Infor Research: Closing the Agentic AI Execution Gap

While business confidence in AI is stronger than ever, for many enterprises the dive into the deep end of execution is riddled with barriers. 

New research from Infor says as much, with data showing that more than half of the businesses are struggling to scale AI. 

Aimed at closing this widening gap between ambition and value realised, industry cloud provider Infor has introduced new capabilities across its Velocity Suite and Infor Agentic Orchestrator delivering niche, industry specific solutions. 

“At Infor, agentic AI isn’t a feature we bolted on. It’s the culmination of two decades of deliberate foundation building,” says Kevin Samuelson, CEO of Infor. 

“Our industry specific platforms, multi-tenant architecture and deep process intelligence give our agents a level of contextual precision that generic AI simply cannot replicate.

“A purchasing agent at a healthcare provider and one at a discrete manufacturer aren’t the same agent, they shouldn’t be.”

The emerging international AI execution gap

The study, conducted across the US, UK, France and Germany, shows that making the technology work at scale is a major operational challenge.

Around 70% of businesses in the US and 74% in the UK report that they have the capability to manage AI implementation, yet this readiness does not translate into reality, with structural barriers blocking effective execution. 

Legacy systems, risks associated with inconsistent governance and fragmented data that pulls down the effectiveness of AI are among these barriers. 

A point to note is that this execution gap appears in distinct ways in different industries, for instance, legacy infrastructure plagues manufacturing the most, governance and compliance hinders healthcare while distributors are burdened by fragmented data across various supply chains.

Embedding generic AI into these workflows, without accounting for their underlying operational reality, produces unscalable outcomes.

Closing this execution gap is hence pivotal to realising value from AI which requires deep domain specific industry knowledge. 

“That specificity is what allows us to clearly articulate the ROI and deliver on it,” Kevin notes.

“We’re not selling automation for its own sake. We’re selling measurable outcomes for the industries by meeting our customers where they are with AI and providing a clear, simple and efficient path to where they want to be.”

Data security concern and agent distrust 

A recurring theme and the most prominent barrier to AI implementation, is concerns of data security, the Infor study reveals. 

About a third (34%) of the businesses in US and Germany, 32% in France and 45% in the UK all point to variants of data sovereignty, security and privacy concerns as a factor that prevented them from advancing their AI strategies. 

As AI models are only as good as the data that they are trained on, having fragmented data on multiple systems could result in AI models that struggle to deliver positive outcomes. 

With only 25% of the businesses saying that their data is mature enough to support AI that can be reliable, this gap could not be more crucial. 

Secure, responsible and compliant deployment of AI agents is a real necessity in the current security climate. 

The lack of internal AI talent (25%), unclear ROI (23%) and high cost of AI (23%) are some other barriers to implementation.

Orchestration, interoperability and observability

The ability for AI to perform tasks autonomously was ranked by 32% as a factor for long term AI success. 

The importance of security here is paramount, and within the Infor Industry Cloud Platform, Infor’s Agentic Orchestrator facilitates a “trusted, transparent infrastructure layer that enables Industry AI Agents to move from isolated tasks to coordinated workflows”.

The updated capabilities released by the company acts across three areas – orchestration, interoperability and observability. 

The platform offers multi agent operation with supervisor agents and specialised task agents. Here, the supervisor agents are pre-trained to flag anomalies which the human in the loop can then can act upon. 

Infor Agentic Operator frees up worries of AI integration, with standardised Model Context Protocol (MCP) offering secure actions and access of data. 

Observability is critical in agentic use and Infor’s latest visibility features offer 3 particular capabilities – inline thoughts, evaluation frameworks and focus mode – allowing complete user control and oversight. 

Speaking to the success of Infor’s approach to agentic implementation, Mickey North Rizza, Group Vice-President of Enterprise Software for IDC notes: “It is very clear that Infor’s clients are finding sustained economic value with their path to the agentic enterprise and they love the journey with Infor.”