Don’t Ignore the Threat: AI-Generated Receipts and the AP Control Gap
As accounts payable professionals, we’ve long known that fraud is a rising
risk in AP and expenses. What’s new, and alarming, is how generative
artificial intelligence (AI) is fundamentally shifting the risk landscape.
What used to require photo-editing skill or insider access now takes seconds
with text prompts. In many cases, fraudulent expenses can now pass through
reimbursement before anyone has time to look closely. That raises a serious
question: is your AP and expense control environment keeping up?
What’s happening now
-
According to vendor-reported data from AppZen, AI-generated receipts
represented approximately 14% of detected fraudulent
documents processed on its platform in September 2025, up from near zero the
year prior. [2]
-
Surveys show that nearly 70% of CFOs believe it is likely,
or already confirmed, that employees are using AI tools to falsify travel
and expense receipts. [3]
-
Third-party AP and expense platforms report flagging over
US$1 million in suspected fraudulent invoices and expenses
in a 90-day period using AI-enabled detection systems during late-2025 pilot
deployments. [4]
These figures are being reported by vendors operating at scale, processing
millions of expense and invoice submissions annually, which provides early
visibility into AI-driven fraud patterns before they appear in traditional
loss statistics.
What’s striking is the low barrier to entry. No advanced Photoshop skills are
required anymore. A user can prompt a generative-AI model (such as those from
OpenAI or Google LLC) to output a receipt image with realistic textures,
logos, timestamps, and even signatures. [7]
Why this matters for AP teams
Here are some of the direct implications for AP and expense processes:
-
Traditional receipt review based on visual inspection alone is increasingly
unreliable. As one controls leader put it,
“Do not trust your eyes.” [8]
-
Expense reimbursement processes sit squarely within AP or AP-adjacent
workflows. Fraud here translates into direct financial losses, audit
findings, and internal control failures.
-
The growing sophistication of fake receipts elevates regulatory, tax, and
compliance risk, including false business-expense reporting, improper tax
deductions, and policy violations.
-
As AP teams migrate toward automation and digital workflows, the attack
surface expands. More electronic submissions, remote approvals, and faster
processing can reduce time available for manual controls.
-
For APPG members advising or operating in mid-market firms (10–30+ invoice
or expense submissions per month), the assumption that small size equals low
risk no longer holds. AI-enabled fraud scales quickly and cheaply.
It is also important to note that AI-detection tools are not perfect.
Screenshots can strip metadata, some generators leave fewer artifacts, and
human judgment still plays a role. Effective defense increasingly requires a
hybrid approach combining technology, analytics, and informed review.
Key control gaps AP teams should assess today
Here are control areas AP leaders should audit or strengthen:
| Control Area |
Risk Gap |
Remediation Focus |
| Receipt and image verification |
Visual inspection only; no metadata or image-artifact checks |
Use software that analyzes image metadata and pixel-level artifacts to
flag suspicious patterns. [9]
|
| Expense submission workflows |
Delayed submission and weak policy enforcement |
Require receipt upload at time of expense, enforce corporate card usage,
and restrict cash reimbursements.
|
| Approval and vendor verification |
Approvers lack vendor familiarity; fictitious vendors go unnoticed
|
Strengthen vendor master controls, review high-frequency expense
vendors, and monitor items outside normal patterns.
|
| Data analytics and monitoring |
Spot audits only; anomalies remain undetected |
Implement analytics to flag repeated vendors, out-of-band amounts, rapid
post-trip submissions, and shared image metadata.
|
| Audit and detective controls |
Fraud discovered after reimbursement |
Deploy real-time alerts alongside retroactive sampling and integrate
findings with AP and risk teams.
|
What AP teams can do now (90-day action plan)
Here is a practical action plan AP leaders can execute in the next 90 days:
-
Schedule a focused control-risk review addressing AI-generated receipts and
document authenticity within the AP and expense framework.
-
Inventory expense submission channels, including paper versus digital,
number of reviewers, and payment methods such as corporate cards, P-cards,
or personal reimbursements.
-
Evaluate AI-detection or image-artifact analysis capabilities now available
in many AP and expense platforms. [10]
-
Update expense and vendor policies to reinforce receipt requirements,
defined submission time windows, corporate card usage, and random audits of
high-risk claims.
-
Build analytics dashboards to monitor unusual patterns such as frequent
small vendors, clustered submissions, or receipts sharing similar metadata
or image characteristics.
-
Communicate clearly with AP teams and business units that AI-driven document
fraud is a priority risk area for FY26, and that approvers are a critical
part of the control environment.
-
Conduct a retrospective audit focused on high-risk expense categories, such
as cash reimbursements, repeat vendors, and frequent low-dollar claims,
rather than attempting a full historical review.
Why this matters for APPG members
For professionals working in accounts payable and expense management, this
issue touches core APPG themes: process integrity, risk management,
automation, and advisory value for internal stakeholders and clients.
As generative AI continues to improve in realism and accessibility through
2026 and beyond, staying reactive is no longer enough. Proactively
strengthening AP controls around document authenticity is quickly becoming a
core competency for modern AP teams.
Discussion prompt for the APPG community
Let’s turn this into a practical discussion. I invite APPG members to respond:
-
When was the last time your organization reviewed its expense-reimbursement
controls specifically for fraud and document authenticity?
-
Do you currently use any tool or process to detect AI-generated or
manipulated receipt images? If yes, what works; if not, what is the barrier?
-
What is the biggest manual bottleneck in your expense workflow, and how
might it be increasing fraud risk?
If you would like to help build an APPG peer checklist or benchmark on AP and
expense fraud controls, reply below or send me a DM and we will organize a
short member survey.
Thanks for reading. Let’s stay ahead of the fraud curve and continue elevating
the strategic value of AP.
💡 Support Our Community:
Our sponsor helps AP teams reduce errors and speed approvals (fewer late
fees and happier vendors).
Want tailored guidance on AP automation and vendor spotlights?
Book a 15 minute chat
|
Robert Ruhno
Director
Accounts Payable Professionals Group
|
|
Back to top ↑