The conversation in British boardrooms and tech hubs is currently dominated by frontier AI models. While the intellectual exercise of these massive systems is fascinating, it is creating a distraction. For the British enterprise in 2026, the obsession with what AI might do tomorrow is masking an urgent reality. Productivity is flatlining, and budgets are being drained by experiments without a finish line.
The Office for Budget Responsibility recently trimmed its medium-term productivity growth forecast to 1%, as the UK lags behind the US, France, and Germany in output per hour. UK firms are spending an average of £15.9 million on AI this year, with many boards planning to increase that by 40% over the next twenty-four months.1 Yet, despite this massive outlay, only about 12% of UK businesses report any actual increase in revenue from these projects.2 The gap between the “pilot” and the “profit” has become a chasm.
The Cost of the “Experimentation Era”
For the past few years, the country’s tech sector has lived in an era of “vibe coding” and speculative proofs of concept. It felt necessary to keep up with the global pace. But the market is losing its patience. Recent data suggests that 82% of executives have seen significant, unplanned increases in their cloud and generative AI costs. Even more telling is that 61% of organisations have had to cancel other vital projects just to keep these AI experiments afloat.3
The total spend is up 8% year-on-year, specifically in the SaaS market, but the number of applications in the average stack has remained flat. Businesses are not looking for more tools, but rather for tools that work. They are tired of “black box” frontier models that require expensive prompt engineers and endless fine-tuning just to produce a slightly better email draft.
Predictable ROI vs. Frontier Uncertainty
The primary difference between a frontier model and a practical SaaS platform is the “Return on Investment” (ROI) timeline. A frontier model is a bet on the future. A well-designed SaaS platform is a tool for the present. In the current economic climate, the luxury of waiting three to five years for a “transformative” result is disappearing.
UK business leaders are beginning to realise that the most successful implementations are those that solve specific problems with high efficiency rather than reimagining everything. We see this in the rise of vertical SaaS, software built for specific industries like manufacturing or healthcare, which is projected to grow to $194 billion by 2029.4 These platforms promise to reduce your inventory errors by 15% or speed up your invoicing by 40%.
Why Specialisation Wins
The problem with frontier models is their lack of context. They are “jacks of all trades” but masters of none. In contrast, specialised platforms are designed around a business’s specific workflows. This is where companies like RedoQ find their niche. They develop “applied” technology by focusing on practical, scalable products rather than just chasing the latest theoretical breakthrough. When a firm uses a tool like Kuick Connect for remote access, it gets a 100% reliable connection so its team can work collaboratively.
Evidently, 71% of UK businesses now prefer “ready-to-use” external software over building their own AI systems.5 The “build-it-yourself” era of 2024 and 2025 led to a graveyard of abandoned projects. Gartner recently predicted that at least 30% of generative AI projects would be scrapped after the initial proof of concept because they simply could not scale or prove their worth.
The Skills Gap
One of the biggest hurdles to the frontier model is the human element. The UK is currently facing a massive AI skills gap.5 The country lacks enough skilled professionals who can safely manage and oversee autonomous agents. In fact, only about 20% of companies have the resources to govern these systems.
This is why SaaS platforms with “baked-in” intelligence are so much more effective. They allow a workforce to benefit from advanced technology without needing a PhD to run it. The workers just do a better job without thinking much about AI under the hood.
The Case for “Boring” Technology
The most successful CEOs I speak with are no longer asking what a specific model can do. Instead, the conversation keeps returning to the idea of “boring” technology. Boring means the system does not crash when you need it most. Boring means costs are predictable month over month. The real questions now are how a platform can lower customer acquisition costs or whether a piece of software can guarantee 99.9% uptime for a supply chain. These are necessary questions because 52% of organisations overspent on their software budgets last year, often due to complex contracts.3 We need platforms that provide transparent, usage-based pricing that aligns with the actual value we receive.
The financial discipline must also lead to faster speed-to-value as the companies can’t afford to wait twelve months for a pilot project to conclude. Instead, the return on investment should be visible within the first ninety days of implementation. The sector must also address operational security, as 80% of businesses cite ethical and security concerns as a primary barrier to adopting new technology.5 Platforms that have the necessary guardrails built into their core are needed today.Â
Closing the Gap
The UK economy has a £232 billion opportunity if our SMEs can digitise to the level of the top 20% of firms.7 This will not happen by waiting for a frontier model to solve all the problems. It will happen by adopting reliable, high-performance SaaS platforms that solve the problems we face today. Technology must be treated as an infrastructure project rather than a science project. Let the frontier models continue their experiments in sandboxes while the rest of us wait and watch for what comes out.
- https://news.sap.com/uk/2025/10/uk-business-investment-in-ai-to-rise-by-40-on-average-over-the-next-two-years-but-a-long-term-strategy-and-people-focus-is-needed-to-make-it-a-success/#_ftnref1Â
- https://www.gov.uk/government/publications/ai-adoption-research/ai-adoption-researchÂ
- https://zylo.com/blog/saas-statistics/Â
- https://www.forrester.com/blogs/saas-as-we-know-it-is-dead-how-to-survive-the-saas-pocalypse/Â
- https://www.ncs-london.com/blog/ai-adoption-in-uk-in-2026Â
- https://www.gartner.com/en/newsroom/press-releases/2024-07-29-gartner-predicts-30-percent-of-generative-ai-projects-will-be-abandoned-after-proof-of-concept-by-end-of-2025Â
- https://committees.parliament.uk/writtenevidence/157291/html/Â
