AI, ERP and the operational model in between.

Does your team work in Exact Synergy daily? If so, a large part of the day is spent processing documents, requests, and manual actions. An AI agent working directly in your Synergy environment — selecting data, choosing an assistant, giving instructions in natural language — changes that. Concrete applications for account management, contract management, project managers, finance, and HR. You choose the data, the AI analyzes, and Synergy executes.

The honest answer to "what does AI do in your ERP?" starts with what it doesn't do: a chatbot on top of Exact, AFAS, or Business Central doesn't solve any real problem. The shift happens when AI can perform actions within your ERP via tool calling — where the AI itself chooses which integration to call to complete a task, turning years of integration work into autonomous capability.

The fear in every boardroom: "we don't have the people who can build this." Justified, if you are thinking in terms of programmers and Python scripts. But that's not how it works anymore. Anyone who can type an email can operate an AI agent. The people who know the process direct the agent — not the other way around. How finance, service, sales, and HR employees are using this in practice.

Every company that has existed for more than five years has the same problem: thousands of documents in the ERP, and no way to find anything specific in them. The classic solution is a DMS migration — expensive and often unsuccessful. AI agents are fundamentally changing this: documents stay where they are, but become searchable by content. Concrete use cases for legal, sales, service, and finance teams.

"Why don't we just use ChatGPT?" is a valid question that misses something essential. ChatGPT is brilliant for brainstorming, writing, and summarizing, but it doesn't touch your systems. An AI agent has tools at its disposal: it retrieves data, compares, writes back, and sends actions, within your permissions and with an audit trail. Both are useful. Both are different. When do you use which?

For twenty years now, process digitization has been a near-miss for employees working on the road. The mechanic, the driver, the inspector. What happens: the registration is added in the evening, or not at all. Voice-driven AI agents remove that friction: speak it in, the agent transcribes, links, creates the work order line, and schedules follow-up actions. The paradox: to become smarter with AI, you need better data — and you get that by lowering the barrier to record it.

Most organizations now have a single AI assistant in their ERP for questions and summaries. But that is only the first step. The real shift is a network of specialized agents — procurement, sales, logistics, finance — working together to execute complete processes, while humans retain decision-making authority at the moments that matter.

Using n8n or Make to connect a language model to your ERP sounds appealing: cheap, flexible, and fast. It works fine for simple tasks, but for serious ERP processes, it grinds to a halt after three weeks and unnoticed errors occur. Why domain knowledge, the right partners, and strict security are indispensable when building AI on ERP data.

An AI agent does not access your database directly. It calls a controlled tool — a dataflow or API — that performs the actual queries or write actions. You determine what is possible; the agent operates within those boundaries. The value is not in retrieving data. It is about the agent deciding which action your process requires — and activating the right tool to make it happen. Controlled, traceable, and repeatable.

Outlook AI is an Outlook add-in that allows you to send emails and attachments directly to custom-made AI agents. Where tools like Copilot and ChatGPT stop at summarizing, Outlook AI goes further: through tool calling, the agent triggers data flows that have been running at clients for years, executing actions directly in ERP systems such as Exact Synergy, AFAS, and Business Central. Think of creating documents, registering service requests, or processing expense claims — all from within Outlook, without copy-pasting. The difference: not an assistant that talks, but an assistant that acts.

This demo shows how AI agents handle service tickets from start to finish. Five emails with attachments are forwarded with one click from Outlook to Exact Synergy via Outlook AI. Synergy AI then analyzes the tickets, sets priorities, and performs — with a single instruction — three actions: creating an urgent Synergy request assigned to an engineer, sending a reply email to the customer, and scheduling an Outlook appointment. The crucial difference: these agents do not just summarize, they execute — connected to ERP data via 'tool calling' and pre-built data flows.

Most organizations want to use AI, but are hesitant to connect it to their ERP — and understandably so. Any2info directly addresses that hesitation. With MCP (Model Context Protocol) as its foundation, the Any2info platform acts as a controlled orchestration layer between AI and business systems such as Exact (Globe, Synergy, Online), AFAS, and Business Central. Every connection is secured via OAuth, every AI action runs through predefined data streams, and every interaction is logged and auditable. Critical actions require human approval. Proprietary and external MCP servers can be combined, allowing AI to operate across an organization's entire software landscape in a structured, secure, and transparent manner. The result: not AI that just answers questions, but AI that responsibly handles processes — from analysis to action.