Technology and economic pressures are set to reshape finance functions in 2026, as senior executives predict a shift from AI hype to practical deployment, tighter scrutiny of cyber risk, and a renewed focus on cost control and process simplification.
Leaders across accounting software, tax technology, infrastructure and cybersecurity say finance teams will move beyond experimentation with AI. They expect firms to embed automation more deeply into reporting, tax and risk workflows, while regulators and boards increase their oversight of new tools.
AI in finance
Hugh Scantlebury, Founder and CEO of accounting software provider Aqilla, said finance leaders are moving away from anxiety about AI and towards questions of implementation and control.
“As we head into 2026, the conversation around AI will become far more grounded and pragmatic, and far less tied to job-loss anxiety or fear of the unknown. That shift will refocus discussions on sensible, responsible use; whether that’s processing high-volume data and speeding up reporting or reducing manual entry and freeing teams to focus on analysis.
Scantlebury said the main issue is how organisations put AI into practice.
“So, the defining question won’t be whether finance and accounting teams use AI – spoiler alert, most already are, even if they don’t realise. It will be around implementation. And to that end, we advocate a responsible, transparent approach with clear oversight and guardrails. That way, as the year progresses and finance teams rely more heavily on AI outputs, they can be confident that they understand how those outputs are generated. Without that clarity, they risk creating systems that are fast but untrustworthy, with all the accompanying reputational and ethical risks,” said Scantlebury.
He said demand from customers is moving towards practical use cases that reduce manual workloads.
“In 2026, customers want more practical help: how to remove repetitive work so they can focus on analysis, interpretation and decision-making. AI’s value lies in freeing people to apply judgement, not replacing it. That’s why the future of the finance function isn’t machine-led accounting; it’s people-led automation, where human thinking directs automated processes to deliver speed, accuracy and clarity,” said Scantlebury.
He added that developers of finance software face growing expectations over how their products behave.
“As developers of AI-enabled finance software, we have a responsibility to support users in this shift by ensuring our systems behave predictably, transparently and in ways that reinforce good financial practice. What we don’t want is for 2026 to become the year of over-automation – because customers aren’t asking for fully machine-led accounting. They want clarity, confidence, and tools that remove repetitive work while keeping judgement firmly in human hands. Remember the ‘A’ in AI – the ‘artificial’ in ‘artificial intelligence.’ It’s programming, subroutines, and logical analysis. It isn’t human – at least not yet, and hopefully not ever,” said Scantlebury.
Prompt literacy
Charis Thomas, Chief Product Officer at Aqilla, said finance professionals will need new skills as AI tools become more common in day-to-day work.
“2026 will be the year when a new kind of tech literacy becomes essential. Remember how we once had to learn to search the internet effectively by refining queries, judging sources and understanding how information was surfaced? Finance teams will now need to learn how to interact with AI in the same way. This prompt literacy isn’t about tricks or shortcuts; it’s the modern equivalent of learning how to research properly. As with all computing scenarios, the quality of the question shapes the quality of the answer. Or more colloquially, put rubbish in, and you’ll get rubbish out. Developing that skill across finance teams will be every bit as important as the technology itself,” said Thomas.
Thomas said the integration of AI with search tools will put fresh emphasis on transparency and traceability.
“Alongside this, the long-running convergence of search engines and AI is becoming far more visible. Large language models are now deeply woven into search, and search is increasingly feeding back into AI systems. It’s a powerful combination, but it makes traceability and underlying data even more important. LLMs excel at giving you an answer – often the next logical answer. But discovery and verification still matter. The foundations behind every insight remain relevant, and will be even more so over the next 12 months,” said Thomas.
She said that the main divide will be how thoughtfully organisations apply AI, rather than whether they adopt it.
“So as 2026 unfolds, the dividing line won’t be between organisations that use AI and those that don’t. It will be between those who apply it thoughtfully and those who let AI guide them unquestioningly. The advantage will sit with teams that embrace transparency, develop prompt literacy, apply sound judgement and maintain control as they automate. AI won’t run the finance function. But it will elevate the people who do,” said Thomas.
Tax and agentic AI
In tax, vendors expect more practical deployments of agent-based AI systems within controlled workflows.
“Agentic AI will be one of the most significant shifts in tax over the next year as we start to see more meaningful applications of the technology. We are moving beyond experimentation to genuinely useful and practical deployments of AI agents into workflows to solve real problems. We also expect to see regulatory bodies, including HMRC, release guidance on the technology’s use, which will be critical in giving organisations the confidence to move forward, particularly in regulated and risk-sensitive areas like tax,” said Russell Gammon, Chief Innovation Officer at Tax Systems.
Gammon said early use cases will concentrate on lower-risk tasks.
“In the early stages, agents will be deployed in low-risk use cases where trust can be built gradually. Data is one area in particular where agents can add immediate value, freeing up significant amounts of time spent on input and number crunching without disrupting existing controls. As confidence grows, those same agents will start to take on more tasks and responsibilities. One area with significant potential is the review process. Agentic AI can support first and second-level reviews, flagging anomalies, highlighting risks and preparing summaries, which will save a significant amount of time for senior leaders, whose capacity is both limited and extremely valuable,” said Gammon.
He expects the impact to grow through 2026.
“There are many opportunities for tax processes to be made more efficient and in 2026, we’ll start to see those changes play out at scale, with agentic AI becoming a core part of how tax functions operate,” said Gammon.
Cost pressure
Broader economic conditions remain a central concern for finance chiefs as they plan for the year ahead.
“As we move into 2026, UK businesses will remain under intense cost pressure. Inflation may be easing, but prices are still rising, compounding the pressure of higher national insurance and other increasing employment expenses from 2025. Organisations can’t absorb those pressures indefinitely, so they will continue to shape business decisions, constraining investment and slowing recruitment,” said Bruce Martin, CEO of Tax Systems.
Martin said many firms will reassess complex processes that have built up over time.
“In this economic environment, businesses will need a sharper view of what genuinely creates value. After years of layering new policies and processes onto existing structures, many organisations are left with a knot of complexity that’s expensive to run and difficult to change. In this case, a reset is good business hygiene, and the most effective leaders will take a more disciplined look at which processes are essential and which can be removed altogether,” said Martin.
He said this review will link directly to investment in digital tools.
“Once that focus is established, business leaders must ensure that they are equipping their teams with the right tools. You wouldn’t send a construction worker to dig a hole with a teaspoon; yet we routinely do the equivalent in back-office functions by asking people to deliver more with outdated systems and manual workarounds. 2026 should be the year that organisations address that gap and make the investments needed to drive efficiencies. AI will be pivotal in this regard, enabling organisations to do more with less as budgets tighten. Importantly, however, it isn’t all about speed. The real opportunity with modern technology, particularly AI, is the ability to do things that simply weren’t possible before, unlocking new insights, new ways of working, and fresh avenues for innovation,” said Martin.
“In a year defined by cost pressure, organisations that streamline how they operate and equip their people with genuinely effective tools will be best positioned to make confident decisions and move forward with intent,” said Martin.
Fintech infrastructure
Market infrastructure providers anticipate that financial firms will keep prioritising latency, uptime and security as trading and payments volumes grow.
“Whether in trading, payments or currencies, fintech companies want competitive advantage. To do this they require speed and quick reaction time – so their choice of infrastructure provider is absolutely key,” said Terry Storrar, Managing Director at Leaseweb UK.
He said the operational demands on finance and trading companies leave little tolerance for disruption.
“Finance and trading companies rely heavily on uptime, latency, reliability, security and scale. All of which are non-negotiable when they are looking at where to place their workloads. Data changes in this industry in milliseconds and they are constantly looking at ways to improve performance and security for their clients. AI and ML technologies will undoubtedly be considered, but they would need to prove real productivity advantages to provide a competitive edge as this industry looks for new ways to stand out and doesn’t generally just follow trends for the sake of it,” said Storrar.
Cyber risk and RoM
Security specialists say finance leaders increasingly regard cyber incidents as a direct financial threat, rather than as a pure technology issue.
“2025 has shown that cyber risk is financial risk. In November, a supply-chain breach at real-estate finance vendor SitusAMC forced Wall Street banks, including JPMorgan, Citi, and Morgan Stanley, to scramble to understand how much mortgage and customer data had been exposed. At the same time, UK Finance estimates that almost £100 million was lost to AI-driven investment scams in just the first half of the year,” said Bertijn Eldering, Associate Sales Engineer at HackerOne.
Eldering cited the growing cost of breaches in the sector.
“Against that backdrop, the average cost of a data breach in financial services now sits at around $5.56 million per incident, before you factor in litigation, regulatory fines, and long-term reputational damage. But it doesn’t have to stay on this trajectory. As we move into 2026, financial institutions will need to treat cyber spend less as a sunk cost and more as core protection for earnings, capital, and customer trust,” said Eldering.
He said recent experience shows that external security researchers can contribute to risk reduction.
“Over the last 12 months, security researchers have proven they can help firms raise their defences by finding and mitigating vulnerabilities before cyber criminals get there. The latest Hacker-Powered Security Report revealed that these researchers saved businesses a total of $128 million in the last year alone. When it comes to cybersecurity, return on mitigation (RoM) should be front and centre for cyber-defence teams,” said Eldering.
Eldering said both defenders and attackers are increasing their use of AI.
“Inevitably, as technology develops, AI integration will stay at the forefront of business decision-making, with 41% of businesses already testing AI assets as part of their work. At the same time as banks embed AI into fraud operations, underwriting, and customer service, attackers are using the same tools to industrialise phishing, deepfake scams, and vulnerability discovery,” said Eldering.
He expects this contest to intensify in 2026.
“In 2026, AI-driven cyber activity will scale on both sides of the fence. Financial institutions must move forward with a RoM-focused mindset. Those that fail to put offensive security testing of their AI and third-party attack surface on an equal footing with product innovation will increasingly find themselves on the wrong side of the next headline,” said Eldering.