36 budgets scored on day 14 of 31: 17 are projected past their limits by month-end, worth $231K of overshoot. Spend shows the projection, the burn behind it, and the requests still waiting on an approver.
The agent clears routine spend — 57% of reviewed volume — and escalates the rest with the reason printed: “amount ≥ $250”, “first-time vendor”, “hard-block policy hit”. No silent approvals, no silent denials.
spend.cedxsystems.com — live build
Runs on demo data — the burn board labels itself “day 14 of 31” on the card.
$231K projected overshoot17 budgets past limit by month-end
57% auto-approval rate20 cleared this week · 69 escalated to humans
16.2% top-vendor sharetop-3 at 36.6% · HHI 708 · 7 high concentration
What it is
Projection, automation and concentration on one board.
Overshoot projected, not discovered
Each budget carries spent, limit, end-of-month projection, overshoot and risk. Support tooling: $34K spent of $37K, projecting $76K, overshoot $39K, risk 76. Linear burn extrapolation — the alert says the method out loud.
69 escalations pending, $46K held for review — each with its reason chips: a $894 software line for “amount ≥ $250”, a $226 airfare for “first-time vendor”, a $98 ground line for “hard-block policy hit”. Approve and dismiss sit on the row.
Escalations 69 · auto-approved 91 · all reviewed 160
Reason chips on every escalation
Under $100 cap · recurring vendor · receipt matched
spend — screen-2
Vendor concentration measured
40 vendors scored with share, spend, risk, band and tier: one software vendor at 16.2% of spend, tier Preferred, risk 96; a legal vendor at 9.1% on Watch. Diversifiable spend in the high band: $47K.
Top-1 / top-3: 16.2% / 36.6% · HHI 708
Tiers: Preferred · Watch · Approved · Blocked
Sim top-1 falls to 12.9% with 2 levers
spend — screen-4
Product tour
Four screens, captured from the running build.
Not a mockup and not a concept deck. This is what opens at /app/spend.
spend.cedxsystems.com
01 — Overview
The month's spend with its future attached
MTD spend $29K across 320 transactions, $95K of requests pending, 24 budgets at risk of which 17 project over. Alerts price the problems: 27 requests past the 3-day SLA worth $81K, and 2 cards at 80%+ of their monthly limit.
Budget health ring at 50, drivers listed
Category mix: professional $35K leads
141 policy-flagged transactions · $88K
02 — Budget burn
Day 14, and the end is visible
The burn board lists every budget's trajectory: community sponsorships already $9.1K over projection, contractor hardware at $12K over, recruiting ops $23K over. Status is a word — Over — with the arithmetic next to it.
12 budgets at high burn risk
Delta saved $90K vs baseline after levers
Burn board · burn levers · scoring model tabs
03 — Agent review
What the agent did and why it asked
Three tabs tell the whole story: 69 escalations awaiting a human, 91 auto-approved, 160 reviewed in all. The auto-clear rules are chips on the screen — under-$100 cap, recurring vendor, receipt matched — and 20 lines cleared this week with no human touch.
Escalation value $46K held for review
40 reviewed today by the agent
Approve / dismiss on every escalation
04 — Vendor risk
Concentration you can act on
The risk board ranks 40 vendors: a marketing network at 11.1% of spend, an airline at 9.3% — each with a tier and a band. The diversify levers simulate the top-1 share falling from 16.2% to 12.9%.
7 vendors in the high band
$47K diversifiable in the high band
Blocked tier exists and is used
Who runs it
Three roles keep spend inside the lines.
Roles, not references. We have no named customers yet, so nobody in these photographs is quoted, credited or claimed as one.
The cardholder
Photographs the receipt, gets matched, and only hears about it again if the agent escalates — which it does with the reason attached.
57% auto-approved
Budget owner
Sees the end-of-month projection on day 14, not day 31 — their budget's overshoot is a number with a risk score, weeks early.
17 over · $231K
Procurement
Watches top-vendor share and the blocked tier, and works the 27 requests aging past SLA before they become maverick buys.
27 past SLA · $81K
The shape of it
What the demo workspace actually looks like.
Every figure below is legible in the captures above. Nothing here is a projection of your estate — it is the state of the demo data.
$29KMTD spend320 transactions booked
$95Kopen requests34 pending · 27 past 3d SLA
$231Kprojected overshoot17 budgets · day 14 of 31
57%auto-approval rate160 reviewed · 69 escalated
Spend by categoryall sources · $K · from the overview capture
Professional$35K
Travel$29K
Software$26K
Marketing$24K
Hardware$18K
Facilities$18K
Training$17K
Lodging$17K
Budget healthcomposite · printed on the overview
50
Inside limits
24 at risk · 17 projected over
Review volume auto-approved91 of 160 reviewed
Auto-approved · 91 — routine, matched
Escalated to a human · 69 · $46K
How it runs
A dollar of spend, in the order it actually happens.
01
Request
Spend starts as a request or a card swipe — 34 requests are pending now, $95K, and 27 have waited past the 3-day SLA.
02
Clear
The agent auto-approves what the rules cover — under-$100, recurring vendor, receipt matched — 57% of reviewed volume, with the rules printed as chips.
03
Escalate
Everything else goes to a human with its reason: amount over $250, first-time vendor, hard-block hit. 69 are waiting, $46K held.
04
Project
Budgets extrapolate their own burn: on day 14, 17 project past their limits for $231K — the overshoot conversation happens while it can still change.
One record
The spend is one record across the estate.
These captures come from one demo workspace — Northline Systems — so the vendors and amounts here are the same ones the rest of the finance apps carry.
Finding this out on the third call is worse for you than reading it here, and worse for us.
Spend is not generally available. What opens today is the live build running on demo data — the 36 budgets and 40 vendors are the demo book, not your cost centres.
We have no named customers to show you, so this page shows none. Northline Systems in the captures is the software's demo workspace, not a reference.
The auto-clear agent's rules are printed on screen; we are not claiming a 57% rate for your spend — that figure is the demo book's, and your policy will differ.
Burn projections are linear extrapolations, as the alerts themselves state. They are a warning light, not a forecast service.
No audit or compliance certification has been issued for Spend. What we can evidence about hosting, encryption and access is on the security page.
Yes. Every screenshot is a capture of the running build and you can open the same build at /app/spend. It runs on demo data.
What does the agent auto-approve?
What the rules cover, and the rules are chips on the screen: under the $100 cap, recurring vendor, receipt matched. Everything else escalates to a human with its reason printed — the 57% auto-approval rate is the demo book's outcome, not a promise.
How does the overshoot projection work?
Linear burn extrapolation, stated on the alert itself: day 14 of 31, current burn extended to month-end. 17 budgets project over for $231K, and two levers simulate that down to $141K.
What is the vendor concentration measure?
Share of spend per vendor plus HHI — top-1 at 16.2%, top-3 at 36.6%, HHI 708 — with tiers from Preferred to Blocked and a diversifiable-dollar figure for the high band: $47K.
Is Spend audited or certified?
No certification has been issued. What we can evidence about hosting, encryption, tenant isolation and production access is written up on the security page.
The budgets are projecting. Go and look at them.
Live build, demo data, no card. Then ask which of your budgets breaks first this month.