95% see no results with AI (and why that's normal)
TL;DR
- 95% of organizations see no AI returns because they’re in the J-curve: first you go down, then you go up
- The initial drop is normal: learning curve, system integration, broken expectations
- Companies abandon ship right before it starts working
- Solution: realistic expectations, 12-month metrics, start small
The uncomfortable truth
95% of organizations that have adopted AI see no measurable returns yet.
Read that again. Ninety-five percent.
And yet companies keep investing billions. Are they stupid? Not exactly.
The J-curve of productivity
When you adopt AI, your productivity follows this shape:
Productivity
│
│ ╭────────────
│ ╱
────┼─╯
│╲
│ ╰──╮
│ │
└────┴──────────── Time
↑ ↑
Drop Recovery
First you go down, then you go up.
Why?
- Learning curve: Your team doesn’t know how to use the tools
- Integration: Systems don’t talk to each other
- Broken expectations: “AI does everything” → it doesn’t
- New processes: You have to change how you work
Only after crossing that valley do you start seeing real benefits.
What I see day to day
I use AI for coding. My model usage graph in 2025 looks like a rainbow: Claude, GPT, Gemini, back to Claude.
Why so many changes? Because at first, each new model made me slower:
- Learning its limitations
- Discovering which prompts work
- Finding where it lies
- Adjusting my workflow
Now I’m more productive than before using AI. But the first months were frustrating.
The corporate mistake
Companies buy AI expecting:
- Immediate results
- ROI in the first quarter
- For it to “pay for itself”
And when the initial drop of the J-curve hits, they panic:
- “It doesn’t work”
- “We wasted our money”
- “Let’s go back to the old way”
And they abandon ship right before it starts working.
How to survive the J-curve
1. Realistic expectations
The first quarter you’ll be LESS productive. Accept it.
2. Long-term metrics
Don’t measure ROI in 3 months. Measure in 12.
3. Start small
One team. One process. One specific problem.
Not “let’s put AI in the whole company.”
4. Tolerate the initial chaos
Your team will complain. They’ll say it was better before. That’s normal.
5. Document what works
When something goes well, write it down. You’ll need those wins to justify the investment while you’re in the valley.
Conclusion
95% see no results because they’re in the middle of the J-curve and don’t know it.
Some will abandon ship. Others will persist.
Those who persist with strategy (not blind faith) will be the 5% who see returns.
AI isn’t magic. It’s a tool that requires time investment before returning value.
If your company is in the J-curve valley, it’s not that AI doesn’t work. You just haven’t reached the other side yet.
You might also like
The AI bubble: 7 trillion looking for returns
Who wins, who loses, and why you should care. Analysis of massive AI investment and its bubble signals.
36% of Spanish companies now use AI: what it means for you
AI adoption data in Spain in 2025 and what it means for those who work with data. A view from Andalusia.
OpenAI launches Prism: the end of the generic chatbot has begun
A free workspace with GPT-5.2 for writing papers in LaTeX. But the real news isn't Prism: it's what it represents about where AI is heading.