What Happens After an AI Opportunity Assessment

You have the findings from an AI Opportunity Assessment. The report identifies two or three high-leverage process areas, a rough effort estimate for each, and a suggested starting point. Now what? This post walks through what typically comes next: how to read the findings, which path to take, and what tends to stall even when the assessment itself was solid.

Read the findings as priorities, not a to-do list

An assessment typically surfaces five to ten potential AI applications across your business. Treating that as a backlog leads to a familiar problem: nothing actually gets done. Every item competes with every other item, and the result is a document that gets reviewed in meetings but never acted on.

The findings are a starting point for a conversation, not a deployment plan. The first question is: of everything identified, which one process, if you changed only that, would have the clearest impact on how the business runs? Start there. Leave the rest on the list for later.

The criteria for picking first: high frequency (it happens often enough to feel the change), repetitive (the task is similar enough each time that AI can help), and someone on the team is motivated to change it. That last one matters more than most people expect.

Three paths people typically take

Build the first workflow internally, using the assessment as the spec. Works well when one person on the team has enough curiosity to experiment, and the process is clear enough to describe. The assessment gives them the brief; they figure out the implementation. Slower to start but builds internal capability from day one.

Run a focused workshop on the highest-priority process. A structured session where the team works through the workflow together with a facilitator. Good for teams that want to move quickly on one thing and need someone to hold the space for it. The output is a working version of the workflow, not just a plan.

Start a transformation program that works through the priority list over three to six months. The right fit when multiple people need to change habits, not just one process needs automating. The assessment becomes the roadmap for a longer engagement that moves through the list in order.

What tends to stall

Three patterns come up repeatedly after assessments, regardless of how good the findings were.

• Waiting for the right tool. Claude works for most of what assessments identify. The delay while evaluating ten tools is usually longer than the time it would have taken to just start.
• Trying to automate a process that isn't stable yet. If the process changes depending on who does it or what day it is, AI makes it faster but not better. Stabilize the process first, then automate.
• Assigning implementation to someone who wasn't in the assessment conversation. The person who understands the context needs to be involved in building the first version. Handing off a document to someone who wasn't there rarely produces something the team actually uses.

The first 30 days

One workflow live. Documented well enough that the person it was designed for can run it independently. Used at least five times in the first month. That is enough. The goal in the first 30 days isn't transformation, it's one working example that makes the case for the next one.

The teams that move fastest after an assessment are the ones that resist the urge to implement everything at once. One thing done well creates momentum. Ten things started creates noise.

If you're deciding whether an assessment is the right first step, this post on Claude for small business gives a broader picture of what working with AI looks like before you get to the assessment stage.