When ChatGPT launched in November 2022, it marked AI's leap from the preserve of technology giants and research scientists to a tool seemingly accessible to any organisation. The subsequent proliferation of advanced models, including lower-cost ones like DeepSeek, has only amplified expectations about AI's transformative potential. Yet, this enthusiasm can be misplaced. 

As we wrote recently in Harvard Business Review, an unquestioning faith in AI's transformative powers may lead executives to make decisions that undermine their organisation’s capacity for profitable growth. 

The technology trap

Consider Snapchat's experience. The platform, launched in 2011, built its popularity on distinctive features such as disappearing messages and casual, in-the-moment visual communication. At its zenith, Snapchat was one of the world's most downloaded apps and had hundreds of millions of daily active users.

In 2023, the company introduced My AI, a generative chatbot powered by OpenAI that was pinned by default to the top of users' chat feeds. The assumption was that adopting cutting-edge AI technology would naturally enhance the user experience and drive engagement.

This assumption, however, proved flawed. Rather than viewing My AI as an innovative feature, users found it intrusive and unsettling, like an unwanted presence monitoring their conversations. The backlash was immediate and severe. App store reviews were inundated with one-star ratings specifically citing the unwanted chatbot. Snapchat's iOS rating in the United States plummeted to 1.67 stars. Google searches for "delete Snapchat" surged by 488 percent in subsequent months, reflecting widespread user dissatisfaction.

Nordstrom's experience with Trunk Club offers another cautionary tale. The American retailer acquired this personalised styling service for busy professionals and sought to embed AI into operations to scale the business. However, this approach diluted the service's core appeal of providing a high-touch, bespoke experience. The result was high return rates and sluggish sales. Nordstrom ultimately shut down the operation.

Leading with strategy

Both examples illustrate the same error: Charging ahead with AI technology without aligning it with the organisation’s strategic objectives. Now, contrast these failures with business moves that start with a clear strategy for providing a leap in value to buyers, and then use AI as an innovative tool to deliver that leap.

Take, for example, Yunji Technology's AI-powered delivery robots. In Chinese cities, instant delivery operates round the clock. However, hotel guests ordering snacks, replacements for forgotten chargers or other last-minute essentials late at night faced a problem: Security protocols typically prevented delivery workers from accessing guest floors, forcing guests to dress and collect orders from the lobby. 

Identifying this "last-inch" logistics problem, Yunji Technology developed a strategy to create cost-effective, personable delivery robots for direct room delivery, then tapped AI to actualise it. These robots navigate autonomously to guests' doors, announcing their arrival in childlike voices designed to elicit smiles. After guests retrieve their deliveries, the robots cheerfully wheel away, even cracking playful jokes with other lift passengers.

The robots quickly gained popularity. Positive hotel reviews mentioning the robots have exceeded 6 million, with 35 percent featuring user-uploaded images – far exceeding typical rates. Today, Yunji Technology robots operate in more than 330 cities and 34,000 hotels. 

Duolingo provides another compelling example. The founders recognised that traditional in-person language classes were expensive and constrained by geography and scheduling. Online platforms offered greater flexibility but suffered from one-size-fits-all approaches that couldn't adapt to individual needs. The lack of real-time interactivity meant many learners disengaged as programmes failed to maintain attention or meet individual requirements.

Duolingo’s strategy? Deliver a language-learning experience that would transcend geographical and scheduling constraints while remaining engaging, interactive and tailored to individual needs. It leveraged AI to create an online platform with personalised gamification. The technology analyses and gamifies user interactions, creating leaderboards and dynamically tailoring language lessons and feedback to each user's requirements.

Duolingo thus delivers bespoke education at scale while making language learning engaging and enjoyable – although its CEO drew controversy when he said last year that the company would become "AI-first". By the end of 2024, Duolingo had attracted over 103 million monthly active users, 950 million total downloads and 575 million registered users, making it by far the largest language-learning platform by user base and reach. Revenue hit US$748 million in 2024, up 40.8 percent from the previous year, with net income of US$86.9 million.

Levelling the playing field

As technology advances, AI agents themselves underscore why investing in AI technology without a clear strategy is unwise. Shoppers increasingly use AI agents like ChatGPT to inform purchase decisions. The agents scour data from every site – large and small, distant and near – to identify options, summarise reviews and assess offerings' objective value. 

Previously overlooked smaller companies can now more easily prevail over established players and brands if they offer compelling value. This raises the imperative for all organisations, large and small, to create strategies that deliver exceptional value, with or without AI.

Four critical questions

While AI continues evolving and its full potential remains unfolding, every company should ask four questions before reaching for the technology:

1. Are we adopting AI due to a fear of missing out or to deliver an objective leap in value?
2. Are we embracing AI hoping to maintain relevance, assuming it will generate growth? Have we fully considered whether and how it aligns with and supports our strategy and value proposition?
3. Are we conflating AI's technological innovation with genuine value? Will customers be convinced of the value it delivers?
4. Is AI helping us build a viable, sustainable business model aligned with our strategy? Or will its impact on our value proposition risk undermining profitability?

These questions serve as essential filters. How they are answered will largely shape whether AI applications will contribute to growth and profitability, or trap an organisation in a swirl of hype that leads nowhere.

This is adapted from an article published in Harvard Business Review.