Spotting AI Opportunities

By the end you will have a longlist of credible AI opportunities for your org.

See it first

The losing move is technology-first: "we bought access to a powerful model, now what can we do with it?" That produces gimmicks. The winning move is problem-first: start from valuable, painful, repetitive work, then ask where AI fits.

The value map

Scan your organization's workflows and look for work that is both high value and AI-amenable. Walk the major functions (sales, ops, support, finance, marketing, product) and list where time and money pool, especially repetitive, language-heavy, or prediction-shaped tasks.

Good fit versus poor fit

AI is not equally suited to everything. The patterns:

Good fit

Repetitive, high-volume tasks
Language-heavy work (drafting, summarizing, classifying)
Prediction from lots of historical data
Work where a strong first draft saves time

Poor fit (today)

One-off tasks with no volume
Work needing guaranteed exactness with no review
Decisions requiring deep real-world judgment and accountability
Anything with no data and no clear pattern

Source ideas from the front line

The best opportunities are rarely visible from the boardroom. The people doing the work know where the tedious, repetitive pain is. Ask them: "What part of your week is repetitive drudgery?" Those answers are an opportunity goldmine, and involving the front line also seeds adoption later.

What you can now do

Approach AI opportunities problem-first instead of technology-first
Use a value map to scan workflows for high-value, AI-amenable work
Recognize patterns of good fit versus poor fit
Source candidate use cases from the front line and build a longlist

Check your understanding

1 / 2

problem-first

What is the recommended way to find AI opportunities?

What's next

A longlist is not a plan. Next, turn it into a prioritized shortlist: Prioritization: Value vs. Feasibility.