Here’s a controversial thought: If we stopped all AI development right now—no GPT-6, no Gemini 4, no new models ever again—we’d still have enough power to change society in a thousand positive ways.
Yet leaders keep telling us AI “can’t do enough yet.” The real issue is a lack of imagination.
We’re drowning in features and benchmarks but starving for vision—mainly because the sector is doing a terrible job showcasing what’s possible. We lack a Steve Jobs of AI to tell us what it’s actually for.
🎧 Tired of reading all day? Me too. Listen to the 10-minute podcast version for extra stories—from a Rhode Island magician helping his brother communicate to Rwandan teachers reaching thousands of students via text message—and bookmark this article for the data and charts.
🎯 The Vision Crisis: Benchmarks Without Bicycles
I’ve briefed over 18,000 people, and they keep telling me the stories told by large enterprise software companies are dull and uninspiring.
First, the tech‑jargon pitch: benchmark charts and feature lists. Then the tiny stories: AI assistants that help you draft emails, clean up meeting notes, or polish slide decks — useful, but around the edges of the real work, not at the core of how we serve patients, repair equipment, teach kids, or run factories.
And on the other side, we get big, far‑off scenarios about AI “changing everything” that feel optimized for investors and headlines, not for what people actually do on Monday morning.
The result? A massive gap between what AI can do and what organizations are doing with it.
We never start an AI Transformation project by discussing prompt engineering or process maps. Instead, we show the team exactly what is possible today without writing a single line of code; we inspire them.
| What the Data Shows | Source |
|---|---|
| 88% of organizations use AI in at least one function | McKinsey 2025, AI4SP 2025 |
| Only 33% to 39% report any enterprise-level EBIT impact | McKinsey 2025, AI4SP 2025 |
| Only 6% to 10% qualify as “AI high performers” | McKinsey 2025, AI4SP 2025 |
| 95% of Top-down enterprise AI pilots fail to deliver measurable P&L impact | MIT NANDA Report, AI4SP Report |
| 42% of companies abandoned most AI initiatives before reaching production | S&P Global |
The technology isn’t the problem. MIT researchers state that “the gap (in results) isn’t driven by model quality or regulation. It is determined by approach.”
Most of us remain stuck in first gear. We apply AI to basic tasks because that is what the leading voices discuss. In 2025, we worked with seven global enterprises to change this. We engaged their frontline teams and saw strong results:
| Agents Created | Tasks Completed | Financial Impact |
|---|---|---|
| 3,800 | 4.2 million | $47M |
The secret to our success in 2025 was, paraphrasing Steve Jobs, that we gave people bicycles for their minds.
🧮 The Maker Gap: Inspiring people to get past the 2%
That $47M impact didn’t come from everyone building everything or from a single top-down initiative. It came from identifying and empowering the critical 2% of Makers within the frontline. The 2% of makers is expanding into double digits as more people learn from what others have done; those 3,800 agents sparked the imagination of the masses.
From our data across 250M+ touchpoints, the AI user distribution looks like this:
| User Type | Percentage | Behavior | Outcome |
|---|---|---|---|
| Passive Users | 43% | Occasional, basic queries | Minimal impact |
| Active Users | 55% | Regular use, standard features | Moderate productivity gains |
| Makers | 2% | Build custom agents, curate knowledge | Transformational results |
When teams have access to simple tools, the gap between user and maker isn’t technical. It is imaginative. The only barrier is the decision to stop being a consumer and start being a creator.
Why frontline workers are your secret weapon: They know where the friction lives. They speak in natural, conversational language, which is precisely how AI works best. The people closest to the work don’t need to learn “prompt engineering”; they need the right, easy-to-use tools and to describe their problems the way they already do.
One of our favorite examples: Agent Luke W., built by a newly hired engineer at a large construction firm. He created an “Apprentice Coach” that let junior field technicians tap into the collective wisdom of senior experts: repair histories, troubleshooting patterns, and institutional knowledge, without waiting for callbacks or scheduling site visits.
Agent Luke generated $5M in incremental revenue by enabling junior hires to complete jobs independently, nearly doubling the field support team’s capacity.
No code written. No IT development cycles required, other than pre-approving a set of AI tools that passed security requirements. Just a frontline worker solving a problem that leadership didn’t even know existed.
🎧 Hear Agent Luke W’s story and other examples of frontline-led transformation in the podcast version.
💰 The Measurement Crisis: Why “Hours Saved” Misses the Iceberg
CFOs are staring at dashboards, asking: “How do I capture all the value? How do I put this on a spreadsheet?”
Here’s the issue: We’re using industrial-age metrics to measure a cognitive revolution. The obsession with “hours saved” is trivializing this technology.
| What Organizations Track | What Actually Matters |
|---|---|
| Hours saved | Problems solved and Tasks completed |
| Workflows automated | Business outcomes achieved |
| Adoption rates | Value created |
Our data shows that “hours saved” translates into actual profit only 30% of the time. If you save an hour but fill it with nothing, you gain nothing. Often, those hours get reinvested in hard-to-measure outcomes: better-quality work, innovation, or a much-needed mental break.
The Value Iceberg
We are making a mistake by measuring only what we can easily see. MIT researchers recently introduced the “Iceberg Index,” which shows that while visible AI adoption in tech is just the “tip” (2.2% of wage value), the massive “hidden” exposure is in administrative and professional services (11.7%).
We can apply this same “Iceberg” logic to how we measure ROI.
If you only measure the surface (time saved), you miss the massive value below the waterline, where the real business transformation happens.
Example: A car factory automates quality control. The old metric says “we saved labor hours.” But the ripple effects—fewer defects, smoother logistics, fewer warranty claims—don’t show up on the workforce spreadsheet. If you are only measuring time, you are missing the iceberg.
📈 What High Performers Do Differently
In 2025, we found some patterns among the 6% to 10% achieving significant enterprise value from AI:
| Practice | High Performers vs. Others |
|---|---|
| Intend to use AI for transformative change | 3x more likely |
| Have fundamentally redesigned workflows, from a bottom-up approach; the SMES are the ones driving the changes | 3x more likely |
| Set growth/innovation as objectives (not just efficiency) | Significantly more likely |
| Commit >20% of digital budget to AI | More than 1/3 of high performers |
| Have scaled AI across the organization | 75% vs. 33% |
Focus on orchestrated mini agents versus top-down Large agents in cloud deployments shows 4x higher success rate
High performers think beyond incremental efficiency gains. They treat AI as a catalyst to transform their organizations, redesigning workflows and accelerating innovation.
🔮 One More Thing: Stop Waiting for GPT-6
If you’re not getting results from AI, the next version won’t fix it. GPT-6, Gemini 4, Claude 5…none of them will solve a vision problem. You’d better focus on unleashing and guiding your people’s imagination and redesigning your structures.
While we lack a Steve Jobs of AI, we have thousands of them—in hospitals, factories, schools, and government offices. People who just need a nudge of inspiration to see what’s possible with the tools we already have and paint that picture for others.
The technology is ready. The question is: Are we imaginative enough to use it?
🚀 Ready to Take Action?
- AI Management Certification – for enterprise groups of 15-20 individuals.
- AI Compass – assess grassroots AI maturity, and opportunities to channel shadow AI:
- Workshops & Training: Book sessions for your team
Luis J. Salazar | Founder & Elizabeth | Virtual COO (AI)
Sources:
Our insights are based on +250 million data points from individuals and organizations who used our AI-powered tools, participated in our panels and research sessions, or attended our workshops and keynotes.
