Incoming email
Detect requests, check attachments, prepare draft replies. Your staff review and send, instead of working through every email from scratch.
We automate workflows between CRM, ERP, documents and email with AI where it measurably saves time. Every automation is tested with your team and handed over to your IT with documentation.
In many companies, enquiries, documents and orders arrive by email and are read, checked and transferred into CRM or ERP by hand. These steps repeat, follow fixed rules and cost time every day. That is exactly where AI automation comes in.
Classic process automation reaches its limits as soon as content is unstructured: a construction plan, an email text, a form that keeps changing. AI can read and classify this content. We connect it with your existing systems where it measurably saves time. Where a professional decision is needed, the automation prepares it, for example as a draft reply.
Detect requests, check attachments, prepare draft replies. Your staff review and send, instead of working through every email from scratch.
Invoices, forms and plans contain data that is often still typed in by hand. An AI pipeline reads the fields, matches them against existing data and passes them to the ERP. In AI invoice processing, that is exactly the core: recognise, check, prepare the booking.
When quotes are put together from many product variants, quote automation saves a particularly large amount of time. The project for a timber group further down shows what that looks like.
A RAG system (retrieval-augmented generation) looks up the relevant passages in your own documents for each question and answers on that basis. This makes manuals, data sheets or contracts searchable. The data foundation is decisive: tables must be connected as structured data, not as running text.
We map the workflow with the people who do it every day: which systems, which formats, which exceptions. At the end it is clear which part gets automated and how success is measured.
We build the automation and connect CRM, ERP, document storage or email. Testing uses real cases from your operations, together with your team.
In operation, monitoring shows whether the automation delivers the expected results. You receive hand-off documentation for your team or IT.
The people who work with the automation are trained. They need to be behind the solution, otherwise it goes unused.
Automations run within your Microsoft ecosystem and your company environment. Processing takes place in the EU, and GDPR data protection is part of the architecture from the start. This has been tested where data protection counts most, from social-services organisations to the regulated mid-market.
An international timber group with over 1,000 product variants needed up to two weeks for a quote based on construction plans. The agent system we built reads the plans and assembles a finished quote via the in-house CRM in around ten minutes.
View projectIn the energy advisory of a social-services organisation, funding requests arrived by email. Our workflow analyses the emails, checks the documents and prepares draft replies, for 350 emails in the test phase. It is not yet in live operation, because the processes have grown over years and involve many formats and intake channels.
View projectTechnical delivery is led by Nikolay Krastev. He studied Artificial Intelligence at JKU Linz and has spent several years building software from prototype to production hand-over. Simon Micheler handles the coordination with your team and is there at the hand-over.
Classic automation follows fixed rules and needs structured data. AI automation also processes unstructured content such as emails, PDFs or construction plans and turns it into structured data for CRM or ERP.
Best suited are workflows that repeat often and where content is read, checked and passed on: email enquiries, document checks, invoice processing, quote creation or searching data sheets. In the workflow audit we check whether automation pays off for your workflow.
AI can read invoices, recognise fields such as supplier, amount and date, and match them against existing data. Approval stays with your team. Whether it pays off depends on the volume and variety of your invoices.
RAG stands for retrieval-augmented generation. Before each answer, the system looks up the relevant passages in your own documents and answers only on that basis. For it to work reliably, the data must be connected in a clean, structured way.
To the systems you already use: CRM, ERP, document storage and email. The solutions run within your Microsoft ecosystem and company environment, with processing in the EU.