From Dot-Com to AI: Have We Learned Anything?
Back in the late 1990s, as the dot-com boom heated up, I was building software for financial services firms. The more I looked at what was happening, the clearer it became: this wasn’t really a technology problem at all. It was a business strategy problem.
That insight pulled me into strategy work just before the dot-com crash. I wish I could say I predicted the bust — I didn’t. But I could see that companies were throwing technology at the wall, hoping it would stick, without addressing the fundamentals of data, culture, or business models.
Fast forward to today, and I feel it's déjà vu with AI transformation. Some say this time it’s different. Others insist it’s history repeating itself. The truth is, it’s a bit of both.
What’s the Same: Lessons That Still Apply
Business first, technology second
The fatal flaw of early digital projects was being technology-led instead of business-led. Many initiatives failed because no one defined ROI or aligned tech to strategy. Today, many AI projects risk the same fate: experimentation without a business strategy.
“Digital transformation was often messy, with manual workarounds everywhere, because businesses jumped on technology without aligning it to strategy or people.”
20-ish years into the digital revolution, COVID-19 exposed many of these flaws. As organisations struggled to work remotely, it became evident that many processes, whilst appearing to be digital from the outside, were still fundamentally manual on the inside.
Data is still the foundation
Whether digital or AI, if your data is poor, it’s garbage in, garbage out. Without investment in governance, quality, and access, no shiny AI tool will deliver.
Culture and capability matter most
You can’t just drop new technology into an old organisation. Both digital and AI transformations demand upskilling, culture change, and leadership buy-in.
Risk must be designed in
Digital raised security and privacy concerns. AI adds model risk, bias, hallucinations, and IP leakage. The lesson holds: governance belongs at the core, not bolted on at the end.
What’s Different This Time
From deterministic to probabilistic
Digital systems were predictable: a website or app did exactly what you coded. It did the same thing each time. AI systems are probabilistic. They generate outputs that can surprise you — sometimes helpfully, sometimes problematically. That makes governance and trust fundamentally different.
Two uses of AI
There’s a world of difference between:
-
Personal productivity AI (email drafting, calendar assistants, copilots).
-
Enterprise AI transformation (reconfiguring manufacturing, optimising logistics, redesigning customer offerings).
The former is flashy and easy to adopt, but productivity hacks are shallow and don't scale. The latter delivers real competitive advantage. Leaders mustn’t confuse the two.
Cost structures feel familiar but hide complexity
-
The early days of digital often meant big upfront CapEx. It was only much later that SaaS-style and cloud subscription-based models became more prevalent.
-
AI usually starts as SaaS-style OpEx — “just another subscription.” That makes adoption deceptively easy.
-
But behind the scenes, token usage, environmental costs, and bespoke enterprise systems (like Masonite’s AI-driven configurator for billions of door designs) reshape the economics of businesses.
The true cost picture of AI is still evolving.
AI creates more AI
Digital transformation brought tools like Stack Overflow that did improve and speed up coding. But AI can literally write more AI. As a result, the pace of change is much greater. This recursive acceleration is unprecedented — and we don’t yet know where it leads.
Energy and geopolitics
AI’s energy demands are enormous. They may be decreasing exponentially over time, but at the same time, our usage of them is increasing. AI may become the catalyst for breakthroughs in renewable energy and smart grids. They may also shape global competition, with China, the US, and others racing to secure AI leadership by securing energy innovation.
The true costs of using AI can be obscured by providers selling services at a loss to capture market share in an AI land grab. We need to keep a watchful eye to ensure the benefits from using AI actually offset the real costs.
Cultural risks
Digital transformation automated the mundane administrative elements of many jobs. Calculators might have eroded our ability to do mental arithmetic, but they didn’t erode our critical thinking skills. AI risks encouraging intellectual laziness: people rubber-stamping AI outputs without engaging their own minds.
Used poorly, AI could deskill decision-makers just when we need sharper judgment. On the other hand, used well, it could equally help us to think much more critically.
(This is a problem we've grappled with at StratNav, leading to our current "human-in-the-loop" approach.)
Where AI Value Is Emerging
-
Knowledge management. For decades, organisations have wasted hours searching for “the latest deck” or “the right policy.” AI finally offers a practical solution to this.
-
Productivity at scale. Ten engineers doing the work of thirty isn’t a gimmick — it’s a game-changer.
-
Enterprise reconfiguration. From manufacturing configurators to financial modelling, AI can tackle problems too complex for deterministic digital systems.
So, Is It the Same or Different?
It’s both the same and different.
-
The same: AI transformation risks repeating digital’s mistakes — tech-first, ROI-undefined, culture-blind initiatives that fail to deliver.
-
Different: AI’s probabilistic nature, recursive capacity, and energy demands add new risks and opportunities we haven’t faced before.
For leaders, the challenge is two-fold:
-
Look backwards — learn from digital’s mistakes and avoid repeating them.
-
Look forwards — apply creative and critical thinking to the genuinely new dimensions AI introduces.
Final Thoughts
AI transformation isn’t just digital transformation rebranded. It’s a continuation — but with sharper edges, faster cycles, and higher stakes.
The leaders who win won’t be those who jump on the bandwagon. They’ll be those who:
-
Ground AI in business strategy.
-
Build governance into the fabric.
-
Distinguish between flashy productivity hacks and true transformation.
-
And invest in the culture and capabilities to use AI as a lever for growth.
The question is: will your organisation be one of them?
Next steps
-
Book a call to strategise your AI transformation: https://calendly.com/chriscfox/discuss-your-need
Credits:
- With thanks to those members of our #stratChat community who contributed their thoughts to this post. Quotes may be attributable to members on a Chatham House Rules basis.
Addendum: Will we see an "AI crash"?
I’ve seen many arguments suggesting that we won’t see a crash like we saw at the start of the digital revolution in 2001 (the so-called “dot-com bubble”).
Most of these arguments, I believe, rest on either a misunderstanding or a faulty memory of what happened then — or perhaps they’re being made by people not old enough to have experienced it first-hand.
The 2001 crash was not caused by technology itself. It was caused by excessive investment and inflated valuations placed on organisations without sound business models. The economic shock rippled through markets, but it did not stop — or even significantly slow — the advance of digital transformation.
-
Some businesses, like Pets.com, Webvan, Boo.com and eToys, failed because they couldn’t sustain a viable model.
-
Others, like Amazon and eBay, survived — perhaps at much lower valuations — and went on to dominate.
-
And still others, like Google, PayPal and Salesforce, launched straight through the downturn and thrived in the recovery.
I would never claim to predict the future. But I do see many reasons to believe the same pattern could emerge with AI. The cycle has little to do with the specific technology — digital then, AI now — and much more to do with how capital flows into industries under disruption.
Some companies will fail spectacularly. Some will survive. Others will be born in the aftermath and flourish. That’s the nature of disruptive change.
The open question is: have we truly learned the lessons of the past (this time), or are we about to repeat them?
