“Those that fail to learn from history are doomed to repeat it.” Winston Churchill

Introduction

Agentic AI, with its ability to make autonomous decisions and learn dynamically, promises unprecedented opportunities for business innovation. By automating processes with precision and intelligence, it can unlock efficiency, drive cost savings, and elevate customer experiences. 

However, this promise comes amid a backdrop of rapidly evolving technology and the increasing complexity and immaturity of Agentic AI architectures, which demand robust frameworks, scalable infrastructure, and deep expertise to deploy effectively. When applied to the wrong processes—those misaligned with business goals or plagued by inefficiencies—Agentic AI can amplify flaws, leading to wasted resources, degraded performance, and potential reputational harm.

Lucky 7 – A True Story

A major Electric Utility was audited by their Public Service Commission and could not substantiate some of the capital charges that had been applied to their rate base. The CFO asked their Accounting and IT Departments to look into the discrepancies. 

The team started out correctly. They went to the Service Warehouse and observed and documented the end-to-end process of receiving a work order, requisitioning the required materials from the warehouse, Issuing the material to the work order and recording that the material was relieved from inventory. They audited the balance in the inventory system after the fact and the inventory balances were always correctly updated. 

During their review they learned that there was an additional process that required multiple executives to approve the request to setup an account for the capital project before the account number could be provided to the Service Warehouse. They also learned that there were many times when the requisition arrived at the Service Warehouse before the account number was setup and approved. In those cases, the Warehouse workers invoked the “Lucky 7’ rule. They just used a 7-digit number from another work order that had worked previously!

The IT team then recommended that they acquire and implement a Workflow Management System to monitor, manage and speed up this multi-step approval process. 

A consulting company was hired to select and implement the software. Fortunately, the consulting company Project Manager was trained in Lean. As an initial step in their project, the Project Manager met with the CFO of the utility and asked some additional questions about how the capital projects were planned, budgeted and approved. He learned that all of the capital projects were actually approved as part of the annual budgeting process. This meant that the account numbers could be set up as part of the annual planning process and the approval workflow that they were planning to automate was completely unnecessary.

The Bottom Line

The Lucky 7 story had a happy ending. Even though this story happened years before the advent of Agentic AI, there is a valuable lesson here for project teams that are interested in leveraging this powerful technology to automate existing business processes. 

To avoid unnecessarily automating a process that does not need to be automated, or automating a process with a lot of inefficient or unnecessary activities, we strongly advise that you adopt the following guidelines for transforming any existing process with Agentic AI:

• Simplify - Understand and Simplify the Process using Lean Principles
• Automate as many remaining activities as possible using AI and traditional technology approaches 
• Integrate the automated and any required manual activities to create a managed, high quality and efficient end-to-end process

Getting Started

InfiniteIQ has defined an integrated, comprehensive approach to Agentic AI implementations that leverages Lean Thinking and the best thinking from Agile Software Development into our Lean/Agile Solution Delivery Playbook.

We named the first phase of out playbook Concurrent Solution Engineering, modeled after the transformational Manufacturing Product Development Approach Concurrent Engineering which is defined as follows:

• Concurrent Engineering is an approach to product development where design, engineering, and other processes are performed simultaneously rather than sequentially to reduce development time and improve product quality.

An InfiniteIQ Agentic AI Concurrent Solution Engineering project typically has the following deliverables:

• A Lean Analysis of the current process(es),
• The Design of the To-be Lean Process(es),
• The Stories to support development of the Agentic AI Solution,
• The Definition of the Agentic AI Solution Architecture,
• An Implementation Plan; and
• A Business Case

Contact Us

Please contact InfiniteIQ Consulting for more information on our approach to Agentic AI, our Lean/Agile Agentic AI Solution Delivery Playbook or to discuss engaging us for an Agentic AI Concurrent Solution Engineering Project.