95% of AI investment fails today

Why you’re getting this: This is a friend-first newsletter. Faangboss is a weave of three essential threads: the craft of design, the art of influence, and the journey of growth in UX. I poke at connective tissues, drawing from my 20+ years of launching products and mentoring designers in tech NYC.

Long time with no updates! I had a baby girl late July and have been knee deep into motherhood. My three week maternity leave starts after Labor Day, I'll have more time to resume writing content. While I have a lot of personal updates to share, I'll leave it for upcoming publications.

MIT published a new research report called "The GenAI Divide" and this is the summary and my hot takes.

$40 Billion reality check, high adoption, zero impact

Enterprises are pouring $30-40 billion into Gen AI, MIT's NANDA study shows 95% of these organizations see zero return of investments. This is based on 52 interviews and 300+ AI implementation analysis. This sparks the debate about whether this is an AI limitation or poor strategy. Are we forcing new tech onto old process, and it's not good at corner cases?

80% of organizations use AI tools and 5% custom AI tools reach production. For a big tech corporation to use AI, they want to remove human oversight, but the current AI requires human validation for critical tasks. Why use AI in the first place?

The only logical answer is they want to reduce headcount, and then hire humans back at a lower salary. Amazon announced reduction of headcount by 2026.

Tech and media have meaningful disruption, but other industries have yet to see transformations from these pilot activities. AI excels in high volume, transactional work and struggle with nuanced, context-dependent tasks.

The core problems are memory and learning. These AI are sophisticated search engines that reset after sessions. Lacking feedback integration, contextual memory, and adaptation to workflows.

The ROI is replacing outsourced functions, not core employees. The only thing any AI can reliably be used for is to sucker investors out of more money. Eventually, AI companies will stop selling their services as a loss leader, and then the AI 'cost-savings' will rapidly disappear.

Any thoughts?

Why would you want to subscribe? Well, why wouldn't you? If you like this content, consider supporting me, your independent writer with a coffee. ☕️

Have a great week ahead,