I see the same mistake repeated in every large organization I advise: they want to transform with AI before they have optimized with AI. It is like trying to build a second floor on shaky foundations.
Since I have been guiding companies through their AI transformation -- at Renault, Allianz, and Orange Business -- I have formalized a two-step approach. Each step has its own objectives, methods, and success metrics. Confusing the two is a guaranteed path to failure.
Step 1 -- Optimize: do what you already do, but better
AI optimization means taking your current processes and accelerating them. You do not change what you do, you change how you do it. This is where the immediate ROI lies.
A concrete example. A marketing department produces 200 creative briefs per quarter, each taking 4 hours. By integrating an AI assistant into the existing workflow, they cut it down to 1.5 hours per brief. The process stays the same. Throughput is multiplied by 2.5.
Optimization has an underestimated strategic advantage: it trains your teams to work with AI. When you move to transformation, your people already have the reflexes of human-AI collaboration.
Step 2 -- Transform: rethink what you do
Transformation means asking: "if I were designing this process today, knowing that AI exists, would I design it the same way?" The answer is almost always no.
At a major insurance company, optimization had delivered a 30% gain on each step. But the transformation involved completely rethinking the flow: an AI engine analyzes the case upfront, pre-fills 80% of the fields, and routes it to the right expert. From 7 steps down to 3. Lead time divided by 4.
The warning signs
Your teams resist change. Prior optimization builds familiarity. Your data is not ready. Optimization naturally forces data cleanup. You have no baseline. Optimization establishes your reference metrics. Your AI governance does not exist. These questions must be resolved during the optimization phase.
My recommendation: dedicate 3 to 6 months to AI optimization before launching the transformation. Organizations that follow this sequence achieve measurable ROI within the first quarter.
This optimize-then-transform sequence is at the heart of the IAgile approach, the convergence of AI and Agility that I detail in a dedicated article.