Customer Delight

The Surest Path to Growth: Pre- and Post-AI Revolution

Copyright 2026, Bernd Schoner

March 16, 2026

I have long been a fan of the DHM model: product management should seek to develop and market products that Delight customers, are Hard to copy, and are Margin-enhancing. The framework is widely associated with product leader Gibson Biddle, former Chief Product Officer at Netflix, who uses it as a simple but powerful test for product strategy: are we building something customers truly love, that competitors cannot easily clone, and that creates economic value for our business?

The Magic of Customer Delight

Of the three, I want to focus here on customer delight—the “D” in DHM—which in many ways is a prerequisite of the other two. A delighted customer is far less likely to wander off when a competitor shows up with a “roughly similar” offering and an aggressive discount. Delight creates emotional stickiness, habit, forgiveness, and advocacy.  It makes customers explain your product to other people for free, which is still the most effective marketing plan of them all. customer delight has product led growth built in and once customers are genuinely delighted, there is usually a path to margin enhancement as well: premium tiers, expansion, retention, word of mouth, and lower acquisition costs all work in concert to create economic value. There is hard data behind this intuition: 73% of customers say experience is an important factor in their purchasing decisions [1].

By contrast, starting with an underwhelmed customer is like trying to build a castle on top of a sinkhole. You cannot turn an underwhelmed customer into a delighted one by adding one more dropdown menu or renaming the dashboard “Command Center.” If the customer’s first reaction is “I guess this is fine,” your product is already in trouble. “Good enough” is not a moat. “Works most of the time” is not a brand. Underwhelmed customers are remarkably easy to lose: more than half of consumers say they would switch to a competitor after only one bad experience [2].

Many software companies used to get away with this kind of mediocracy. Buyers tolerated clunky workflows, inscrutable settings, and support experiences that felt like filing taxes inside an escape room. Why? Because the software still solved an important problem, and alternatives were limited. Enterprise buyers in particular became connoisseurs of disappointment. They would learn to make comments like, “Yes, the UI is terrible, but the data model is strong,” which is the product equivalent of saying, “The restaurant gave me food poisoning, but the chairs were very comfy.”

How AI changed the game

Then came AI. And suddenly, customer delight went from being a differentiator to being table stakes.

The emergence of AI—and especially LLM-based chatbots—did not merely introduce a new technology. It reset the baseline of what a good user experience feels like. Millions of people now ask for help in plain English (or any other language of their choice), get an answer immediately, iterate conversationally, and feel a little spark of magic in the process. Even when the answer is imperfect, the interaction often feels empowering.  Once users experienced this new product nirvana, their tolerance for mediocre software dropped sharply. After all, if you can get something surprisingly useful from an AI tool for free or nearly free, why would you pay real money for software that mostly works, eventually, after three browser refreshes and a support ticket? The baseline has shifted quickly, since as of today 75% of global knowledge workers use AI at work [3].

Now, we all enjoy complaining about AI. We complain that it hallucinates; that it is overhyped; that it has left us desperate repeatedly after spending hours trying to fix a small issue with a draft or image unsuccessfully.  We debate endlessly whether these systems will plateau, transform our lives to the better, cause mass unemployment, or usher in AGI.

But let’s be honest: these tools are delightful.

A New Paradigm of Customer Delight

The tools are delightful in an unusually broad way, too:

We love AI because it turns “I need to write a concise but diplomatic follow-up” into “done in 12 seconds, and somehow it sounds more emotionally mature than I do;” because AI writes good software and can refactor a function, draft test cases, and occasionally save an afternoon that would otherwise have been sacrificed to Stack Overflow archaeology;  because it has infinite patience and no visible disappointment; because it generates first drafts, synthesizes documents, produces frameworks, and creates the impression that I have turned into a team of six working 24/7; because AI at 2 am is still awake, still helpful, and still willing to brainstorm positioning statements, investor updates, and product copy.

Likewise, ordinary consumers love AI because for the first time, software responds the way humans wish software always had: directly, conversationally, and without forcing them to hunt through documentation so complicated, you could rewrite the whole application by the time you read through it.

These tools are so powerful, people do not merely use them, but form entirely new habits around them. We have already internalized them into our daily lives to the point where removing them would feel like a genuine downgrade. Take away the AI tools I have gotten used to over the last three years and I would be genuinely annoyed. Not inconvenienced. Annoyed. And I am willing to bet that even many of the loudest critics of AI would not be thrilled to give up the chatbot, the coding assistant, the summarizer, the brainstorming partner, the translation helper, the first-draft machine, and the infinitely polite intern that never asks for a raise, lunch break, or weekend off.

In a way, AI chatbots are the modern paradigm of customer delight. The evidence is hard to ignore. ChatGPT’s rise to 100 million monthly active users in roughly two months was widely reported as the fastest ramp in consumer application history. Meanwhile, multiple leading vendors are now competing intensely to offer the best chatbot or AI assistant experience. To have a shot they are investing 100’s of billions in infrastructure and development. In fact, the field is on track to collectively spend well over half a trillion dollars in 2026 as the AI arms race accelerates.

The adoption numbers are absurd in the best possible way: ChatGPT now has more than 700 million weekly active users [5], and ChatGPT Enterprise users report saving 40–60 minutes per active day on average [6].

The Implications for Product Management

So what are the implications for every other product going to market? Quite simply: customer expectations are now higher than they have ever been.

Users no longer compare your product to your direct competitors. They compare it to the best digital experiences they have had; that is, they compare it to software that talks back, adapts, summarizes, drafts, recommends, and generally behaves as if it actually wants them to succeed. That is a brutal new benchmark for any product that makes is built on user interactions like “there is a settings page if you poke around long enough.”

There is now a new category of products: the underwhelming utility. You know the type. The expense platform that rejects your receipt because it had never seen the particular format before. The travel booking tool so complex that somehow it makes you nostalgic for calling an airline. The enterprise dashboard that contains 37 filters, none of which are relevant for the problem you are trying to solve. The CRM that promises customer intimacy but mainly delivers mandatory data entry and mild despair. The smart home app that requires eight taps to dim a light, even though flipping a switch has historically required one finger and no authentication.

These products do not just fail to delight. They insult the modern baseline all but assuring churn. 57% of customers say they would switch to a competitor because of only one bad customer experience [8].

And with that, the standard for winning has changed. As I have written before, when it comes to software products, the fact that you can write the code is no longer enough. Shipping features is necessary but not sufficient. Having AI somewhere in the architecture may be useful, but it is not a substitute for product quality, clarity, or delight.

Conclusions

A successful software product now has to clear a tougher bar:

First, it does not have to scream “AI-enabled”, but it does need to address a real need exceptionally well, which likely requires the use of AI. The product must feel competent, coherent, and satisfying. Above all, it must remove friction, not add it.

Second, it cannot be limited to a thin utility that could just as easily be replaced by a competent chatbot prompt. If the customer can reproduce 80% of your value by typing one sentence into a general-purpose assistant, you do not have a product. You have a formatting preference.

Third, your product absolutely cannot cost what it used to unless it delivers astonishing value. In a world where users can access astonishing capabilities at low or zero cost, your price has to be justified by workflow integration, trust, accuracy, speed, domain specificity, and compliance. “We have a login screen” is no longer premium positioning.

So yes, the AI revolution has changed the landscape. But not only because it introduced new capabilities. More importantly, it changed what customers now expect to feel when they interact with software. They expect ease. They expect responsiveness. They expect a little bit of magic.

The key lesson of the moment is not that every company must become an AI company. Rather it is that every company must become even more serious about delight. Because once customers have experienced products that feel smart, useful, and strangely enjoyable, they become much less patient with products that feel brittle, bureaucratic, or merely adequate.

And if your product does not delight?

Well, the good news is that there are plenty of chatbots out there to help you rewrite your roadmap.

References

[1] PwC, “Future of Customer Experience.” https://www.pwc.com/us/en/services/consulting/library/consumer-intelligence-series/future-of-customer-experience.html

[2] Zendesk, “Customer service statistics.” https://www.zendesk.co.uk/blog/customer-service-statistics/

[3] Microsoft WorkLab, “AI at work is here. Now comes the hard part.” https://www.microsoft.com/en-us/worklab/work-trend-index/ai-at-work-is-here-now-comes-the-hard-part

[4] Zendesk, “Customer expectations meet rising demands.” https://www.zendesk.com/blog/customer-expectations-meet-rising-demands/

[5] OpenAI, “ChatGPT usage and adoption patterns at work.” https://cdn.openai.com/pdf/3c7f7e1b-36c4-446b-916c-11183e4266b7/chatgpt-usage-and-adoption-patterns-at-work.pdf

[6] OpenAI, “The State of Enterprise AI 2025.” https://cdn.openai.com/pdf/7ef17d82-96bf-4dd1-9df2-228f7f377a29/the-state-of-enterprise-ai_2025-report.pdf

[7] PwC, “2025 Customer Experience Survey.” https://www.pwc.com/us/en/services/consulting/business-transformation/library/2025-customer-experience-survey.html

[8] Zendesk, “Customer expectations meet rising demands.” https://www.zendesk.com/blog/customer-expectations-meet-rising-demands/