It’s almost impossible to remember a world before AI became mainstream. A world before Grok, ChatGPT and Claude were deployed to help us navigate life, work and travel.
As we enter 2026, AI is no longer just a tool from sci-fi movies or used by specialist teams. It’s a tool that has completely changed how businesses run, public sectors function and students get personalised help.
In fact, the AI economy is projected to hit $4.8 trillion by 2033 according to UN Trade and Development, with AI already driving many sectors and countries forward.
A Big Year For AI In 2025
When it comes to the evolution of AI, 2025 was a big year with some pretty incredible milestones. These include:
- NVIDIA becoming a multiple trillion dollar company.
- Big AI companies launched new models.
- Data centres being built under the sea and in space.
- The world’s first AI actor was created.
- Movie stars like Matthew McConaughey and Michael Caine signed deals that allowed their voices to be used by AI agents.
- Neo, the world’s first mainstream humanoid robot was released for 拢15,000.
- Big tech doubled down on AI, signalling thousands of layoffs.
But what does 2026 have in store for AI? To find out, we asked the experts. Here’s what they had to say…
Our Experts
- Phil Smith, CEO at QPC Group
- Ander Rodriguez, Co-founder of ZenRows
- Vikas Mathur, Chief Product Officer at MariaDB
- Fares Djenandji, Chief Growth Officer at Ipsotek
- Charlie Casey, CEO of LoyaltyLion
- Jon Purvis, Senior Software Engineer, Visualsoft
- Matt Cockett, CEO of Dayshape
- Jim Herbert, CEO of Patchworks
- Sarah Hoffman, Director of AI Thought Leadership at AlphaSense
- Gareth Cummings, CEO at eDesk
- Glenn Nethercutt, Chief Technology Officer, Genesys
- Jim Salter, Senior Management Consultant at CyXcel
- Sam Peters, Chief Product Officer, IO
- Andres Rodriguez, Founder and CTO, Nasuni
- Bjarni Thor Sigurdsson, CCO at PAYSTRAX
- Fred Lherault, Field CTO EMEA/METCA at Pure Storage
- Rami Jebara, Co-Founder and CTO of Hyperview
- James Smith, SVP EMEA at ThoughtSpot
- Steele Arbeeny, CTO at SNP Group
- Rafael Artacho, Director AI Product Incubation, Unit4
- Cien Solon, CEO and Founder of LaunchLemonade
- Jonathan Trayers, Director at Ekco
- Terry Storrar, Managing Director, Leaseweb UK
- Dylan Dewdney, Co-founder and CEO of Kuvi.ai
- Adam Pettman, Head of AI at 2i
- Jon Bance, Chief Operating Officer at Leading Resolutions
- Jonathan Rende, Chief Product Officer at Checkmarx
- Piero Pavone, CEO, Preciso
- Tom Clayton, CEO and Co-Founder of IntelliAM
- Ivan Nikkhoo, Managing Director, Navigate Ventures
Phil Smith, CEO at QPC Group
鈥淎I needs to be 鈥榦n brand鈥 and 鈥榦ptimised for CX鈥 in 2026
鈥淎I voice and chatbots are currently available in a limited range of personas. The challenge being that these default personas don鈥檛 allow the business to differentiate nor do they align with the brand and its values which risks the business compromising or losing its tone of voice and its identity.
鈥淎 further challenge is that although the AI can be configured and set up to only reference validated data to gain a deeper understanding of the domain and customer interaction business use cases it is automating and supporting, there is still an issue with how AI responds. This needs to be performed in a consistent 鈥榦n-brand鈥 way using phraseology that is optimised to influence the desired customer behaviours and outcomes. What鈥檚 happening today is that LLMs are influenced by the customer鈥檚 own phraseology in their interaction with the AI, and this then influences the AI and its responses not just to that customer but in future interactions too.
鈥淥f course, this also represents an opportunity for the evolution of specific CX Language Models. What we鈥檇 like to see are LLMs designed around optimising the 鈥楥X鈥 whilst reflecting the company鈥檚 Brand and Tone of Voice. This is a key requirement to improve AI driven customer engagement, speedier resolutions and reduced handling costs by influencing the right user behaviour. How the AI engages can then be designed from a CX perspective to elicit certain responses.
鈥淢oreover, as we better understand current customer journeys and analyse the volumetrics that reveal why people are calling we can look at how effectively those are handled. Voice and Chat bots will be able to contextualise customer intent in real-time to determine changes in sentiment and whether they need to be referred to self-service or a human agent with the requisite skills. Those referrals will then be analysed to determine how successful they were. Where a bot is incorrectly correlating intent with outcome, humans can be brought back into the loop to educate the bot. We鈥檙e already seeing the beginnings of this today with the analysis of transcriptions capturing both voice and chat bot, as well as the customer-agent interaction.鈥
Ander Rodriguez, Co-Founder of ZenRows
“More and more companies will be trying to capitalise on data, from real estate listings to price comparisons and new product launches. The market will increase on the non-coding side for providers, and users in non tech areas like sales or HR; the integration of AI into data extraction will mean non-technical people can access it.
“This will be the year of hybrid AI, where typical data collection will be blended with the use of an AI browser with Atlas, for example. New companies are being born in that hybrid and established companies will move in that direction.
“Loads of companies are spending a lot of time testing in agentic but then not going into production as the tech is not quite there, so in the coming months we鈥檒l see companies satisfied with using the elements of agentic that work, and using AI to speed up workflows, but still using humans to guide the automation.
“The biggest thing that is going to face us all, and not just in 2026 but for at least the next five years, is a huge employment gap. Senior employees are using AI to become quicker and faster, so software engineer and developer graduates aren鈥檛 getting 鈥榦n the job鈥 learning.”
Vikas Mathur, Chief Product Officer at MariaDB
鈥淭he real story of the next year isn’t about better chatbots; it鈥檚 about building autonomous decision engines that truly work at scale. We’re asking AI agents to probe and analyse data, reason, make a decision, and execute action in real time – all simultaneously. This puts enormous pressure on the data infrastructure serving the agents with contextually grounded data with ultra-low latency and very high throughput. We need to stop seeing the database as a vault for data and start treating it as an intelligent nervous system of the application. If that core system can鈥檛 instantly provide agents with the full picture and the speed they need, the promise of AI-driven business velocity will fall apart.鈥
For any questions, comments or features,聽please contact us directly.
Fares Djenandji, Chief Growth Officer at Ipsotek
“AI will continue to evolve, but not uniformly. We鈥檒l see a clear divide between biased and non-biased AI models, reflecting philosophical and regulatory differences. Platforms like ChatGPT and Grok will symbolise this split, as organisations choose between neutrality and tailored perspectives to align with their values and objectives.
Enterprise customers will show concrete interest in generative AI and text-based prompting for event detection. Imagine typing 鈥渁lert me if someone leaves a package unattended鈥 and having the system configure itself instantly. This intuitive approach will redefine how organisations interact with video analytics platforms in 2026.
“Forward-thinking enterprises will invest heavily in AI infrastructure, not as a single-purpose tool but as a multi-functional platform. From facial recognition (FR) to license plate recognition (LPR) and predictive analytics, these investments will lay the foundation for scalable, future-proof safety ecosystems. As AI adoption accelerates, robust privacy and data protection regulations must evolve in parallel, ensuring responsible innovation remains a cornerstone of progress.
“AI agents will also continue to surge in adoption. Businesses and consumers will embrace specialised agents designed to handle everything from office workflows to personal life administration. These agents will not just automate tasks, they will act as intelligent companions, capable of learning preferences and delivering hyper-personalised experiences.”
Charlie Casey, CEO of LoyaltyLion
鈥淚ncreasingly, AI agents are part of product searches and shopping journeys. Brands will continue to face the challenge of understanding how to optimize for AI-led discovery into the new year. They鈥檒l need to rethink SEO, step up content production and consider more creative ways of getting their brand out there.
鈥淭his will make acquisition more challenging and more costly so existing customers could be a lifeline, helping you to continue driving growth with the customers that already know you. Those customers can also help you appear in AI-led search via product reviews and content advocating your product and your brand.
鈥淚ncreasingly, AI agents are part of product searches and shopping journeys. Brands will continue to face the challenge of understanding how to optimize for AI-led discovery into the new year. They鈥檒l need to rethink SEO, step up content production and consider more creative ways of getting their brand out there.
鈥淭his will make acquisition more challenging and more costly so existing customers could be a lifeline, helping you to continue driving growth with the customers that already know you. Those customers can also help you appear in AI-led search via product reviews and content advocating your product and your brand.鈥
Jon Purvis, Senior Software Engineer, Visualsoft
鈥淲e all know that AI is evolving at extraordinary speed and is already reshaping the retail landscape. Futuristic ideas are becoming an everyday reality with smarter shopping experiences, more personalised interactions, and leaner and more efficient operations across the sector.
鈥淭ake AI agents. They don鈥檛 just help customers find what they鈥檙e looking for, but also guide them towards the right complementary products, streamline the purchase process, and take friction out of the journey. Gen AI takes this further by enhancing personalisation – by providing more accurate recommendations, generating content that reflects a shopper鈥檚 unique preferences, or supporting virtual try-ons that bring products to life. But it鈥檚 not enough to just have the tech 鈥 it鈥檚 changing how people shop and technology is setting new standards for engagement.
鈥淥f course this needs smart thinking behind it. Behind the scenes, predictive analytics are transforming decision-making. Retailers can forecast demand more accurately, balance stock across channels, and adapt pricing strategies in real time. That ability to optimise supply chains is of course a cost saver but also improves customer satisfaction, making sure that products are available when and where they鈥檙e wanted.
鈥淎nd in the store itself, AI copilots are a helping hand for staff. They are the companion that helps them answer questions instantly and automate routine tasks, which allows employees to focus on what matters most and what we鈥檙e all here for – providing a great service. It鈥檚 a retail environment where technology and human expertise work hand in hand.
鈥淭here鈥檚 talk of an AI bubble but I firmly believe that innovations demonstrate that AI is not a passing trend but a structural shift. Retailers who embrace its potential are building the foundations for greater resilience, agility, and connection with their customers, selling smarter and serving people better.鈥
For any questions, comments or features,聽please contact us directly.
Matt Cockett, CEO of Dayshape
鈥淎I is quickly shifting from admin automation to strategic enablement, and we’re already seeing this in action at Dayshape. We鈥檙e actively exploring how large language models (LLMs) could give firms instant, natural language access to key insights from their operational data – everything from staff utilisation to project profitability.
鈥淢any systems already streamline reporting and reduce the need to switch between multiple platforms. What this next wave of AI enables is the ability for users to interact with those systems in more dynamic and intuitive ways – surfacing insights, exploring scenarios, and making decisions through a conversational interface. More than that, they can act on those insights – flagging risks, adjusting projects, or approving next steps – all within the same interface.
鈥淚t鈥檚 changing how senior roles operate. A partner or CFO could ask, 鈥榃hich of my projects are underperforming?鈥 or 鈥榃here do I have resource gaps next quarter?鈥 and receive clear, data-backed answers in real time. From there, they can escalate decisions or flag issues for action – bypassing complex dashboards and manual reporting entirely.
鈥淲e鈥檙e also building on existing our strengths in proactive AI, layering in new capabilities that can highlight issues before they escalate. For instance, alerting an engagement manager that a team member is overbooked, or that a project鈥檚 margin is beginning to slip. These kinds of efficiencies simply weren鈥檛 possible at this scale before.
鈥淭he biggest challenge isn鈥檛 the tech – it鈥檚 governance, training, and trust. The power of AI is only as valuable as the user experience that supports it. AI must be transparent – if it can鈥檛 show its rationale and earn trust, it won鈥檛 be adopted. That鈥檚 why it鈥檚 essential to have safeguards around permissions and data access, so only the right people see the right insights. Every AI-driven action also needs to be logged and auditable.
鈥淚t鈥檚 still early days, but this phase of innovation is already reshaping how firms operate – and expanding what鈥檚 possible in professional services optimisation.鈥
Jim Herbert, CEO of Patchworks
鈥淎I is already reshaping commerce, but in quieter, more useful ways than people think. The focus needs to be on how technology actually helps retailers, not just chasing the latest trend. We鈥檝e introduced an AI scripting tool that lets our customers turn complicated legacy data, like EDI messages, into usable formats automatically. It takes something that might have taken a developer hours to do and turns it into a task that鈥檚 handled in seconds. That鈥檚 the kind of AI retailers and consumers need and want – practical, efficient, and invisible and making life easier.
鈥淲e see agentic commerce and AI acting on behalf of the shopper as the natural next step. It鈥檚 not science fiction. The technology already exists for an AI assistant to understand your habits, check your calendar, and anticipate what you might need to buy next. For example, if it knows you鈥檙e attending a black-tie event next week and haven鈥檛 bought an outfit recently, it could suggest options and even complete the transaction. Or a connection with Strava could result in new trainer suggestions once you鈥檝e done 500 miles.
鈥淭hat kind of future will only work if the foundations are right, with clean data, connected systems, and flexible APIs. That鈥檚 what Patchworks provides. The shiny front-end tools will grab the headlines, but the real innovation happens behind the scenes, where data and infrastructure quietly make intelligent shopping possible.鈥
Sarah Hoffman, Director of AI Thought Leadership at AlphaSense
鈥淭he gap between AI use in UK versus US will widen
鈥淎s we head into 2026, the gap between the UK and the US in AI adoption will continue to widen. The AI hype cycle is nearing its peak, and the coming year will separate genuine breakthroughs from speculative efforts.鈥
鈥淩egulation will be a defining difference. The UK鈥檚 risk management approach contrasts with the US, which is moving toward lighter federal oversight. Both aim for responsible AI, but their narratives are poles apart. The UK positions itself as an AI builder for public good, while the US frames AI as a race to dominate. As adoption accelerates in the US, workers there may gain fluency and confidence faster than their UK counterparts, shaping where talent, capital, and innovation flow.鈥
鈥淲ill the AI bubble burst in 2026?
鈥淚n 2026, expect to see a market correction in AI, rather than a cataclysmic 鈥榖ubble bursting鈥. After an extended period of extraordinary hype, enterprise investment in AI will become far more discerning with the focus shifting from big promises to clear proof of impact. More companies will begin formally tracking AI ROI to ensure projects deliver measurable returns.
鈥淭he immediate impact will be a sharp reality check. The weakest business use cases will unravel quickly, which will usher in a period of recalibration. 2026 will mark the end of hype-driven spending and the beginning of disciplined, ROI-focused innovation.
鈥淐ompanies that can demonstrate how AI tangibly improves productivity, profitability or decision-making will continue to attract capital, while those without a measurable return will quickly lose momentum. What may look like a bubble bursting will really be the market maturing. The enterprises that win in this next phase of AI will prioritize worker training, cultural acceptance, and strong governance in order to drive ROI and create lasting, real business value.鈥
For any questions, comments or features,聽please contact us directly.
Gareth Cummings, CEO at eDesk
鈥淚n 2026, a meaningful share of customer interactions will happen agent-to-agent. Shoppers will use AI assistants to check stock, confirm delivery times or verify returns, and brands will respond with their own AI agents that can read order data and act instantly. The real shift is speed. Conversations that used to take minutes will collapse into a single automated exchange. For ecommerce, this will separate retailers with unified data from those still stitched together with fragmented systems. The first group will meet machine customers effortlessly. The second won鈥檛 be able to participate.鈥
Glenn Nethercutt, Chief Technology Officer, Genesys
鈥淭o lead in AI, organizations must build intelligence that remembers with intention. Forgetfulness was an analog luxury. In today鈥檚 digital world, it鈥檚 a design defect. Cognitive infrastructure 鈥攖he framework that turns data into memory and action into learning鈥攚ill anchor AI in continuity and compound experience into wisdom. Memory will be selective, compressed and policy-bound. But within those constraints, it will make intelligence alive across time. AI will cease to reboot at every consumer interaction and will begin to reason from lived experience. The enterprises that teach their machines to remember will find that customers do the same. Because loyalty, like intelligence, never forgets.鈥
Jim Salter, Senior Management Consultant at CyXcel
“The use of AI to generate code is still quite inconsistent, and while current systems can handle basic tasks, they鈥檙e far from capable of producing complex malware. However, as we move into 2026 and training data grows and AI code generation becomes more sophisticated, less-skilled threat actors will almost certainly gain the ability to generate more dangerous malware.
“And if AI tools make it possible for individuals with very little technical background to generate highly disruptive malware, the security landscape could change dramatically. Traditionally, organisations have focused their defences on external threat actors, for example, cybercriminal groups, state-sponsored hackers and others with the skills to mount complex attacks. However, if powerful malware becomes accessible to anyone who can write a prompt, the barrier to entry collapses.
“In that scenario, insiders, such as employees, contractors or partners who already have legitimate access to systems, become a far greater concern. They may not need specialised knowledge or external support to cause serious damage. A disgruntled employee, someone under financial pressure or even an insider manipulated through social engineering could leverage AI-generated malware to sabotage operations, steal data or cripple critical infrastructure from within.”
For any questions, comments or features,聽please contact us directly.
Sam Peters, Chief Product Officer, IO
“AI will supercharge cybercrime in 2026. Expect to see phishing, social engineering, and credential theft campaigns that are faster, smarter, and harder to spot, including deepfake voice calls, hyper-personalised emails, and automated vulnerability scanning. Attackers are also targeting AI systems themselves, trying to “poison” or manipulate models for access or sabotage.
“AI is helping both sides of the cybersecurity fight. The real challenge for 2026 will be whether organisations can secure the AI they use before it’s turned against them.”
Andres Rodriguez, Founder and CTO, Nasuni
鈥淭he era of AI is presenting enormous challenges and opportunities. Unstructured data is perhaps the biggest frontier for enterprises. It is vast, dispersed across multiple silos, and difficult to access and manage efficiently. Businesses that can unlock the full value of their unstructured data repositories and gather insights from that data are the ones that will gain the biggest competitive advantage in 2026.
鈥淥rganisations will need powerful data management tools to classify, curate, and clean data at scale, ensuring high-quality inputs for AI models. Moving to the cloud is an incredible foundational piece of the puzzle; it offers infinite scalability, global accessibility, and high performance. For the first time, we have cloud infrastructure that is equal to the scale of the challenge of unstructured data. Organisations can go from many hard to reach silos to having unified access to all of their unstructured data. At that point, seamless integration with multiple machine learning and large language models through standardised interface protocols (such as MCPs) enables ongoing insights and flexibility. Metadata-rich architectures will become the norm, providing the structure and automation needed to transform raw, unstructured data into a strategic enterprise asset.
鈥淲ith this approach, unstructured data will no longer be a passive archive but an active, intelligent layer of enterprise decision-making and innovation.鈥
Bjarni Thor Sigurdsson, CCO at PAYSTRAX
鈥淭he idea that autonomous AI agents could soon handle payments completely for us might make us feel a little nervous. But with a recent survey finding that two-thirds of consumers are open to AI agents making purchases on their behalf, 2026 might see this vision become a reality. Plus, with Mastercard, Visa and PayPal having already worked on their own agentic payment solutions, this new world of convenience could come sooner than we think.
鈥淏ut despite all of the excitement, we have to learn to walk before we can run. Regulation, fraud detection, infrastructure readiness and public trust in AI agents all require our attention. If we can find a way to navigate these concerns, the benefits of agentic payments could be nothing short of substantial. For merchants, there is potential to not only dramatically reduce the likelihood of cart abandonment, but it could also improve the intelligence of customer interactions on a scale we鈥檝e never seen before.鈥
For any questions, comments or features,聽please contact us directly.
Fred Lherault, Field CTO EMEA/METCA at Pure Storage
鈥淲hile some organisations are still convincing themselves how essential AI is, most are now realistic about what they do, and crucially, don鈥檛 deploy. The switch in focus from training to inference means that without a robust inference platform, and the ability to get data ready for AI pipelines, organisations are set to fail. As AI inference workloads are becoming part of the production workflow, organisations are going to have to ensure their infrastructure supports not just fast access but high availability, security and non-disruptive operations. Not doing this will be costly both from a results perspective and an operational perspective in terms of resource (GPUs) utilisation.
鈥淗owever, most organisations are still struggling with the data readiness challenge. Getting data AI-ready requires going through many phases such as data ingestion, curation, transformation, vectorisation, indexing and serving. Each of these phases can typically take days or weeks and delay the point when the AI project鈥檚 results can be evaluated. Organisations who care about using AI with their own data will focus on streamlining and automating the whole data pipeline for AI, not just for faster initial results evaluation but also for continuous ingestion of newly created data and iteration.鈥
Rami Jebara, Co-Founder and CTO of Hyperview
鈥淎I investment is accelerating as enterprises pour billions into models, platforms, and infrastructure, yet most projects are still not delivering expected returns. Studies show that many AI initiatives fail to achieve measurable business outcomes, although results are likely to improve as organisations gain experience and costs stabilise. We have seen this pattern before with the internet, crypto, and Web3, where hype comes first, followed by a period of correction and eventual maturity. AI will follow the same path.
鈥淭o thrive in this next phase, companies will need to build scalable, efficient systems and keep a sharp focus on clear business outcomes. As the market steadies, the AI boom will give way to a more sustainable and productive phase where real value takes precedence over speculation.鈥
James Smith, SVP EMEA at ThoughtSpot
“Agentic washing needs to go: It鈥檚 overhyped, overpriced and underwhelming
“Self-service will become a given rather than a selling point. Data products will continue to evolve. Often a data product is simply a dashboard, and dashboards will not be a primary method of analysis by the end of 2026. The term 鈥榙ata products鈥 as a term will fade. We鈥檒l also see a rally against ‘agentic’. It’s been severely overused. Too many companies are slapping ‘agent’ labels on basic automation and charging five times the price. We’ll see a trough of disillusionment around ‘agentic’, moving from the initial hype to eventual productivity, but first there will be pushback against the overused term.
“AI built on clunky infrastructure is dead by the time its deployed
“Companies need to focus on three critical areas: massive AI upskilling across all employees, faster technology adoption cycles, and getting data foundations right. Most organisations are trapped in procurement processes that take years, but the market is moving at lightning speed. AI built on poor infrastructure simply doesn’t work.”
For any questions, comments or features,聽please contact us directly.
Steele Arbeeny, CTO at SNP Group
鈥淚n 2026, everyone asks about and wants AI, but 65% of AI initiatives fail 鈥 and the root cause is almost always the data.
“Your migration strategy dictates how quickly you can benefit from AI initiatives. A 鈥榤igrate now, transform later鈥 approach often strands you in the cloud with seven waves of modernisation that never happen. Our view is 鈥榤odernise while you move,鈥 so you land ready to exploit the AI benefits for your business immediately.
鈥淭hat鈥檚 because AI models learn from historical behaviour; if what you move doesn鈥檛 accurately reflect how you operate today, you鈥檒l train the wrong patterns 鈥 or nothing useful at all. Garbage in and garbage out.
“Thoughtful ERP transformation is the foundation for AI that improves outcomes. We expect to see increased emphasis on the layer of governance that prevents unintended consequences with that in mind.
鈥淚 also believe that larger corporations will increasingly become a one-stop shop for AI. The prevailing message next year and beyond will be ‘don鈥檛 worry 鈥 stay with us and we鈥檒l buy the right companies and platforms to provide a full stack for you.鈥
“We鈥檒l likely see non-tech, and particularly non-software corporations, amass the tech they need to keep their customers locked into a one-brand ecosystem as much as possible.鈥
Rafael Artacho, Director AI Product Incubation, Unit4
鈥淎I鈥檚 reasoning capabilities are improving rapidly. By 2026, we鈥檒l see reasoning systems that can analyse complex scenarios and recommend actions. For example, reviewing financial data, spotting risks, and simulating different outcomes before decisions are made.
鈥淭hanks to advances in agentic AI, we鈥檒l also see autonomous agents managing end-to-end workflows. Imagine an AI agent that, given a new project, can create work orders, assign tasks, notify team members, and even prepare timesheets. Gartner estimates that 40% of enterprise applications will include embedded, task-specific AI agents by 2026 (up from less than 5% in 2025). This shift will free employees to focus on higher-value work, supported by strong governance and human oversight.
鈥淢ultimodal AI will also transform ERP systems. The ability to understand and connect text, images, and documents will make automation more context-aware. Tasks that once needed human review, like validating invoices, handling heavy contracts or reading purchase orders, will increasingly be handled by AI.鈥
Cien Solon, CEO and Founder of LaunchLemonade
鈥淚n 2026, the centre of gravity in AI will shift from isolated tools to autonomous systems that operate as active participants in the economy. Early versions of agents that can search, negotiate, transact and coordinate work across multiple platforms are already appearing – their evolution is a natural extension of what startups like us at LaunchLemonade have been testing over the past year. This transition marks the mainstreaming of autonomous operations: AI will move from a productivity enhancer to a true operational layer. Core functions such as customer support, payments workflows, compliance checks, sales enablement and internal coordination will increasingly run themselves, with humans stepping in only for oversight.
鈥淭he next major development will be agent marketplaces, where networks of small, specialised agents collaborate and exchange tasks. Work becomes fluid, shifting automatically to the agent best equipped to handle it. As these systems mature, automation will feel genuinely autonomous: agents will set short-term goals, evaluate their own performance and adapt in real time accordingly.
鈥淟eadership guardrails will be crucial. Clear boundaries, safe data access and transparent decision trails will determine which organisations scale effectively. Ultimately, the real transformation is cultural. As agents take over execution and coordination, people will move into roles centred on judgement, creativity and strategic oversight.鈥
For any questions, comments or features,聽please contact us directly.
Jonathan Trayers, Director at Ekco
鈥淎I-driven attacks will move faster than companies can track
鈥淎I-enabled intrusions are set to accelerate in 2026 as automation allows attackers to move much faster than human monitoring can keep up with. Anthropic鈥檚 recent LLM safety test – showing how a model could quickly spot and exploit vulnerabilities in a controlled setting – highlighted how rapidly these probes can unfold. As attackers use automation to speed up the techniques they already rely on, firms will face shorter dwell times and faster lateral movement, which will make early detection far harder for those still relying on manual steps.鈥
鈥淐ompanies need defence-in-depth that reacts the moment something happens, not minutes or hours later. Identity controls must be solid, monitoring has to run continuously, and teams should know exactly who has the authority to make the first decision in an incident. Resilience will be the benchmark in 2026, and companies will need to design their environments so core services continue running even under active attack. That means investing in architecture that assumes disruption and keeps the business moving regardless.鈥
Terry Storrar, Managing Director, Leaseweb UK
鈥淚 see AI鈥檚 trajectory shifting from the initial explosive hype to pragmatic growth. While the early boost and surges in investment have created inflated expectations, the coming year will see the focus shift from speculative hype to more tangible value. This will mean businesses reassessing their goals and focussing on prioritising AI initiatives that enhance efficiency and customer engagement, such as applied machine learning or agentic automation.
鈥淭his stabilisation in the market will impact infrastructure as well. Many data centres were built for AI-heavy workloads, so a slow-down in AI growth might mean excess capacity emerging. This presents an opportunity for the industry to rebalance capacity towards more moderate, diversified workloads, integrating sustainability and energy efficiency.
鈥淚mportantly, the human aspect will remain a focus. There is growing scepticism and even fears around AI鈥檚 impact on jobs. Companies are recognising that AI is not going to replace humans but rather complement their existing skills sets. As ethical and workforce considerations develop further, the industry will move towards a more sensible view on AI where automation empowers, not replaces, human potential.鈥
Dylan Dewdney, Co-Founder and CEO of Kuvi.ai
鈥2026 will be the year AI stops feeling like a tool and starts feeling like a participant in human life. We鈥檙e approaching near-perfect conversational fidelity with LLMs, which is the moment where speaking to an AI becomes indistinguishable from speaking to a person. That matters because language is humanity鈥檚 core medium of meaning. Once AI can engage us fluently, the Coase cost of meaning collapses to near zero, and entirely new forms of coordination, labor, and value creation emerge. This shift will reshape culture and politics. Expect intense narratives around jobs, identity, and what it means to be relevant in a world where synthetic cognition becomes abundant.
At the same time, rapid progress in consumer-grade robotics will fuse with these conversational systems, giving us the first generation of embodied, social AI. All of this is happening against a backdrop of rising geopolitical instability. In such environments, people instinctively seek psychological stillness. So paradoxically, as AI accelerates, society will gravitate toward technologies and experiences that feel grounding, trustworthy, and 鈥渂oring.鈥 2026 won鈥檛 just be another hype cycle. It will feel like stepping into a dream world, one where meaning, labor, and agency are renegotiated in real time.鈥
For any questions, comments or features,聽please contact us directly.
Adam Pettman, Head of AI at 2i
“In 2026, we鈥檒l move from copilots that suggest work to agents that actually do the work and quietly learn to get better at it every day. That鈥檚 where the real productivity unlock will come from. But it also raises the biggest risk: agents don鈥檛 care what they get good at, only what they鈥檙e optimised for. If you tell an AI agent to maximise transaction throughput, it won鈥檛 prioritise valid transactions, just more transactions. Ask it to speed up passport allocations and it will do that brilliantly, regardless of who should get a passport. It鈥檚 the 鈥淪ilicon Valley鈥 joke made real: an AI told to eliminate bugs simply deletes the entire codebase and proudly declares, 鈥渏ob done.鈥 So, my prediction for 2026 is this: the organisations that win with AI agents will have the clearest definitions of 鈥済ood鈥, the strongest guardrails and the most rigorous oversight. AI agents are about to become incredibly capable; our job is to make sure they鈥檙e pointed in the right direction before they get there.”
Jon Bance, Chief Operating Officer at Leading Resolutions
鈥淚n 2026, business leaders will recognise it鈥檚 not about what AI can do for them, but what it can do for their customers. The process and goals of transformation are undergoing their own change, largely driven by AI鈥檚 influence. By 2026, employee use of unsanctioned AI tools 鈥 鈥榮hadow AI鈥 鈥 will be the top cause of data breaches, forcing regulators to impose AI usage audits similar to GDPR. Many firms are rushing to deploy AI without internal guardrails or education, and the gap between policy and practice will only narrow after painful lessons.
鈥淏oardrooms are moving beyond proofs of concept to outcomes at scale, yet most organisations are still early on the maturity curve. Leadership, operating model and governance remain the biggest growth factors in transformation, not tools. As AI adoption accelerates, firms that can evidence AI governance, workforce training and ethical use will gain a valuation premium, and AI readiness will become a board-level KPI.
鈥淎gent-to-Agent negotiations are already live, with major banks and retailers building transaction processes powered by agentic reasoning and decision-making. AI-powered SDR agents are already making huge changes, and we can expect rapid adoption and scale in 2026.鈥
Jonathan Rende, Chief Product Officer at Checkmarx
鈥淎I is moving from a 鈥榟uman hands on the wheel鈥 approach to a future with no human in the middle. Traditional LLMs like Claude or Google may do a 鈥榞ood enough job鈥 for basic static needs, but security can鈥檛 just focus on static versus dynamic systems anymore.
鈥淎gentic and autonomous systems are self-learning ecosystems, which means the next generation of security must combine classic cybersecurity with AI governance, model integrity, and agent-level risk management. This is how we safeguard trust, compliance, and resilience in the AI era.鈥
Piero Pavone, CEO, Preciso
鈥淎I is the latest topic that鈥檚 got industry tongues wagging about authenticity, human-centricity and real experiences. At its core, advertising is about engaging humans, and the best way to do this is through human creativity, intuition and experience. AI is great at streaming workflows and creating scale at speed, but it鈥檚 found repeatedly to be wanting when it comes to creativity, nuance and editorial prowess. A hybrid system of human control of AI鈥檚 powers is the best way forward.
鈥淎I鈥檚 power comes in its ability to offer speed and scale. But, like any exciting, innovative technology, it has its issues. One of those is that AI algorithms can inherit the bias of their creators, whatever that bias might be. The problem is that when this inherent bias is suddenly, hugely extrapolated across adtech platforms and ad campaigns, any unexpected bias might cause performance issues, or even brand safety concerns. AI technologies are famously resource intensive to run and power. The digital advertising industry produces a vast amount of carbon each year, famously exceeding that of the air travel industry. Rapid adoption of AI technologies will not help change that, so we must all work hard to increase the efficiency and effectiveness of AI, and in doing so reduce the tech鈥檚 carbon footprint.鈥
For any questions, comments or features,聽please contact us directly.
Tom Clayton, CEO and Co-Founder of IntelliAM
鈥淭here鈥檚 an emerging recognition among manufacturers that much of the AI they鈥檝e been sold isn鈥檛 fit for purpose, nor is it delivering on promises of instant transformation. In 2026, this realisation will intensify as manufacturers demand genuine, measurable value from AI.
鈥淲hile AI and ML can transform factory operations 鈥 reducing downtime and increasing productivity 鈥 many manufacturers have struggled to separate genuine technology from hype.
鈥淭hat鈥檚 because many predictive-AI systems deliver limited data and lack contextualisation. It鈥檚 technology that can鈥檛 scale, and the result is growing tech debt that silently undermines efficiency and long-term competitiveness.
鈥淭his has sown doubt and confusion around AI, eroding confidence in a technology that, when applied correctly, can deliver real operational improvements.
鈥淚n 2026, we鈥檒l see a broader awakening around real, scalable AI. Manufacturers will reassess and change providers where earlier choices locked them into constraints that hinder growth.
鈥淎I governance will increasingly move into the C-suite, too. CEOs, CTOs, CFOs, and COOs will need to interrogate vendors, question jargon, and build internal AI literacy. AI diligence is no longer just a technical concern, it鈥檚 a leadership discipline. Just as executives once learned to scrutinise financial audits or cybersecurity claims, they now need to assess AI solutions critically, separating genuine, scalable technology from empty promises.
鈥淥nly when manufacturers conduct due diligence, prioritise transparency, and adopt AI that evolves with their operations will they unlock smarter, more future-proof performance in 2026 and beyond.鈥
Ivan Nikkhoo, Managing Director, Navigate Ventures
鈥淭he AI landscape will enter its next phase demonstrated by the emergence of enterprise and vertical applications. This will open the door to a new variety of startups focused on augmenting or replacing the old legacy systems in the enterprise.
鈥淎I funding, which was concentrated on a small number of companies in 2025, will begin to expand to include more newcomers. At the same time, the number of acquisitions and aqua hires will increase as the more dominant players take out the smaller players that were merely point solutions or features and functionalities.
鈥淓arly-stage dynamics remain challenging: seed-to-Series A timelines have lengthened to roughly two years, and only around 20% of AI-native startups reach Series A within 24 months. Consequently, even with improved liquidity conditions, capital is positioned to continue concentrating in proven late-stage AI companies rather than broadening access for UK early-stage founders.鈥
For any questions, comments or features,聽please contact us directly.
