At the Gestal­ter­kreis Invoice & Agentic AI, user companies discuss concrete appli­ca­tion areas of auto­no­mous AI in the Purchase-to-Pay process (P2P) together with insiders.

Insiders, a tech­no­logy-leading provider of software for Intel­li­gent Auto­ma­tion (IA), discussed concrete use cases for auto­no­mous AI agents in the Purchase-to-Pay process together with customers as part of its „Invoice & Agentic AI“ Gestal­ter­kreis. The two-day event focused on the question of how new approa­ches to Agentic AI can be meaningfully inte­grated into existing business processes.

Col­la­bo­ra­tion in Practice

Agentic AI is currently regarded as the next promising stage of deve­lo­p­ment in the use of arti­fi­cial intel­li­gence in companies. In the future, auto­no­mous systems are expected not only to assist with indi­vi­dual tasks, but to inde­pendently take over process steps and prepare decisions.

A brief accom­panying survey among the Gestal­ter­kreis par­ti­ci­pants paints a clear picture: there is strong demand for further auto­ma­tion of manual process steps. At the same time, only a few companies have so far deployed AI agents in their ope­ra­tional workflows. Accor­dingly, there is con­siderable interest in concrete and practical use case scenarios.

„I was initially a little skeptical and wondered what these agents could actually do from a technical stand­point,“ reports Patrick Heckenstaller, Head of Invoice Receipt at MAN Truck & Bus SE. „But at the Gestal­ter­kreis it was really inte­res­ting to see how far the deve­lo­p­ment has already come.“ Sebastian Bernard, European Sales, Opti­miza­tion of Order Manage­ment Processes at Bosch Rexroth AG, adds: „Before the joint workshop and the pre­sen­ta­tions, I didn’t have enough concrete ideas. It was really during the workshop that you got truly immersed — and came away with com­ple­tely new ideas.“

At the Gestal­ter­kreis, Insiders presented the current deve­lo­p­ment status and the roadmap of its solutions for automated invoice pro­ces­sing. In sub­se­quent workshops, par­ti­ci­pants discussed potential use cases for auto­no­mous systems across the entire P2P process together with product managers and AI experts.

The focus was par­ti­cu­larly on process steps that still require manual inter­ven­tion today — such as excep­tions, unclear documents, or more complex review processes. A key outcome of the dis­cus­sions: AI agents deliver their greatest value where they meaningfully com­ple­ment existing auto­ma­tion and reduce manual effort.

„The agent that was presented to us there was very, very impres­sive to me. I could see at any time: where am I right now? Has my delivery arrived yet? Am I still under-delivered? And in the end I could trigger actions — for example, auto­ma­ti­cally notifying the supplier,“ reports another par­ti­ci­pant respon­sible for managing accounts payable and the invoice review process in their company.

„Agentic AI is being widely discussed right now. However, what matters is not the tech­no­logy alone, but its concrete value within the business process,“ says Jan Oevermann, Head of Sales Unit Industry at Insiders Tech­no­lo­gies. „That is why we develop cor­re­spon­ding solutions not in isolation, but together with the companies that operate these processes on a daily basis.“

Insiders‘ Path to Agentic AI

Formats such as the Gestal­ter­kreis are a central component of this customer-oriented approach to driving inno­va­tion. Customers, partners, and product teams work together directly, discuss real-world requi­re­ments, and provide feedback on new approa­ches. The outcomes feed directly into the further deve­lo­p­ment of the solutions.

On the tech­no­logy side, Insiders relies on a platform archi­tec­ture that combines various AI methods and can be flexibly inte­grated into business processes. These include the AI inte­gra­tion platform OvAItion, the auto­ma­tion platform OmnIA, and the no-code service builder GenerAItor.

OvAItion enables companies to run different AI models in parallel — from classical machine and deep learning methods to current large language models from various providers. A distinc­tive feature is that Insiders also offers a private LLM in which the data never leaves the Insiders Cloud. This allows the most suitable models to be used for different tasks without being per­ma­nently tied to indi­vi­dual tech­no­lo­gies.

OmnIA handles the orchestra­tion of processes. The platform connects indi­vi­dual AI services with real business workflows, controls processes, and inte­grates human review and decision-making steps where needed. This allows auto­no­mous functions to be incre­men­tally inte­grated into existing processes without losing trans­pa­rency or control.

With GenerAItor, Insiders also provides a no-code tool that enables spe­cia­list depart­ments to configure their own AI services. Clas­si­fi­ca­tion and extra­c­tion functions can be created without any training effort and inte­grated directly into automated workflows.

From Feedback to Practical Imple­men­ta­tion

The dis­cus­sions, feedback, and ideas from customers flow directly into the deve­lo­p­ment work at Insiders Tech­no­lo­gies. This gives rise to practical, scalable AI solutions that address real process chal­lenges and con­ti­nuously advance the path toward auto­no­mous agents in the P2P process.

The com­bi­na­tion of close col­la­bo­ra­tion with customers, platform archi­tec­ture, and flexible AI inte­gra­tion allows inno­va­tions in Agentic AI to create targeted added value for companies.