PLATFORM · POLICY ENFORCEMENT

Rules checked before approval.
Not after payment.

REME enforces your expense policies at the moment of submission — not weeks later during quarterly audit. Threshold caps, category limits, receipt requirements, approval routing, and exemption workflows, all checked in real time before a claim moves forward. The result: 95% of expenses arrive in your accounting system already in policy.

Book a demo See pricing
Real-time enforcement Configurable rules engine SOC 2 Type II ISO 27001

Most "policy enforcement" enforces nothing. It documents.

Legacy expense tools enforce policy through audits. An employee submits an out-of-policy expense, the approval flow lets it through, payment processes, and weeks later, during quarterly review, finance flags it. By then, the money is gone. The "enforcement" was actually documentation — a record of what wasn't caught in time. Pre-transaction enforcement is a different category: rules checked before payment, exceptions handled before money ever moves.

Audit-time enforcement (legacy)
✗ Employee submits expense — policy not checked
✗ Approver clicks "approve" without rule context
✗ Payment processes
✗ Weeks later: quarterly audit flags violation
✗ Finance opens recovery conversation — money already spent
✗ Result: documentation, not enforcement

Time from violation to detection: 30–90 days

Pre-transaction enforcement (REME)
✓ Employee submits — policy engine checks instantly
✓ Approver sees rule context with the claim
✓ Out-of-policy expenses flagged before approval
✓ Exemptions go through structured workflow with justification
✓ Payment only processes for in-policy claims
✓ Result: enforcement, not documentation

Time from violation to detection: 0 seconds

14% of submitted expenses

are non-compliant on average across mid-market organizations. With audit-time enforcement, most are caught only during quarterly review — weeks after payment cleared. Pre-transaction enforcement catches them at the moment of submission, before money ever moves.

Source: Forrester research commissioned by Navan, 2026.

Five rule categories. One real-time engine.

REME's policy engine runs five categories of checks at the moment of submission. Configure them once; they enforce automatically across every claim.

01

Threshold rules

The most fundamental category — caps and floors on amounts. Policy engine compares the submitted amount against the configured threshold for that expense category, employee role, or trip context. Violations flag automatically; exemptions route to designated approvers with justification.

EXAMPLES

  • · Food expenses capped at $50/meal for individual contributors, $80 for managers
  • · Hotel rates capped at $200/night, $300 in tier-1 cities
  • · Mileage reimbursement capped at IRS standard rate ($0.67/mile in 2026)
  • · Single-claim ceiling of $5,000 — anything above routes to CFO

02

Category rules

Some expense categories require additional documentation, justification, or approval. The engine identifies the category from receipt OCR and applies category-specific rules — no employee tagging needed.

EXAMPLES

  • · Alcohol requires meal context (who, why, how many people)
  • · Personal entertainment subject to client/team trip justification
  • · Software subscriptions require IT pre-approval
  • · Travel expenses require trip booking reference

03

Documentation rules

Receipt requirements vary by amount, category, and jurisdiction. The engine checks whether a receipt is attached, whether it's legible, whether it matches the claim amount, and whether it includes the data your jurisdiction requires for tax recovery.

EXAMPLES

  • · Receipts mandatory for all claims above $25
  • · Tax invoices required for VAT recovery in EU/UK
  • · GST tax breakdown required for India and Singapore claims
  • · Receipt date must be within 30 days of submission

04

Approval routing rules

Different claims need different eyes. The engine routes each claim based on amount, category, employee, project, or department — and escalates automatically if the first approver doesn't respond within configured SLAs.

EXAMPLES

  • · Up to $500: direct manager approval
  • · $500–$2,000: department head approval
  • · $2,000+: CFO approval
  • · Client entertainment: account manager + CFO dual approval
  • · Auto-escalate to backup approver after 48 hours

05

Timing & freshness rules

When did the expense occur, when was it submitted, and is the claim period still valid? The engine flags late submissions, claims outside fiscal periods, and stale receipts that violate your retention policies.

EXAMPLES

  • · Submission deadline: within 14 days of expense date
  • · Receipts older than 60 days flagged for explanation
  • · Claims for closed fiscal periods automatically blocked
  • · Holiday/weekend submissions allowed but routed differently

How the engine works — from submission to approval

Every expense submitted via WhatsApp passes through this flow. Sub-second latency. No employee training required.

01

Submit

Employee sends receipt photo via WhatsApp. OCR extracts amount, merchant, date, tax breakdown.

02

Identify category

AI categorizes the expense. Looks up applicable policy rules for this employee, this category, this jurisdiction.

03

Run rule checks

Engine evaluates all 5 rule categories in parallel. Threshold, category, documentation, routing, timing.

04

Decision

Pass → routes to approver. Flag → routes to approver with rule context. Block → returned to employee with explanation.

05

Audit log

Every check, every result, every approval — logged automatically. Full audit trail without finance team involvement.

Total latency from submission to decision: typically 200–400 milliseconds. The employee sees the result before they put their phone down.

Configure policies in plain English. Or import what you already have.

Most policy engines force you to translate your written policy into rigid rule trees and conditional logic. REME does it in plain English — describe the rule the way you wrote it in your employee handbook, and the engine configures itself. Already have policies in Concur, Expensify, or Spendesk? Export the JSON and import directly.

  Add a policy rule

"Cap food expenses at $50 per meal during business trips, except client dinners which are unlimited but require manager approval above $200."

INTERPRETED RULE

✓ Category: Food (when trip context = business)
✓ Threshold: $50/meal default
✓ Exception: Category = "client dinner"
✓ Conditional approval: client dinner > $200 → manager review

Exemptions are the hard part. Most tools punt on them.

Real-world expense flows have exceptions. The CEO takes a client to a $300 dinner — outside the $50 food cap, but legitimate. A field engineer needs a same-day flight that violates the 14-day advance booking rule — necessary for the customer escalation. Most policy engines block these or route them through generic approval flows that frustrate everyone. REME handles exemptions as a first-class workflow.

EMPLOYEE FLOW

Submit with justification

When the engine flags a violation, the employee sees the rule, the gap, and a justification field — right inside WhatsApp. They explain ("client dinner with VP of procurement at Acme") and submit. No app to open, no separate workflow.

APPROVER FLOW

Approve with full context

Approvers see the claim, the rule it violates, the employee's justification, and historical patterns ("this is the 2nd exemption this quarter"). Decision happens with full context, not blind approval. One-tap approve, deny, or request more info.

AUDIT FLOW

Trail without effort

Every exemption is automatically logged with: rule violated, justification provided, approver identity, decision timestamp, and surrounding pattern data. Quarterly exemption review takes minutes, not hours.

Policy enforcement and fraud detection — two layers, one platform.

Policy enforcement catches what the rules say is wrong. Fraud detection catches what the rules don't say but should. Together they cover the full spectrum.

LAYER 1

  Policy Enforcement

What it catches:

  • • Threshold violations (over the cap)
  • • Missing receipts on required claims
  • • Out-of-jurisdiction tax submissions
  • • Late submissions outside grace period
  • • Claims for closed fiscal periods

Speed: 200–400ms at submission

Action: Block, flag, or route to approver

You're on this page →

LAYER 2

  AI Fraud Detection

What it catches:

  • • Duplicate submissions across angles
  • • AI-generated fake receipts (ChatGPT-made)
  • • Merchant fabrications (vendors that don't exist)
  • • Anomalous employee patterns
  • • Receipt alterations and edits
  • • Round-number amount manipulation
  • • Cross-employee collusion patterns

Speed: 200ms across 7 AI agents

Action: Block, flag for review, or surface to security

Read the full deep dive → AI Fraud Detection

Both layers run on every claim. In parallel. Before any approval.

Policies that flow into your accounting stack

Policy decisions don't live in isolation. REME pushes the full enforcement record — claim, rule check result, exemption notes, approver identity — directly into your accounting and ERP system. Audit trail by default.

QuickBooks Online
Xero
NetSuite
Sage Intacct
Microsoft Dynamics 365
Oracle ERP Cloud
SAP S/4HANA
QuickBooks Desktop
2-way sync Real-time posting Automatic GL coding Exemption notes preserved

FAQ

Policy enforcement — common questions

Your audit committee asks: "Why did this slip through?" The answer should be: "It didn't."

Enforce before, not after.

Real-time policy checks at submission. Plain-English rule configuration. Five rule categories covered. Exemptions handled as first-class workflows. Backed by our adoption guarantee — if your team doesn't hit 80% in 30 days, we waive the next 60 days of paid usage.

Book a demo See pricing
Real-time enforcement 5 rule categories Plain-English config 80% adoption guarantee

80% adoption guarantee

If your team doesn't reach 80% submission adoption within 30 days, we waive the next 60 days of paid usage. No small print.