Accenture Calls for Knowledge Graphs to Bolster AI Trust

Investment in AI technologies peaked in 2024, driven by sectors aiming to overhaul operations through machine learning and automation systems.

Yet, the journey from initial pilot projects to full-scale production has highlighted numerous challenges.

Many enterprises face difficulty in delivering tangible returns on AI investments, while technical intricacies and integration issues have impeded adoption progress.

This disparity between AI's potential capabilities and its real-world business applications has become a central concern for technology executives.

However, Accenture identifies a potential solution with the application of knowledge graphs.

These graphs organise data as networks of entities and their relationships, enabling AI to access contextual information and ensure precision when addressing complex queries.
Responding to these challenges, major consulting firms, software vendors and technology providers now focus on integration services and enforce data governance frameworks to support sustainable AI deployment.

“If your AI strategy doesn’t start with data governance and semantic integration, you’re skipping a critical step,” says Kristalyn Warren Mumaw, Accenture’s AI Evangelist.

The role of physical AI

The concept of “Physical AI” gained visibility at the Consumer Electronics Show, epitomising the integration of foundational AI models with robotics.
This innovation facilitates autonomous functions through natural language orders, advancing past traditional robotic process automation (RPA) systems that required predefined protocols.

Chris McDivitt, Global Solution Lead for Autonomous Supply Chain at Accenture, illustrates the shift with a practical example: “You can ask a robot to ‘find all the green widgets in Aisle 3,’” he explains.

The system can execute this task even when physical locations change or when “green” refers to product names rather than colours.

Physical AI allows higher-level directives to be executed by combining model planning capabilities with robotic mobility.

Industries like supply chain and manufacturing present ripe opportunities for transformation. This technology redefines workflow dynamics, presenting new avenues for human-machine collaboration.

Trust as a constraint for AI expansion

Chris' analysis puts trust forward as the main factor hindering the expansion of AI applications.

The level of user confidence varies based on the application context and the consequences of errors.

“Would you trust ChatGPT to tell you why the sky is blue? How about diagnosing a medical condition? What about explaining why you're short 30 units for a critical outbound order?” Chris illustrates.

Confidence in AI largely hinges on the system's accuracy and the implications of possible mistakes, he adds.
Users are typically more accepting of generic knowledge outputs than in scenarios demanding expert knowledge or real-time enterprise data.

Current interactions with AI technologies typically entail users directing chatbots or AI systems where they hold decision-making roles.

Chris anticipates a shift towards strategic operations where humans will direct AI systems for complicated task resolutions.

This progress from operational to strategic roles demands technological advancements, including improvements in large language models (LLMs) and agent frameworks.

This shift necessitates robust policies alongside increased trust and verifiable evidence that AI agents deliver appropriate results.

Chris asserts that these advancements will redefine human-AI collaborations within enterprises, with a future workforce increasingly viewing AI as a dynamic partner rather than just a tool.

Enabling AI with Digital Brain

To enhance accuracy and trust, Accenture suggests embedding AI agents in a “Digital Brain” environment.

This framework aims to provide AI systems with access to enterprise knowledge and up-to-date data.

Recognising the limitations of LLMs in holding specialised knowledge and current data essential for enterprise functions, Accenture posits this framework, specifically within supply chain and manufacturing sectors where precision and trust are crucial for success.

Chris says: “To increase accuracy, usefulness and subsequently trust of AI, we need to put AI agents in the right environment to operate in,” he says.

The Digital Brain concept represents Accenture’s approach to tackling enterprise AI deployment challenges, promising operational success where reliability is key.

“If you want to learn more about building secure, autonomous supply chains that enhance resilience and efficiency, let’s have a conversation,” he concludes.