THE MECHANISM
How REME finds, blocks, and shrinks expense leakage
Three components working together. This page explains each one in the depth CFOs and finance leaders need to evaluate REME properly. Read in 5 minutes or skim the section headers.
Six AI agents · 90-day baseline · 9-month reduction target · Dashboard-tracked outcome
What REME actually is
Most expense platforms are receipt scanners with approval workflows. REME does that too, but it's not the point. REME's job is to measurably reduce the money leaking from your expense process, and to prove that reduction over time on your own dashboard.
Three components work together to make that happen:
- 1
Frictionless submission (via WhatsApp, email, or web upload) so 100% of your team's expenses actually enter the system.
- 2
Real-time AI fraud detection (six specialized agents) so no fraudulent claim gets through the approval gate.
- 3
Measured leakage reduction (baseline in 90 days, reduction in the next 9 months) so you have an outcome to prove and track.
Component 1 is the ease of use. Component 2 is the technical capability. Component 3 is the business outcome. Every successful REME deployment relies on all three.
FIND
Your first 90 days
The baseline is your true leakage rate, measured on your own spend, without any historical data required.
Traditional consulting engagements ask you for 12 months of clean historical expense data before they can benchmark anything. Most mid-market companies can't produce that. Even when they can, benchmarks against 'industry averages' don't tell you your company's specific leakage rate.
REME solves this with a different approach: measure attempts, not historical catches. From day 1, every expense claim your team submits gets screened and logged. Even blocked ones count. By day 90, you have your baseline: your company's leakage rate expressed as attempted leakage per $100k of submitted spend.
What gets measured
- Every submitted claim, logged with metadata
- Every blocked claim (attempts caught by AI, still logged)
- Category concentration (which fraud types drove the attempts)
- Employee or department concentration
- Vendor concentration
- Total leakage rate expressed per $100k of submitted spend
The adoption caveat, stated honestly
Your baseline accuracy depends on your team actively submitting during those 90 days. If only half your team is on REME, your baseline reflects half your team's behavior. We work with you during onboarding to drive full team adoption typically within the first 4-6 weeks so the baseline is accurate for your whole company by day 90.
If adoption is genuinely slow, we can extend the baseline window rather than deliver a baseline based on partial data. The 90-day timing is the target. An honest baseline is the priority.
The Day-90 Leakage Report
At day 90 you receive the Leakage Report: a generated PDF containing your baseline number, your category concentration, your employee/department/vendor patterns, and your 12-month reduction target. It's designed to be forwarded to your board, audit committee, or CFO's leadership team as a single-page summary of your leakage situation and REME's plan to fix it.
The Day-90 Leakage Report is REME's signature moment. It's the artifact that proves the mechanism works on your specific data. From here forward, you're comparing your live rate against this baseline every quarter.
BLOCK
Every claim, checked in under 200 milliseconds
Six specialized AI agents work in parallel on every submitted claim. Fraud gets blocked before approval, not caught weeks later during audit.
Every expense claim submitted through REME gets checked by six AI agents simultaneously. Each agent catches a different fraud pattern. Combined, they cover the major categories of expense fraud we've seen across our customer base.
Duplicate detection
Same receipt submitted twice, from different angles, or after minor modification. Image forensics catch what human reviewers miss.
Learn more →Handwritten claim validation
Handwritten receipts with confidence scoring on the extracted amount. Inflated amounts flagged. Cross-checked against employee history and vendor patterns.
Learn more →Currency mismatch detection
Receipt in one currency, claim in another. Exchange rate validated against transaction date, not today's rate.
Learn more →Out-of-country claim flagging
Claims from locations the employee wasn't authorized to be at. Cross-checked against travel authorization and role scope.
Learn more →Disallowed multi-currency validation
Claims in currencies your policy doesn't authorize for this employee or role. Configurable per policy.
Learn more →Data mismatch detection
Receipt shows one amount, claim shows another. Vendor names don't match. Dates inconsistent. Every field cross-checked.
Learn more →All six agents run in parallel in under 200 milliseconds. When any agent flags a claim, finance sees the specific flag reason with the evidence. Legitimate exceptions get approved with a note. Actual fraud gets blocked before approval.
Plus your policy, enforced automatically
Beyond AI pattern detection, REME has a configurable controls layer. Your finance team sets rules the AI enforces on every claim. Two controls live today (high-risk vendor rules and auto-approval). New controls added quarterly based on customer requests (category ratios, location-aware, time-of-day).
SHRINK
Cut your leakage rate by more than 50%
The 9 months after your baseline are where the ROI actually happens. Two mechanisms work together to drive down your leakage rate.
Detection is the feature. Deterrence is the ROI. Here's what that means in practice.
Mechanism 1: AI detection keeps working
Our six AI agents keep catching attempts every day. As your baseline data grows, our AI learns your specific vendor patterns and gets more accurate. Detection catches more of what does happen.
Mechanism 2: Deterrence changes behavior
This is where most of the reduction comes from. Once your team knows every claim is checked, most stop trying. Fraud attempts drop because deterrence is the strongest fraud control there is. The dashboard tracks the falling attempt rate month by month.
The 9-month target
By month 12 (day 90 + 9 months), most REME customers see leakage rates that are less than half of their day-90 baseline. That's the target we design the platform to achieve.
No legal guarantee, honest positioning
We don't offer a formal legal guarantee on the 50% reduction. Reduction depends on how your team responds to knowing every claim is screened, which varies by company culture. What we do provide: transparent dashboard tracking, quarterly reviews with your account team, and configurable controls your finance team can tune as patterns emerge. If your team is engaged, 50%+ reduction in 9 months is what we design for.
The dashboard tracks it live
Your leakage rate curve is visible on your dashboard from day 1. Baseline gets drawn at day 90. From that point forward, every quarter you see whether your rate is on the reduction trajectory. If it's not, the dashboard tells you why (which categories, which departments, which employees) so your finance team can act.
What you actually see on the dashboard
The Insights dashboard is where the Find/Block/Shrink story becomes visible to your finance team every day. Key elements:
- Live leakage rate meter (per $100k spend or per 100 claims)
- Baseline line drawn at day 90
- Reduction curve tracking rate over time
- Category breakdown showing where leakage concentrates
- Employee/department heat map of high-attempt patterns
- Vendor watchlist showing repeatedly-flagged vendors
- Savings tile showing leakage stopped (hard dollars)
- Time saved tile showing automation impact (hours saved)
- Configurable alerts for finance team when patterns change
[Insert dashboard screenshot here when marketing team provides]
Common questions about how REME works
Then your baseline reflects only the portion of your team that's active. We work with you during onboarding to drive adoption through WhatsApp submission (designed to be frictionless, so most teams reach high adoption within 4-6 weeks). If adoption is genuinely slow, we can extend the baseline window to give you a fair measurement. The 90-day timing is the target; the honest baseline is the priority.
Practical difference: leakage is the softer, more accurate word for what actually happens on most expense claims. Real fraud (deliberate deception) is a small piece. Most expense leakage is duplicates from confusion, amount discrepancies from carelessness, category miscategorization from unclear policy, and small policy bending that adds up. Our AI catches all of it regardless of intent. When we talk about the dollar impact, we call it leakage because that's what it actually is.
Total flagged claim value divided by total submitted spend, expressed per $100k. Example: if your team submits $500k in claims during the 90-day baseline and $17,500 is flagged, your leakage rate is 3.5% or $3,500 per $100k. We use a rate rather than a raw dollar figure so your leakage measurement stays comparable even as your company grows.
Yes, encrypted in transit and at rest. Historical receipts enable cross-claim comparison (matching a new receipt against previously submitted ones) and audit trail reconstruction. Retention period is configurable per company policy.
There's no legal penalty because there's no formal guarantee. What we do: work with your account team on quarterly reviews to understand why the reduction curve isn't hitting target and what to tune. Common reasons for slow reduction include incomplete team adoption, aggressive attempt patterns from specific employees or departments, or policy gaps we can tighten through configurable controls.
Yes. Book a demo and we'll walk you through a sample report based on illustrative data. When your own report is generated at day 90, it will be structured identically but populated with your actual data.
See the three components on your own data.
A twenty-minute demo walks through Find, Block, and Shrink tailored to your team's expense patterns, currencies, policies, and existing accounting stack.