What should finance teams automate first? Start with receipts, coding and reconciliation

Think about everything that has to happen after an employee makes a purchase.

The receipt has to get to the right place, the transaction needs to be matched, coded and reviewed and the completed information has to make its way into the accounting system. If something is missing, someone has to track it down. 

It’s mundane work. But multiply it across hundreds or thousands of transactions and it becomes a significant drain on finance teams. According to PEX’s State of Finance 2026 report, 31% of finance teams spend more than 20 hours a month collecting receipts, coding transactions and reconciling expenses. Yet 67% automate 20% or less of their transactions. 

That makes receipts, coding and reconciliation a practical first target for automation. Receipts and reconciliation are the most commonly cited finance operations challenge, named by 30% of respondents, and the report identifies receipt capture and transaction categorization as the lowest-lift place for teams to begin with AI. 

If the technology exists to automate more of this work, why aren’t finance teams using it? Trust is one reason. Thirty-six percent of finance leaders name trust in AI accuracy as their biggest barrier to expanding AI, more than any other barrier in the survey. 

But finance teams don’t have to hand an entire process over to AI at once. They can start with one everyday workflow, deciding which routine steps to automate, what employees should own and where finance needs to review or step in.

Start with the workflow

Start by mapping what needs to happen from the moment someone makes a purchase until that transaction is ready for reconciliation:

Purchase → receipt capture → matching → coding → review → accounting

From there, you can identify where automation can keep routine transactions moving and where people still need to be involved.

What an automated transaction workflow looks like

1. The employee makes an approved purchase. Pre-set spend controls can define where, when and how funds can be used, giving employees clear parameters before they spend.

2. The employee submits the receipt. They submit it through the available receipt capture method soon after the purchase, putting responsibility for documentation closer to the person who made the purchase.

3. AI matches the receipt to the transaction. Receipt-matching technology connects the submitted receipt with the corresponding transaction, reducing manual matching for finance.

4. AI suggests the category or GL code. Based on the transaction information, AI suggests a code that the employee reviews and confirms.

5. Finance reviews the completed transaction. With the receipt and coding already attached, the team can review the record rather than assemble the information itself.

6. The transaction moves into the accounting workflow. Transaction data flows into the accounting system, reducing manual data entry and keeping expense information moving throughout the month.

7. Exceptions take a different path. If a receipt is missing, required information is incomplete or a transaction needs additional review, it can be flagged and routed to the appropriate person. Once the issue is resolved, the transaction can continue through the workflow.

Configure the workflow

For that model to work, finance has to define the rules before the workflow goes live. 

Start with a few questions:

• What information is required? Define what employees need to provide, such as a receipt, expense category, GL code or other transaction details.
• Who owns each step? Decide what the employee is responsible for, what automation or AI can handle and when finance needs to get involved.
• What requires human review? Determine which transactions can move forward once required information is complete and which need an additional approval or review.
• What happens when something is missing? Establish how exceptions will be identified, who is responsible for resolving them and what needs to happen before the transaction can continue.
• Where does the completed information go? Determine how reviewed transaction data will move into the accounting system or other downstream workflows.

These decisions become the guardrails for automation, allowing routine work to move with less intervention while preserving finance’s visibility and control.

Establish a baseline

Before implementing automation or AI, document how the process works today. You don’t need an exhaustive analysis: just enough information to tell whether you’re reducing the manual work you set out to address.

Look at a few practical measures that are meaningful to your team: 

• How many transactions are missing documentation? 
• How often does finance have to step in? 
• How much time does the team spend coding or correcting transactions? 
• How many exceptions occur and what causes them? 
• How long does reconciliation take?

Measure the first 30 days

After the workflow has been running for a month, return to the same measures to see whether it’s reducing the work required to process everyday transactions.

Compare the first 30 days with your baseline. Are fewer transactions arriving without documentation? Is finance spending less time coding or correcting transactions? Are fewer routine transactions requiring manual intervention? Are exceptions being resolved earlier? Is reconciliation taking less time?

You may also learn where the workflow itself needs adjustment. If the same exception keeps occurring, for example, that could point to an unclear requirement, a missing control or a step that needs a different owner.

If the workflow reduces manual effort without sacrificing visibility or control, you’ve proved its value. From there, you can decide where automation could help next. That pattern of starting small and expanding shows up in PEX’s research: teams scaling AI average 3.7 live capabilities, compared with 0.5 among teams that haven’t started. And 59% of scaling teams report ROI, compared with just 3% of teams that haven’t started.

How PEX helps automate everyday transaction work

PEX brings the steps of the transaction workflow together, so information can move from the Cardholder to finance and into the accounting system with fewer manual handoffs.

• Capture and match receipts. Cardholders can submit receipts through the PEX mobile app, text or email. AI-powered receipt matching connects receipts with the corresponding transactions, reducing the manual work of tracking down and matching documentation.
• Speed up transaction coding. Finance teams can use pre-programmed GL code mappings to automatically associate transactions with the right codes. AI-powered GL coding suggestions can also recommend codes based on historical transaction data.
• Build controls and approvals into the workflow. PEX spend controls can set parameters around specific merchants, spending amounts, locations and time periods. Multi-level approval workflows give finance teams another layer of oversight when a transaction or request requires review.
• Focus attention on exceptions. PEX can surface transactions with missing receipts or other required information. Finance teams can identify what needs attention without checking every transaction for missing information.
• Automatically sync data to accounting tools. PEX offers 50+ pre-built accounting integrations that automatically sync transaction data with accounting systems, reducing manual data entry and helping keep records up to date.

Case in point: Compass to Care

Compass to Care provides free transportation assistance to families traveling for childhood cancer treatment and clinical trials. The nonprofit needed a way to distribute funds quickly while maintaining visibility and control over spending.

With PEX Visa® Prepaid Cards, Compass to Care can set merchant-specific spending controls, while families submit receipts by text or email. AI-based receipt matching and real-time transaction reporting reduce the administrative work required to track spending.

Compass to Care reports a 95% increase in efficiency in fund management and accounting and saves more than two hours per week on card funding. PEX has also helped the organization scale its programs without significant increases in staffing.

Make manual intervention the exception

The payoff from automating everyday transaction work is more time for the work that requires finance’s expertise.

When routine transactions arrive with the documentation, coding and approvals they need, finance doesn’t have to touch every step. The team can focus its attention on exceptions, decisions and higher-value work.

That’s where AI and automation can create capacity without sacrificing control.

Explore more findings on how finance teams are using AI and automation in the PEX State of Finance 2026 report.