AI and data reshape finance, tax and hiring practices

Finance and technology leaders are adopting artificial intelligence tools across compliance, hiring and operations. At the same time, practitioners are urging businesses and individuals to use new data more actively in tax and planning decisions.

In the UK, Financial Planning Month is drawing attention to the impact of Making Tax Digital on landlords and sole traders, alongside the growing use of AI in recruitment and post-trade operations. Commentators from Taxd, Nexteam and Duco describe a shift towards more frequent use of data while warning against overconfidence in automated outputs.

Arjun Kumar, Co-Founder and Co-Chief Executive Officer of Taxd and a former PwC accountant, said landlords and self-employed workers still ask basic questions about quarterly submissions under Making Tax Digital. Under the rules, these groups will need to log income and expenses four times a year instead of once. Each submission is a summary of digital records rather than a full tax return.

"A lot of landlords and sole traders want to know what quarterly reporting under MTD actually involves. Put simply, it means logging your income and expenses four times a year instead of once. These quarterly submissions aren't full tax returns; they're summaries of the digital records you're already keeping. The bigger opportunity is what that regular view makes possible. Most people only look at their tax once a year, when it's too late to change anything.

Every pound you put into a pension or claim as an allowable expense reduces your tax at your highest rate, so for higher earners, a bit of planning can mean tax relief of 45% or more. We've built our services to flag the reliefs and allowances people didn't realise applied to them. October is a good time to get ahead of MTD and check your ISA and pension allowances properly rather than guessing," said Arjun Kumar, Co-Founder & Co-CEO, Taxd.

Tax specialists see the new reporting cycle as a way to bring retirement saving and expense claims into sharper, more regular focus. Their comments come as financial planners highlight unused pension and ISA allowances among many higher earners.

Hiring practices in finance teams are also changing as recruiters add AI summarisation tools to established processes. Nexteam, which recruits for finance and operations roles, uses machine learning systems to structure information and highlight relevant experience from candidate records.

"The distinction I care about is between organising evidence and inventing certainty. An AI summary can make a finance candidate's experience easier to compare, but it should not turn a job title into proof that someone has owned a close, built a forecast or managed a team. We still need to verify those points through interviews, assessments and discussion of the work.

My advice is to keep three things separate: what the record supports, what is missing and what a person still needs to check. A clearer shortlist is useful; a more confident-looking shortlist is not necessarily more accurate," said Sergio Ermacora, Co-Founder & Chief Operating Officer, Nexteam.

Ermacora's comments reflect concerns that hiring managers may treat AI-generated summaries as conclusive. Recruiters report faster shortlisting, but human validation remains central for senior finance roles.

At the other end of the financial workflow, Duco Founder and Chief Executive Officer Christian Nentwich links recent KPMG research on agentic AI in finance to activity in post-trade operations. The KPMG report finds that financial services and technology organisations are deploying agentic AI more broadly than other sectors.

"The figures show financial services is already ahead on agentic AI, and we are starting to see that translate into live operational use. In post-trade, agents can take on work that still consumes huge amounts of human time today, from triaging exceptions and identifying root causes to spotting recurring problems and proposing fixes. The important point is that this does not mean replacing deterministic controls.

Agents can do the reasoning, while rules and approval processes make sure actions remain controlled and auditable. As that model becomes more common, firms will need to rethink how operations teams are structured, with people focused less on routine investigation and more on the cases where judgement and intervention are genuinely needed. This need becomes more urgent with the move to T+1 next year, as shorter settlement cycles leave firms less time to investigate and resolve exceptions manually," said Christian Nentwich, Chief Executive Officer & Founder, Duco.

Market participants are preparing for the move to T+1 settlement cycles across major markets. Operations leaders face pressure to handle exceptions more quickly while keeping audit trails and approvals under tight control.

Taken together, the comments from Kumar, Ermacora and Nentwich point to a common theme: more data and automation in finance, paired with a greater need for active human judgement. Recruiters, operations teams and taxpayers all face new information flows and shorter decision cycles.