Finance teams demand auditable AI & stronger data in 2026
Finance and accounting leaders expect artificial intelligence in 2026 to come under tougher scrutiny, with a focus on auditability, data quality and demonstrable productivity gains, according to John Phillips, General Manager EMEA at FloQast.
Phillips said finance teams now view AI tools through an audit lens. He highlighted concerns about the risk of deploying AI systems that cannot show how they reach their outputs.
"AI that doesn't stand the audit test is too much of a liability for finance and accounting teams to consider. Cutting the time spent on manual tasks only pays off if the outputs can be traced and checked. When they can't, efficiency gains disappear as people fall back on manual recalculations.
"Teams require a clear view of what the agent did, what data it pulled, and why it reached a particular answer. Baking that transparency in from day one will earn the most trust. Those who don't will struggle, because it is too risky to stake reports and filings when based on guesswork," said Phillips, General Manager EMEA, FloQast.
Vendors and internal technology teams in finance now face higher expectations over explainability and traceability. Audit and compliance requirements sit alongside traditional goals such as speeding up reconciliations or financial close processes.
Changing finance roles
Phillips said AI will reshape how accountants interact with IT systems and how finance projects run. He described a shift away from traditional models where technology specialists led transformation programmes.
"Finance transformation used to be driven by people who understood IT systems but not the everyday pressures of accounting. AI closes this gap.
"In 2026, accountants will be able to shape automation in their own language without long technical hand-offs. Finance teams will start to build agents to support their workload, which will call for new skills across the function.
"As these capabilities grow, finance and accounting roles will start to shift from documenting requirements to actively engineering solutions, using an understanding of the workflow that only accountants have," said Phillips.
This shift places more design and configuration work inside finance teams. It also raises questions over training, governance and how organisations balance specialist IT oversight with growing in-house experimentation in accounting departments.
Data foundations
Phillips said data quality will remain a structural constraint in 2026 despite rapid progress in AI tools. He pointed to the complexity of enterprise resource planning systems in mid-market and larger organisations.
"Everyone talks about full automation, yet the real blocker in 2026 won't change: getting clean, usable data out of the systems that matter. Most ERPs have been customised over years to reflect how a business really operates, meaning no two data sets behave the same.
"AI will speed up decision-making, but only when the underlying information is consistent, accessible and well-organised. That's why more mid-market firms will move towards central data warehouses giving them one place to pull from and reducing surprises when they start building AI-driven workflows.
"In 2026, the teams that invest in their data foundations will be the ones who see meaningful automation," said Phillips.
His comments reflect a broader move among finance departments towards data engineering projects. Many teams now consolidate information from multiple ERP instances and operational systems before they deploy AI tools for close management, forecasting or compliance work.
'Prove it' mindset
Phillips expects a more sceptical buyer in 2026 as finance leaders test AI promises against their own data. He said this shift is altering how AI products are evaluated and procured.
"After all the hype, it's natural that teams in 2026 will be more cynical about AI's benefits. They'll want to see the productivity gains for themselves by running their own data through a tool before they commit, and that expectation is reshaping how AI products are sold.
"Sales cycles will lengthen because buyers want hands-on support, transparent benchmarks, and real proof that the product fits their workflow. This pressures vendors to show substance and helps buyers see what AI can actually deliver. This 'prove it' mindset will become the default in 2026 - less 'trust me' and more 'show me'," said Phillips.
The focus on demonstrable outcomes suggests finance teams will place more weight on pilot projects and controlled roll-outs. Vendors face longer evaluation periods and closer inspection of how their tools handle workflows, audit requirements and existing data environments.
