'The progress is real': Deloitte report claims some firms are finally seeing return on AI investment — but still lack insight into what actually works

Deloitte finds 28% of businesses report current revenue growth, up from 20% earlier this year Productivity and improved decision-making are still performing better Sovereignty has become an integral part of AI strategies New Deloitte data has claimed that when it comes to business AI, we're finally

<![CDATA[ <article> <ul><li><strong>Deloitte finds 28% of businesses report current revenue growth, up from 20% earlier this year</strong></li><li><strong>Productivity and improved decision-making are still performing better</strong></li><li><strong>Sovereignty has become an integral part of AI strategies</strong></li></ul><p>New <a href="https://www.deloitte.com/content/dam/assets-zone3/us/en/docs/services/consulting/2026/the-state-of-ai-in-the-enterprise-2026.pdf" target="_blank" rel="nofollow">Deloitte</a> data has claimed that when it comes to business AI, we're finally at the stage of ROI, with more organizations than before now reaping the rewards of their costly deployments.</p><p>As many as 28% now report current revenue growth as a result of their AI strategies, up from 20% compared with earlier in 2026, the firm's report found.</p><p>Realizing higher revenue is largely seen as one of the most important returns from AI – with companies spending so much on services and infrastructure, they need to make that money back.</p><h2 id="ai-adopters-are-finally-seeing-increased-productivity-translate-into-higher-revenue">AI adopters are finally seeing increased productivity translate into higher revenue</h2><p>While other areas like efficiency and productivity (up 74%), decision-making and data insight (up 60%), cost reductions (up 45%) and better client relationships (up 44%) are all moving in the right direction, Deloitte says increasing revenue is still one of the slowest to deliver tangible results.</p><p>Deloitte's research also indirectly implies that businesses could be getting carried away with these early signs, because the majority (86%) of businesses surveyed plan to increase AI investments over the next financial year.</p><p>Additionally, while early ROI markers are now heading in the right direction, Deloitte worries that many businesses still don't have the right visibility into AI's financial impacts. While three in four (71%) are tracking total spend across token consumption, many can only see the total bill and not the breakdown of which activities lead to which spend.</p><h2 id="jobs-are-changing-because-of-ai">Jobs are changing because of AI</h2><p>It's becoming an increasingly familiar concept that employers need to redefine roles as AI becomes more engrained – developers are already seeing a shift from writing code to managing AI agents, for example.</p><p>Deloitte says there's a long way to go on this front, because while three in five companies have already redesigned jobs 'moderate[ly]', only 15% of them have done so 'extensive[ly]' and 1% 'completely.'</p><p>"Many are still just scratching the surface," the report claims.</p><h2 id="the-future-of-ai-investments-looks-different">The future of AI investments looks different</h2><p>Looking ahead, Deloitte expects sovereign AI to become even more of an important consideration. Nearly half of Americas (46%) and EMEA (47%) businesses are building in-house to regain control over their workloads, with three in four (74%) global respondents reporting at least 'moderate' importance of sovereign AI.</p><p>But the paper does acknowledge the benefits of using centralized hyperscalers, noting that going sovereign or building in-house comes with certain trade-offs – usually "capability, speed or cost."</p><p>At the same time, we're hearing more and more about enterprises like <a href="https://www.techradar.com/pro/we-were-highly-skeptical-about-the-cost-of-ai-revolut-reveals-it-cut-ai-coding-costs-by-cutting-out-claude-entirely">Revolut</a> picking up the infrastructure bills themselves, choosing to self-host open-weight models for greater cost predictability, usually at the expense of model performance in scenarios where they can achieve their desired outcomes without leading frontier models.</p><p>The report sets out six key considerations for the next stages of AI: adopters must respond by redesigning work as a whole, not just the workflow; costs need to be traceable to the token degree, not just the full bill; AI strategies must align with actions that produce revenue; governance needs to keep up as systems become more autonomous; early adopters are now starting to look at physical AI like robots, and so should you; and sovereignty should be seen as an ongoing effort consideration over a one-time fix.</p><figure class="van-image-figure pull-right inline-layout" data-bordeaux-image-check="" ><div class='image-full-width-wrapper'><div class='image-widthsetter' style="max-width:676px;"><p class="vanilla-image-block" style="padding-top:31.51%;"><img id="diM9tpwF2Lz85R8q85CT78" name="tr-g_news" alt="Google logo on a black background next to text reading 'Click to follow TechRadar'" src="https://cdn.mos.cms.futurecdn.net/diM9tpwF2Lz85R8q85CT78-1920-80.jpg" mos="" align="right" fullscreen="" width="676" height="213" attribution="" endorsement="" class="pull-rightinline" /></p></div></div></figure> </article> ]]>