With increased coverage and usage of AI tools, questions have been raised about whether the AI bubble may burst. Ben Gilbert, VP at Humara, explores where the warnings come from and what the future holds for AI in the business world.
Warnings from the Bank of England about the ‘AI bubble’ bursting have concentrated minds on where the technology is delivering value in the real world.
The bank’s finance policy committee warned that ‘the risk of a sharp correction has increased’ following the soaring valuations of AI companies. The fear is of a repeat of the dot-com bubble of the late 1990s when valuations collapsed as investors feared many companies might never deliver on their digital technology promises.
The current apprehensions about AI have been triggered in part by reports from authoritative bodies such as MIT and Gartner. MIT believes 95% of organisations receive no ROI from investments in Generative AI, while Gartner has predicted that more than 40% of Agentic AI projects will fail by 2027 due to rising costs, governance challenges and lack of clear ROI.
It’s not all bad news, however. Gartner also believes Agentic AI will resolve 80% of common customer service queries by 2029, leading to a 30% reduction in operational costs. Other commentators simply reject the MIT findings. Who should we believe?
Correction of course – not a crash
While we may see a market correction, AI adoption is certain to continue after businesses have learned how to approach it in a more systematic way. The real skill in designing AI applications lies in identifying opportunities to serve customers more effectively while also generating commercial value.
Currently, most companies are leveraging AI for efficiency, such as automating workflows or streamlining customer support. However, these benefits often take years to deliver tangible returns and are difficult to measure beyond time savings.
Consequently, AI projects with unclear gains or delayed ROI will be the first to be cut. Without definite value, these initiatives risk becoming costly experiments rather than profitable investments. As expenses mount and results lag, businesses will inevitably abandon projects that don’t directly contribute to growth or profitability. With this in mind, there seems no good reason why firms have so readily embraced AI for process efficiency and customer support, yet hesitate to apply it in areas like sales, where the outcomes are clear and immediately measurable.
Unlocking AI’s true potential in sales
Sales is one area where AI agents, guided by human insight, can deliver a bottom line impact that captures the attention of even the CFO. In the telecommunications sector, for instance, AI-led sales journeys have already driven significant uplifts in conversion, increased attachment rates and boosted average order values – all while enhancing customer confidence and satisfaction.
The key to this success lies in ensuring that AI recommendations are rooted in human expertise, sales psychology and real-world experience. To achieve immediate success, businesses deploying AI for sales should adopt a robust, proven and accessible framework from the start.
Bounded vs unbounded problems
Bounded problems are highly defined and predictable, representing the areas where AI has already proven its effectiveness. Examples include automated reporting, routine customer service and churn prediction.
Unbounded problems are more complex, involving fluid inputs and subjective human variables like nuanced decision-making and personal interaction. It is precisely in these areas that AI holds the most transformative potential.
AI’s role in digital sales
Digital sales is a classic unbounded problem, as it requires a sophisticated understanding of human emotion, intent, timing and behaviour. In a highly competitive industry like telecommunications – characterised by a massive customer base and complex product bundles – AI’s ability to navigate and adapt to these unbounded challenges presents a major opportunity for growth and innovation. Customers often, for example, have difficulty working out plans and pricing and may be frustrated about some aspect of a service. These concerns lead to queries that AI can resolve.
An AI sales agent, meticulously trained on extensive user interactions and human sales expertise, can craft a highly impactful and personalised online sales experience.
The alternative is relying on a cumbersome blend of automated systems and human effort to meet customer demand, which is unlikely to foster competitiveness. As other companies embrace advanced AI, maintaining outdated approaches will inevitably leave businesses behind.
Finding the right partners
In today’s fast-changing landscape, companies must act decisively when choosing partners with the expertise to drive success – particularly in areas like digital sales, where the business case is increasingly clear. The right AI solutions can deliver rapid value, continuously improve over time and quickly make a noticeable impact on the bottom line. Meanwhile, businesses clinging to outdated technologies risk falling behind more agile competitors.
