AI Adoption Status Report: Are we seeing the value yet?
Our latest LEADINGThought series has been focusing on the progress in AI adoption and its human consequences. We covered several aspects surrounding this period of technological disruption – the importance of building the adaptability muscle in opening up to possibilities and dealing with a rapidly changing environment, the human consequences of following an automation vs. augmentation path to AI adoption, and the early evidence of AI impact on junior careers. We also heard from leaders who are helping to evolve our thinking on how we preserve what makes us human while adopting AI to solve real problems, and the importance of evolving our workplaces to enable this transition.
It therefore feels like a good time to take stock – review the evidence so far, consider where the biggest opportunities are, and recognise what is needed for humans to remain at the centre of the equation.
AI adoption round-up
According to Stanford’s Erik Brynjolfsson “We are flying blind into one of the most consequential periods in world history”.
He and his team at the Stanford Digital Economy Lab recognised the importance of timely, trusted evidence that can support decision-making at the speed the technology is progressing. Their economic indicators have been utilising data from research by the global payroll firm ADP, as well as Anthropic’s economic index with monthly updates. The latest update has highlighted that whilst there is an increasing trend towards adoption at work, recent surveys appear to diverge, with one survey, Harley et al. [5]reporting a decrease in adoption, while two other surveys report continued increases towards 58% adoption.
Anthropic’s latest economic index and survey report reinforce the patterns around increased usage, while also highlighting that large majorities of people are reporting productivity gains in speed, scope, and quality of their work (86%, 82%, and 69%, respectively), while 27% report gains through cost savings on services they would otherwise have to purchase.
Nevertheless, despite the rate of adoption and reported individual productivity gains, there is currently no decisive evidence for a transformative shift in growth and productivity. The Stanford Digital Economy Lab Transformation dashboard is being refined to reflect what type of transformation this would be and whether it would end up with replacement or augmentation. Recent evidence linking large-scale adoption surveys to administrative payroll records found that the promised benefits of high levels of adoption failed to materialise on the bottom line, as the models were not truly integrated [7]. “Freed-up” capacity was spent on responding to the model’s output and manually wiring it into existing systems and decisions, so despite individual productivity gains these never surfaced as earnings or reclaimed hours.
From a broader labour market perspective, the Stanford Digital Economy Lab’s economic indicators highlight that overall, we are not seeing widespread displacement from AI, although junior hiring has slowed down. This is mainly due to automation of the jobs most exposed to AI, such as software engineering, rather than an overall trend. This work is building on their earlier work that reported significant drop in junior level vacancies[5].
So, which are emerging as the most promising focus areas for AI adoption and what would it take to realise the benefits?
The opportunity in AI adoption
SMEs have received a lot of attention in this area given the potential for efficiency and productivity gains. AI adoption in SMEs is rising fast, and this is supported by figures from both the OECD and, closer to home, the British Chamber of Commerce. According to OECD’s SME Digitalisation for Competitiveness report 39% of global SMEs now use AI applications, while generative AI adoption has risen to 26%. In the UK specifically, the British Chambers of Commerce stated 54% of SMEs are actively using AI in 2026, up from 25% in 2024 and 23% in 2023. Yet only 6% report a transformational impact, fundamentally changing work processes and creating significant value.
HR and recruitment appear to be areas where SMEs are realising benefits, as some form of AI is integrated into workflows to automate repetitive tasks and improve decision-making. Examples are around the use of AI to enhance existing Applicant Tracking systems (ATS) through automated candidate matching, bias free scoring, and secure in system data processing. However, there are also ongoing concerns regarding unintentional bias, because of training algorithms on historical data that might perpetuate gender bias for example. In our latest LEADINGThought leader interview, Jenny Garrett OBE highlighted specifically the importance of understanding what algorithmic bias exists and direct it somewhere else if we don’t want simply to repeat or accelerate the patterns of the past.
According to cross-industry analysis and structured experiments by Accenture and Google reported in the Harvard Business Review in August, the biggest opportunity to close the AI adoption-benefit gap is in the middle office [2]. Middle office accounts for more than four-in-10 working-hour tasks across 18 industries, and as such holds the biggest untapped source of AI value. They highlight that most AI adoption roadmaps jump from the back office, where returns are thinning, to the front office where payoff is much further away. In one live experiment improving intelligence integration in a typical middle-office workflow lifted overall success from 42 percent to 80 percent, and ensured experts focused on where it truly mattered, scaling up their expertise rather than see it disappearing.
The first representative, international firm level data study by Yotzov and others [11], found that on net, firm-level productivity and employment expectations imply a 0.8% boost to output over three years as a result of AI adoption. This is alongside forecasting a productivity boost of 1.4% with an employment cut of 0.7%.
Realising the Benefits from AI adoption
Therefore, on evidence so far, and notwithstanding the continuing refinement of how adoption and gains are measured, access to AI isn’t the problem; neither is willingness to experiment. The issue appears to be focus, and ensuring the right conditions are in place for the value to materialise.
Research from BetterUp Labs highlights the importance of setting the right conditions in determining whether AI adoption produces performance or just activity[3]. They state that the biggest predictors of AI output quality are – how leadership communicates about AI, how much trust exists, and whether people still feel they matter.
This is summarised perfectly by @Adam Grant – ‘Adoption you can buy but trust you have to earn!’. The importance of trust and psychological safety in highly volatile environments, as one of the key cornerstones of adaptability, learning, exploration, and innovation, was also highlighted by Dr Lucrecia Grandolini in one of our recent LEADINGThought leader interviews.
Lack of focus and blunt mandates lead to what Niederhoffer, Robichaux, and Hancock termed ‘AI workslop’ that ends up undermining trust and leads to employees adopting AI ‘performatively’[8]. This then ends up limiting the commercial value derived from AI adoption. AI advantage is built as much by what leaders say no to as by what they fund. As workslop rises, and trust erodes, productivity drops. This is one of several adverse effects from adopting an automation vs. an augmentation path.
The Journey continues…
This latest newsletter focused on progress so far with AI adoption and initial findings on the greatest opportunities, and what’s needed to realise commercial value. This isn’t about availability and usage, it’s about focus and integration into business workflows.
As several commentators are talking about the ‘end of the AI bubble’ this remains a period of profound transformation. However, the pace and nature of the transformation is a critical choice that is being made every single day.
We will continue to track how this evolves through our LEADINGThought series, bringing you evidence-based insights and thinking grounded in real-life experience that can evolve our approach and move us forward.
Resources
1. Anthropic Economic Index report: Cadences, 26 June 2026.
2. 4 steps to transform the Middle Office with AI .H. James Wilson, Chetna Sehgal, Michael Zimmerman, Ali Arsanjani, Blaise Abderholden and Jimmy Priestas (HBR, August 20, 2026).
3. The case for conditions. BetterUp Labs Research 2026.
4. Bharier, D., Etheridge, B., & Morais, P. (2026). AI Adoption and Workforce Change in SMEs. ISER Working Paper Series. Colchester: University of Essex.
5. Brynjolfsson, E., Chandar, B., Chen, R. Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence (November 13, 2025).
6. Hartley, Jonathan and Jolevski, Filip and Melo, Vitor and Moore, Brendan, The Labor Market Effects of Generative Artificial Intelligence and Job Loss Fears (December 18, 2024). Available at SSRN: https://ssrn.com/abstract=5136877 or http://dx.doi.org/10.2139/ssrn.5136877
7. Humlum, Anders and Vestergaard, Emilie, Still Waters, Rapid Currents: Early Labor Market Transformation under Generative AI (May 2025). NBER Working Paper No. w33777, Available at SSRN: https://ssrn.com/abstract=5250742
8. Liebscher, A., Lee, A. Y., Rapuano, K., Kellerman, G., Niederhoffer, K., & Hancock, J. (2026, February 5). Workslop: Examining the prevalence, antecedents and consequences of low-quality AI-generated content at work. Retrieved from osf.io/preprints/psyarxiv/5f78h_v1.
9. Powering Productivity: AI and the Future of UK Work. British Chamber of Commerce, March 2026
10. Stanford Digital Economy Lab. "AI Economic Indicators." August 2026. https://digitaleconomy.stanford.edu/project/indicators/
11. Why Companies that Choose AI Augmentation over Automation may Win in the Long Run. Jan-Emmanel De Neve, Jeffrey T. Hancock, and Kate Niederhoffer. (HBR, April 15, 2026).
12. Yotzov, Ivan and Barrero, Jose Maria and Bloom, Nicholas and Bunn, Philip and Davis, Steven and Foster, Kevin and Jalca, Aaron and Meyer, Brent and Mizen, Paul and Navarrete, Michael A. and Smietanka, Pawel and Thwaites, Greg and Wang, Ben Zhe, Firm Data on AI (February 2026). NBER Working Paper No. w34836, Available at SSRN: https://ssrn.com/abstract=6246334