Let me tell you something.
I have been doing this for over 32 years now. Working with technology, building systems, and helping companies transform. And the thing is, what I am seeing today with AI is not new. The pattern, the mistakes, the hype followed by disappointment. I have seen this before.
Let me take you back a few years. I was one of the lead architects at Capgemini, working on the Capgemini Cloud Platform. Our job was ambitious. We were making cloud migration seamless, more cognitive, more intelligent. We were injecting artificial intelligence into systems to make them self-healing, self-optimising. This was before ChatGPT existed. Before generative AI became a household term. Before every company on the planet decided they needed an AI strategy.
And you know what? Even then, working at the cutting edge of what was possible, I noticed something troubling. The technology we were building was genuinely transformative. But the penetration of AI into actual enterprise operations? It was remarkably small. We were pushing boundaries, but the reality on the ground was that most organisations were nowhere near ready to truly embrace what we were building.
Now, why am I telling you this? Because the same pattern is repeating itself today. Only this time, it is at a much larger and much more expensive scale.
Fast forward to now. Generative AI has exploded. LinkedIn is flooded with posts about agentic AI, autonomous systems, and the future of work. Every conference I attend, people are buzzing about large language models and intelligent automation. Companies are racing to announce their AI initiatives.
The numbers look impressive. According to McKinsey’s 2025 research, 78% of organisations now report using AI in at least one business function. That is up from 55% just two years ago. Sounds great, right?
But here is the thing.
Only 27% of white-collar workers say they frequently use AI in their daily work. Only 16% of workers overall report actually using AI at their jobs. And here is the one that really tells the story: 70 to 85% of AI initiatives fail to meet expected outcomes. In 2025, 42% of companies abandoned most of their AI initiatives. That is up from 17% just one year before.
So what is going on here?
This past weekend, I was at my local debate club. We had a fascinating discussion. And the topic that came up was this: Are companies adopting AI, or are they just buying it?
That question stuck with me. Because from my experience, working across dozens of Fortune 500 projects where AI readiness sits at the centre of the table, I can tell you the answer.
Most companies are not doing AI adoption right. They are getting it spectacularly wrong.
The Restaurant That Never Serves a Meal
Let me paint a picture for you.
Imagine a restaurant owner who dreams of running the finest dining establishment in town. He invests millions. He acquires a struggling bistro down the road for its recipes. He hires a celebrity chef. But not to cook. Just to attend the grand opening and pose for photographs. He installs the most expensive Italian oven money can buy. He redesigns the dining room with fashionable décor. He builds a beautiful website. He even places a robot greeter at the entrance to welcome guests.
Opening night arrives. The restaurant looks magnificent. The press coverage is glowing. The owner is photographed cutting the ribbon.
But here is the problem.
The actual cooks in the kitchen? They have never been trained on that expensive Italian oven. They are still using the old equipment because it is what they know. The recipes from the acquired bistro? They sit in a drawer, never integrated into the menu. The celebrity chef? He left after the photo opportunity. The robot greeter? It impresses visitors for about thirty seconds before they realise it can only say three phrases.
The restaurant looks innovative. It has all the ingredients of success. But the customers are not eating innovation. They are eating the same mediocre food, served the same old way.
Here is the key insight: The customer does not eat the oven. The customer eats the meal.
This is precisely what is happening with AI adoption in enterprises today. Companies are investing fortunes in the kitchen whilst forgetting that the entire point is to serve better meals.
What Companies Think They Are Doing
Now let me walk you through what I am seeing across industries. The patterns are remarkably consistent.
Acquiring AI startups: This has become the latest corporate sport. You see these announcements all the time. According to Georgetown University’s research, annual AI merger and acquisition transactions have more than doubled over the last decade. Apple alone has acquired 28 AI companies. Alphabet has 23. Microsoft has 18. Meta has 16. In 2024 alone, there were nearly 400 AI acquisitions globally.
Companies announce these deals with great fanfare. They believe they have bought AI capability.
But what actually happens? The startup’s technology was built for a different context entirely. The key talent, the very people who made the startup valuable, they leave within eighteen months. Integration with legacy systems takes years and costs millions more. The acquired product often ends up shelved or reduced to a minor feature nobody uses.
You have bought a bistro. But your kitchen still cannot cook.
Creating AI Centres of Excellence: This is another big trend. Organisations establish dedicated teams of AI specialists who will supposedly transform the company.
The reality? These centres become ivory towers. Disconnected from business operations. They produce impressive proofs of concept that never reach production. Here is a statistic that tells you everything: the average organisation scraps 46% of AI proofs of concept before they ever reach production. Only 26% of organisations have the capabilities to move beyond proof of concept to actual deployment.
Business units view these centres as overhead rather than partners. Knowledge remains siloed instead of spreading across the organisation. The Centre of Excellence becomes a centre of expensive irrelevance.
Hiring Chief AI Officers: Executive leadership for AI. Surely this means transformation will follow?
Not quite. The CAIO often has impressive credentials but no real budget, no genuine authority, and no integration with business units. They spend their time in strategy meetings and industry conferences rather than in operations. Tensions arise with the CIO, CTO, and CDO about who owns what. Research shows that 68% of executives report that generative AI has created tension or division between IT teams and other business areas.
After two years of frustration, the CAIO leaves for another company. The cycle repeats.
Rolling out Enterprise Copilot to everyone: Microsoft Copilot, GitHub Copilot, Gemini, Claude. Companies are licensing these tools at scale, announcing that their entire workforce is now AI enabled. GitHub Copilot alone has reached 20 million users. More than 90% of Fortune 100 companies are adopting it.
Sounds impressive, right?
But what actually happens on the ground? IT deploys it, sends an email announcement, moves on to the next project. Here is what the research tells us: 49% of employees say they have to figure out generative AI on their own. Only 45% of employees believe their organisation has successfully adopted AI. Compare that to 75% of executives who think the same thing.
You see the gap? The boardroom thinks AI adoption is working. The people doing the actual work know it is not.
Introducing vibe coding and Visual Studio integrations: The developer community is buzzing about AI assisted development. According to Stack Overflow’s 2025 Developer Survey, 84% of developers are either using or planning to use AI tools. 51% of professionals use AI daily.
But here is the thing. 48% of companies now use two or more AI coding tools. That is creating fragmentation rather than integration. Without proper training, without workflow integration, without measurement, these tools become expensive novelties rather than productivity multipliers.
Building agentic AI and autonomous systems: LinkedIn feeds are dominated by posts about agentic AI. Autonomous systems that can plan, act, and learn on their own. According to MIT Sloan Management Review and Boston Consulting Group’s 2025 research, agentic AI has already reached 35% adoption, with another 44% planning deployment soon.
But here is the reality. Most organisations struggle to move agentic AI from theory to practical return on investment. In any given business function, no more than 10% of organisations are actually scaling AI agents. Without well defined applications, leaders invest in experiments that do not scale.
Creating showcase chatbots and robots: Look how innovative we are! We have AI greeting customers!
The chatbot handles five percent of queries, frustrates customers with its limitations, and escalates everything to human agents anyway. Employees roll their eyes at the innovation theatre. Meanwhile, genuine operational problems remain completely unsolved.
Announcing strategic AI partnerships: This is perhaps the grandest illusion of all.
Companies proudly announce collaborations with NVIDIA to build AI capable computing infrastructure. Joint initiatives with OpenAI to explore generative AI possibilities. Partnerships with Microsoft, Google, and Amazon to co-develop AI solutions.
Press releases flow. Share prices respond. Industry analysts nod approvingly.
But let me ask you a simple question: how has this partnership changed what your procurement team does on a Tuesday afternoon?
The answer, almost invariably, is that it has not changed anything at all.
These partnerships are strategic signalling. They are impressive announcements designed for investors and industry observers rather than a genuine transformation of daily work.
The ovens get more expensive. The celebrity chefs get more famous. The partnerships get more prestigious.
But the kitchen still is not cooking better meals.
AI Indigestion: Swallowing More Than You Can Digest
So what is really going on here?
I call it corporate AI indigestion. Companies are swallowing AI investments faster than they can possibly digest them.
The numbers tell the story. Total corporate AI investment reached $252.3 billion in 2024. Private AI investment climbed 44.5% year over year. The generative AI market alone is projected to reach $59 billion in 2025 and grow to $400 billion by 2031.
All this money is flowing in. And yet, 70 to 85% of AI initiatives fail to meet expected outcomes. Only 6% of organisations qualify as AI high performers with meaningful business impact. And 42% of companies abandoned most of their AI initiatives in 2025.
The symptoms are predictable. Bloated budgets with unclear outcomes. Impressive announcements followed by quiet disappointments. Technology sits unused whilst the daily work continues unchanged. A growing gap between what the board believes is happening and what employees actually experience.
The root cause? Companies are treating AI as something you buy rather than something you do.
They are investing heavily in:
Assets. Startups, platforms, tools, infrastructure, NVIDIA partnerships, and OpenAI collaborations.
Structures. Centres of Excellence, Chief AI Officers, dedicated AI teams, and governance committees.
Symbols. Chatbots, robots, announcements, awards submissions, LinkedIn posts.
But they are investing almost nothing in:
Skills. Training every employee to use AI effectively in their specific role.
Processes. Redesigning workflows to incorporate AI meaningfully.
Culture. Making experimentation safe, encouraging genuine adoption, and celebrating practical wins.
Measurement. Tracking real outcomes rather than deployment metrics.
Here is a statistic that tells you everything. Boston Consulting Group’s research shows that successful AI transformations allocate 70% of their efforts to upskilling people, updating processes, and evolving culture. Most companies invert this ratio entirely. They spend 70% on technology and 30% on people.
That is why they fail.
The investment happens at the top, in boardrooms, press releases, and strategy documents. But the application must happen at the bottom, in daily workflows, individual tasks, and how real people do their actual jobs.
Most companies have poured resources into the top while doing almost nothing at the bottom. They have decorated the restaurant magnificently, whilst the kitchen remains unchanged.
What Real AI Adoption Actually Looks Like
Now let me tell you what genuine AI adoption looks like. Because I have seen it done right. It is rare, but it exists.
The thing is, real AI adoption is far less glamorous than what appears in press releases. It is quiet. It is practical. It is embedded so deeply into daily work that it becomes invisible.
And the results? They speak for themselves. Early adopters see an average 12% return on investment for generative AI implementations. Content marketing teams save around 11.4 hours per week per employee. Companies that moved early into generative AI adoption report $3.70 in value for every dollar invested. Top performers achieve $10.30 returns per dollar. Where AI is properly adopted, labour productivity grows 4.8 times faster than the global average.
In an organisation that has truly adopted AI, the finance analyst uses it to spot anomalies in reports that would have taken hours to find manually. The HR manager drafts job descriptions and screens applications in a fraction of the previous time. The engineer debugs code and writes documentation with AI assistance as naturally as using a calculator. The customer service agent has AI suggesting responses in real time, improving both speed and consistency. The project manager summarises meetings and tracks actions without spending hours on administrative tasks.
There is no fanfare. No innovation awards. No press coverage. Just people doing their jobs better, faster, with less friction and frustration.
The question that reveals genuine adoption is not “What AI have we bought?”
It is “What is AI helping our people do better today?”
If you cannot answer that question with specific, concrete examples from across your organisation, you have not adopted AI. You have merely acquired it.
How to Get It Right
So how do you actually get this right? Let me share what I have learned.
Start with the work, not the technology: Before purchasing any platform or tool, identify specific tasks where AI can genuinely help. What takes too long? What involves repetitive drudgery? What decisions could benefit from pattern recognition? Find the problems first. Then seek the appropriate solutions.
Embed AI into existing tools: Do not create new systems that require people to change their habits entirely. Put AI capabilities where people already work. Within their email, their service management platform, their spreadsheets, and their existing applications. Make AI invisible, not an additional burden.
Train relentlessly and continuously: A single webinar or email announcement is worthless. Real adoption requires ongoing coaching, regular sharing of use cases, hands-on practice, and persistent encouragement. Treat AI skills as seriously as any other professional development.
Remember that statistic I mentioned? 49% of employees say they have to figure out AI on their own. That is not a failure of technology. That is a failure of leadership.
Measure outcomes, not activity: Stop counting how many people logged in or how many licences were activated. Measure what actually matters. Time saved. Errors reduced. Decisions improved. Customer satisfaction increased. Employee frustration decreased.
Research shows that 66% of companies struggle to establish ROI metrics for AI initiatives. This measurement gap is why so many programmes fail.
Democratise rather than centralise: Instead of an AI Centre of Excellence that hoards knowledge, create AI champions within every team. Spread capability outward rather than concentrating it in a separate department that others view with suspicion.
Start small and scale what works: Pilot AI in one specific process. Prove genuine value. Learn from the experience. Then expand to the next process. This patient approach delivers far more than ambitious company wide transformations that collapse under their own weight.
The Challenge
Let me leave you with a challenge.
Forget your AI strategy documents. Forget your acquisition announcements. Forget your Centre of Excellence. Forget your Chief AI Officer’s impressive credentials. Forget your NVIDIA partnership and your OpenAI collaboration.
Walk onto your operations floor on a typical Monday morning. Find someone in finance, someone in customer service, someone in procurement, someone in HR. Ask them a simple question:
“How did AI help you do your job better today?”
If they look at you blankly, you have your answer.
You have not adopted AI. You have merely decorated the restaurant.
It is time to start serving actual meals.
Enamul Haque is Director of Intelligent Solutions at Wipro, a technology author, educator, and adjunct professor at Bangladesh Maritime University, where he teaches AI and Data Science at the Faculty of Ocean and Earth Sciences. He is passionate about bridging the gap between AI potential and practical application. Connect with him on LinkedIn or through his YouTube channel, Digital Deep Dive.
References
- McKinsey & Company (2025). “The State of AI in 2025: Agents, Innovation, and Transformation.” https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai
- Gallup (2025). “AI Use at Work Has Nearly Doubled in Two Years.” https://www.gallup.com/workplace/trends
- Fullview (2025). “200+ AI Statistics & Trends for 2025: The Ultimate Roundup.” https://www.fullview.io/blog/ai-statistics
- Center for Security and Emerging Technology, Georgetown University (2024). “Acquiring AI Companies: Tracking U.S. AI Mergers and Acquisitions.” https://cset.georgetown.edu/publication/acquiring-ai-companies-tracking-u-s-ai-mergers-and-acquisitions/
- MIT Sloan Management Review and Boston Consulting Group (2025). “The Emerging Agentic Enterprise: How Leaders Must Navigate a New Age of AI.” https://sloanreview.mit.edu/projects/the-emerging-agentic-enterprise-how-leaders-must-navigate-a-new-age-of-ai/
- Writer (2025). “Key Findings from Our 2025 Enterprise AI Adoption Report.” https://writer.com/blog/enterprise-ai-adoption-survey/
- Stack Overflow (2025). “Developer Survey 2025.” https://stackoverflow.com/survey
- GitHub (2025). “GitHub Copilot Statistics and Adoption.” https://github.com/features/copilot
- Netguru (2025). “AI Adoption Statistics in 2025.” https://www.netguru.com/blog/ai-adoption-statistics
- CB Insights (2025). “State of AI Report: 6 Trends Shaping the Landscape in 2025.” https://www.cbinsights.com/research/report/ai-trends-2024/
- Deloitte (2025). “AI Trends 2025: Adoption Barriers and Updated Predictions.” https://www.deloitte.com/us/en/services/consulting/blogs/ai-adoption-challenges-ai-trends.html
- Wharton School (2025). “2025 AI Adoption Report: Gen AI Fast-Tracks Into the Enterprise.” https://knowledge.wharton.upenn.edu/special-report/2025-ai-adoption-report/
- IBM (2024). “Global AI Adoption Index 2023.” https://newsroom.ibm.com/2024-01-10-Data-Suggests-Growth-in-Enterprise-Adoption-of-AI-is-Due-to-Widespread-Deployment-by-Early-Adopters
- Aristek Systems (2025). “AI 2025 Statistics: Where Companies Stand and What Comes Next.” https://aristeksystems.com/blog/whats-going-on-with-ai-in-2025-and-beyond/
- Worklytics (2025). “2025 AI Adoption Benchmarks.” https://www.worklytics.co/resources/2025-ai-adoption-benchmarks-employee-generative-ai-usage-statistics