Claude tops the AI tools finance teams use while finance-built rivals are still emerging

Most finance teams still build the forecast in a spreadsheet. Four in five plan and forecast mainly in one, and a further 7 percent have no planning tool at all ^[14]. That means 87 percent of teams either plan mainly in spreadsheets or have no planning tool ^[2]. The same share, 87 percent, runs payroll and HR on a dedicated system ^[18].

The uses teams report for AI sit on that kind of work. Financial analysis leads at 20 percent, followed by reporting, modeling and forecasting at 12 percent each, reconciliations at 9 percent, and data queries and workflow automation at 8 percent each ^[6]. According to the report, AI lets teams search spreadsheets, finance software and documents, and get routine tasks done faster ^[8].

A finance-AI vendor pitches a product built for one finance task. The report's findings cover two different behaviours ^[4][7]. One is an analyst opening a general assistant next to the ledger and working through a workbook. The other is a controller switching on AI features inside accounting software the team already runs. Both count against the specialist vendor. The first competes on breadth. The second is already installed and needs no procurement round.

Software lists are already short. Over 90 percent of teams run six or fewer tools, with 41 percent on one to three and half on four to six ^[12]. A July Deloitte survey of 58 large-company CFOs in the UK found 73 percent optimistic about AI, though their spending stayed cautious ^[9]. "It's obvious that finance teams aren't chasing more software," said Pauline Babel, chief financial officer of Spendesk, which runs CFO Connect ^[10][2]. Babel also said teams want one joined-up place to work ^[11].

The areas with no system to add AI to are where a dedicated finance product has an opening. Almost two-thirds of companies have no dedicated procurement tool ^[16]. For cash, 55 percent track it in spreadsheets and 17 percent use no dedicated system ^[15]. In both areas, what a finance-built product would replace is a spreadsheet or nothing at all ^[15][16].

For the finance lead picking tools, two axes sort most of the decision. The first is where the work lives: in software the team already runs, or in a spreadsheet. The second is whether the AI only needs to read and summarise, or has to change records. Read-only work inside existing software goes to that software's own AI features. For spreadsheet work that only needs reading, such as analysis and first-draft forecasts, a general assistant is the cheap test. The reported uses already cluster there. Reconciliations and other record changes inside existing software are a question for the current vendor first. That leaves spreadsheet work that changes records, cash and procurement among it, as the quadrant where a finance-built tool has a case on this evidence.

I'd start with the general assistant on spreadsheet work, because it adds nothing to a software list the team is trying to keep short. The tradeoff is how thin the evidence behind it is. The survey counts which tools 215 respondents use and for what, in a sponsored report ^[2][3]. It does not measure time saved, error rates or whether teams kept using a tool after trying it.