Processes · AI Agents
AI agents take over recurring processes between your systems, with clear rules, approvals and a log that makes every action traceable.
Definition
An AI agent is a digital case worker for a recurring process: it reads inputs (an email, a form, a record from the ERP), makes a decision or drafts an outcome following defined rules, and writes the result into your systems. Unlike a chatbot, it doesn't hold an open conversation; it completes a concrete work step that a human today performs manually across several programs.
Typical use cases
The agent reads incoming inquiries, matches them against your price list and creates a draft quote for approval.
Overdue invoices are detected, reminders created and sent by your rules, with a full log.
Incoming mail is categorized, urgent cases flagged, routine questions get a reply draft.
What today travels via copy-paste between ERP, CRM and spreadsheets, the agent transfers automatically and logs every step.
Weekly reports, revenue overviews, status updates: compiled and sent without anyone having to remember.
Technology
automation software is an open-source automation platform with over 400 ready-made connectors to ERP, CRM, email, calendar and accounting systems. It's self-hostable: workflows and data can run on your own infrastructure instead of a third-party SaaS platform. The result for you: no vendor lock-in, predictable cost, and the option to take operation in-house if you ever want to.
Guardrails
Critical steps (sending, payment, deletion) only run after confirmation by a human.
Every action the agent takes is traceable: who, what, when, from which source.
The agent does exactly what the process defines, nothing more. Unclear cases go to a human.
An AI agent is a digital case worker for a recurring process: it reads inputs (email, form, system data), makes a decision or drafts an outcome following defined rules, and writes the result into your systems. Unlike a chatbot, it doesn't hold open-ended conversations, it completes a concrete work step.
A chatbot answers questions in dialogue. An AI agent works through a process, with or without dialogue: create a quote, check an invoice, transfer data. A chatbot can be a tool an agent uses internally, but the agent is geared toward an outcome, not a conversation.
Processes that recur regularly, follow clear rules, and are currently done manually between systems: quotes, invoicing runs, email sorting, reports, data transfer. One-off or highly subjective decisions suit it less.
Setup typically 1,500–3,500 € per process, operation from about 150–400 € per month depending on scope and system integrations. The exact figure depends on how many systems need connecting, that's what the potential assessment clarifies.
automation software is an open-source automation platform with over 400 ready-made connectors to common systems (ERP, CRM, email, calendar, accounting). It's self-hostable, meaning workflows and data can run on your own infrastructure. We build on it instead of proprietary black-box tools, which avoids vendor lock-in and keeps costs predictable.
automation software itself is a tool, GDPR compliance depends on how it's operated. Self-hosted on EU servers or your own infrastructure, automation software is as data-protection-compliant as any other software you control yourself. We sign the necessary data processing agreements for connected third-party systems.
It doesn't guess. Cases outside the defined rules are flagged and handed to a human instead of a wrong decision being made. That's a deliberate guardrail, not an exception.
30 minutes, free potential assessment. We'll tell you honestly whether an agent is worth it, and if so, which one first.
Book intro call