AI 𝐄𝐯𝐞𝐫𝐲𝐰𝐡𝐞𝐫𝐞 Is AI 𝐍𝐨𝐰𝐡𝐞𝐫𝐞

Spreading AI across every task feels like progress. Often it just helps you go faster on steps that were never the problem.

BCG studied 1,250 companies in 2025. Only 5% capture value from AI at scale. A full 60% report no material gain at all, despite real spending. (BCG, "The Widening AI Value Gap")

That gap is not a technology failure. It is an 𝐚𝐢𝐦𝐢𝐧𝐠 𝐟𝐚𝐢𝐥𝐮𝐫𝐞.

Most teams sprinkle AI across everything: quicker drafts, quicker analysis, quicker replies. Plenty of motion. Flat results. Accelerating a step that was never the real constraint changes nothing downstream.

Eliyahu Goldratt made this point decades ago in his book The Goal. Every system has one binding constraint, the single point that limits total throughput. Optimize anything except that point, and the system produces exactly as much as before. He called it the 𝐓𝐡𝐞𝐨𝐫𝐲 𝐨𝐟 𝐂𝐨𝐧𝐬𝐭𝐫𝐚𝐢𝐧𝐭𝐬: find the bottleneck, fix the bottleneck, leave the rest alone until it moves.

Growth works the same way. Before you pick an AI use case, 𝐟𝐢𝐧𝐝 𝐰𝐡𝐞𝐫𝐞 𝐫𝐞𝐯𝐞𝐧𝐮𝐞 𝐚𝐜𝐭𝐮𝐚𝐥𝐥𝐲 𝐥𝐞𝐚𝐤𝐬:

→ Acquisition: too few of the right people ever hear of you?
→ Conversion: they arrive, then leave without buying?
→ Retention: they buy once and never come back?
→ Pricing: you sell plenty but capture too little?

𝐀𝐢𝐦 𝐀𝐈 𝐚𝐭 𝐭𝐡𝐚𝐭 𝐥𝐞𝐚𝐤. Nowhere else first.

Spotify is a clean example. Its catalogue is near-identical to Apple Music and Amazon Music. The product is not the constraint. Staying is. When everyone has the same songs, the leak is churn.

So Spotify aimed its AI straight at retention: Discover Weekly, the AI DJ, Wrapped, each designed to make leaving feel like a genuine loss. By its own reporting, listeners who lean on these features stay longer and churn less. Same songs as rivals. 𝐀𝐈 𝐩𝐨𝐢𝐧𝐭𝐞𝐝 𝐚𝐭 𝐭𝐡𝐞 𝐨𝐧𝐞 𝐬𝐭𝐞𝐩 that decides the subscription.

Here is the encouraging part. 𝐀 𝐜𝐨𝐧𝐬𝐭𝐫𝐚𝐢𝐧𝐭 𝐢𝐬 𝐭𝐡𝐞 𝐦𝐨𝐬𝐭 𝐠𝐞𝐧𝐞𝐫𝐨𝐮𝐬 𝐭𝐡𝐢𝐧𝐠 in your business. It pinpoints precisely where one concentrated effort generates the most return. Identify it, and 𝐥𝐚𝐬𝐞𝐫-𝐟𝐨𝐜𝐮𝐬 𝐲𝐨𝐮𝐫 𝐀𝐈 𝐢𝐧𝐢𝐭𝐢𝐚𝐭𝐢𝐯𝐞𝐬 on it.

A 30-minute exercise for your next team meeting:

Name the single place you lose the most revenue: acquisition, conversion, retention, or pricing. Then ask of every live AI pilot, does this touch that step, or a comfortable one beside it?

Most pilots will fail that test. That is not bad news. That is your roadmap.

Where does your business actually leak, and is your AI anywhere near it?

#AIStrategy #Growth #TheoryOfConstraints #DigitalTransformation #Bottleneck

Contact

bruno.gentil@sherpaconsultingasia.com

© 2026. All rights reserved.