AI Agents aren’t the only thing you need
Echoes in the Age of AI: Lessons from ERP Adoption
When I started my career, I worked for a consulting firm that was hired to implement an ERP (Enterprise Resource Planning) system. At the outset of each project, the SME (Subject‑Matter Expert) sat down with the client’s staff to map out the existing processes. Those processes rarely aligned with ERP best practice, so the SME had to recommend changes that would bring them closer to the standard. Naturally, employees resisted because the gap was wide and their daily routines were disrupted. The SME’s response was to introduce small, incremental tweaks that got as close to the standard as possible without causing too much friction.
Implementing those changes became half the battle, the other half being the resistance to them. By letting the client adopt only the minimal adjustments first, the SME eased the transition. Once the organization began using the ERP, the SME would revisit and suggest a second wave of refinements. At that point resistance had diminished because people were already familiar with the system and could see tangible benefits.
I’m sharing this story because I notice the same pattern of pushback today, but with AI.
The Hidden Challenge: Aligning AI Tools with Strategic Workflow Design
The hype around AI Agents has led companies to adopt misconceptions about their capabilities. Leaders often assume AI Agents can solve every organizational challenge, prioritizing tools over processes in their implementation strategies. Currently, AI Agents are deployed as autonomous systems designed to handle specific tasks, with executives expecting these tools to function like “magic” once integrated into workflows. This belief is one of the most significant misunderstandings about AI Agents; particularly that they can operate effectively in isolation from underlying processes.
In reality, these tools function similarly to human employees: if the workflows they interact with are flawed or inefficient, even the most advanced AI Agents will produce subpar outcomes.
Company leaders must shift their focus from merely adopting AI Agent technologies to redesigning and optimizing business processes. Investing in process improvements is critical to ensuring that AI Agents deliver meaningful value rather than exacerbating inefficiencies.
AI Integration That Works: Start with the Rules, Not the Tech
In an ideal world, leaders would streamline workflows before adopting AI tools. But in reality, most dive straight into implementation. Instead, they can start by integrating AI solutions within their current processes, with clear goals and expectations communicated upfront. This initial rollout should function as a diagnostic step, allowing leaders to spot inefficiencies or bottlenecks in real time. As soon as issues arise during the pilot phase, address them by tweaking processes immediately. Once this first iteration is stable, revisit and refine the workflows again. Repeat this cycle (diagnose, adjust, optimize) until every process is tailored for seamless AI collaboration.
