AI Agents in P2P: Phoenix Contact and Insiders Demonstrate the Verification Agent Live

    Published: November 19, 2026

    Key Takeaways

    At SSOW DACH 2026 in Berlin, Insiders Technologies and Phoenix Contact demonstrated how AI agents work within the purchase-to-pay (P2P) process. The focus was on the Verification Agent: It independently reconciles order confirmations with the corresponding purchase orders, retrieves missing information from the ERP system and supplier master data, and identifies discrepancies. It forwards unclear cases to the purchasing department with a documented explanation. Other areas of application include the account assignment of incoming invoices and responding to supplier inquiries. The system is based on the OmnIA automation platform. The more business context the AI can utilize, the higher the dark processing rate. Control remains with humans via approval rules.

    Insiders Technologies, a leading technology provider of Intelligent Automation (IA) software and part of the Proalpha Group, will present new approaches to the use of Agentic AI in the purchase-to-pay process (P2P). The focus is on intelligent AI agents that go beyond traditional document processing and independently handle tasks within defined business processes. Together with its client Phoenix Contact, Insiders Technologies demonstrated in a live workshop how companies can take the next step from Intelligent Document Processing (IDP) to end-to-end, agent-based process automation.

    In manufacturing and retail, high document volumes, complex supplier relationships, and time-consuming reconciliation processes are part of day-to-day operations. Although many companies already use AI for the automated capture and processing of invoices, purchase orders, or order confirmations, downstream process steps often still require manual intervention.

    Agentic AI aims to close this gap. Traditional solutions for Intelligent Document Processing (IDP) classify documents, extract information, and make it available to downstream systems. AI agents go a step further by analyzing business transactions in context, independently retrieving additional information from connected systems, making rule-based decisions, and triggering further process steps. Intelligent document processing is thus evolving from automated data capture to an active component of end-to-end business processes.

    AI agents are increasingly taking on tasks independently within the purchase-to-pay process

    In the P2P process, various AI agents support individual process steps and enable increasingly autonomous processing. In addition to verifying incoming documents such as order confirmations or invoices, they can also be used to automatically respond to supplier inquiries and assign accounting codes to incoming invoices. For example, AI agents can suggest appropriate general ledger accounts and cost centers or provide information on the processing status of an invoice. The goals are a higher rate of automated processing, shorter turnaround times, and a noticeable reduction in the workload for business departments.

    Live Demonstration with Phoenix Contact: The Verification Agent in Action

    Insiders Technologies and Phoenix Contact demonstrated how to successfully transition from individual AI applications to agent-based process automation in their joint workshop “Your P2P Process Is Hiring. Agents Wanted.” In addition to the technological possibilities, the experts also highlighted the organizational and procedural requirements for the successful deployment of Agentic AI in industrial companies.

    The focus was on a live demonstration of the Verification Agent for order confirmations. This agent processes incoming documents such as order confirmations or invoices, independently reconciles them with the corresponding purchase orders, and identifies discrepancies. To do so, it draws on relevant data from ERP systems, supplier master data, and other process sources.

    Unlike a purely rule-based check, the agent can independently research additional information, verify its results, and derive further processing steps from them. It involves employees only in cases of uncertainty or defined exceptions. It transparently documents the reasons for handing off the task.

    “With Agentic AI, we’re reaching a new level of automation. AI no longer just understands documents—it independently handles tasks within complex business processes,” explains Andree Albrecht, CEO of Insiders Technologies. “There is enormous potential here, particularly in the purchase-to-pay process: fewer manual reconciliations, faster decisions, and higher automation rates. Together with Phoenix Contact, we are demonstrating how these technological possibilities can be translated into concrete use cases for the industry.”

    Business Context Determines the Level of Automation

    The technological foundation for the AI agents is Insiders Technologies’ AI-based automation platform, OmnIA. It combines intelligent document processing with modern AI technologies, integrated enterprise systems, and automated workflows.

    Internal measurements conducted with OmnIA demonstrate just how important the business context is for automation. According to these measurements, more invoices can be processed entirely without manual intervention when language models also have access to master data, information from previous transactions, process data, and work instructions. The so-called “dark processing rate” therefore increases when more relevant information is available to the AI. AI agents use this information not only to process documents more accurately but also to independently take on defined tasks and trigger further process steps.

    Humans retain control throughout the process. Clear approval rules, traceable decisions, and the targeted involvement of employees in exceptional cases ensure the necessary transparency and security.