right person, with reminders, and the decision is recorded.
- Integrations: your CRM, accounting, email, messengers and spreadsheets exchange data without copy and paste.
- Scheduled jobs: reports, reminders, invoices and data checks that run on time without anyone starting them.
- AI steps where they help: sorting incoming requests, pulling data out of documents, drafting replies for a person to check.
Where AI helps, and where a plain rule is better
Not every step needs AI. A rule like "orders over a set amount go to the manager" is cheaper, faster and completely predictable. AI earns its place where the input is messy: free-text emails, scanned documents, requests that need reading before they can be sorted. We use each where it fits, and wherever a mistake would be costly, an AI step hands its result to a person to confirm.
Built into your systems, run by your team
Automations stitched together from separate subscriptions are hard to see and harder to change. We build them into your own web app or internal systems instead, in accounts registered to you, with a log of what ran and when. Your team learns to change a rule, add a step or connect a new service with Claude Code, so the automations keep up with the way you work.
This page is about processes and the connections between systems. If your team mainly needs new tools — dashboards, portals, calculators — or you are planning AI across several departments, these guides cover it:
How it works
Process map
We go through your processes with the people who run them, and pick the ones where automation saves the most time first.
Foundation
We build the foundation and the core automations into your systems: the data flows, the integrations, the scheduled jobs and a log you can read.
Training
Your team learns to maintain and extend the automations with Claude Code, on your own processes.
What comes next
An AI automation review of your team's processes lists what to automate next, so your team knows where to take it from here.
Price
The Team plan is $1,900, fixed and paid upfront. It covers the process map, the foundation with the core automations, training for up to 15 people and an AI automation review of your team's processes. If one person will look after the automations, the Solo plan is $450, fixed, with a 100% refund on request. The services your automations use — hosting, AI subscriptions, any paid integrations — are paid by you, in your own name.
Frequently asked questions
Do we have to replace our CRM or accounting system?
Usually not. We connect the systems you already use. If one of them has no way to exchange data, we tell you during the process map, before anything is built.
Why not just use Zapier or Make?
They are great for a simple link between two services, and if that is all you need, we will say so. When a process has many steps, works with your own data or gets more expensive with every run, building it into your own system is easier to change and to see.
Is it safe to put AI into our processes?
AI steps run in AI services under your company's own accounts, get only the data the step needs, and hand their result to a person wherever a mistake would matter.
Who looks after the automations afterwards?
Your team. That is what the training is for: changing a rule, adding a step or connecting a new service with Claude Code, without waiting for a developer.
How much does it cost?
$1,900 on the Team plan for up to 15 people, or $450 on the Solo plan for one person, fixed and paid upfront.