Finance ops · Production

Perplexity's two-person finance team automates 163 hours of work per month with Ramp's Accounting Agent

The problem

Perplexity needed to manage hundreds of vendors and thousands of monthly transactions with a general ledger team of essentially one person, and existing AP tools could not flag unusual vendor charges without manual review at that transaction volume.

First attempt

General-purpose credit cards distributed without structure created cascading fraud exposure across dozens of vendor relationships, and AP software could process bills but could not surface anomalous charges, requiring manual digging that did not scale.

Workflow diagram · grounded in source
1
Vendor-specific card issuance
Integration
Each approved vendor relationship gets its own dedicated card with preset spend controls.
▾ source quote
“each approved vendor relationship gets its own dedicated card with preset spend controls”
2
Accounting Agent auto-codes transactions
Ai action
Nearly 100% of card transactions are automatically coded through a combination of Patricia's rules and AI that handles the exceptions.
▾ source quote
“Nearly 100% of the team's card transactions are automatically coded through a combination of Patricia's rules and AI that handles the exceptions”
3
Agent surfaces anomalies with reasoning
Ai action
The Agent draws attention to anything worth an extra pair of eyes and tells the controller why—triple the spend month-over-month, a new line item description, or a coding question.
▾ source quote
“It draws my attention to anything worth an extra pair of eyes and tells me why—triple the spend month-over-month, a new line item description, a coding question”
4
Human correction and model update
Feedback loop
When the Agent gets something wrong, the team corrects it in plain language and the model learns from each correction.
▾ source quote
“When the Agent gets something wrong, the team corrects it in plain language, explains the context, flags edge cases, and gives the system enough information to update its own understanding. The model learns from each correction. The exceptions list shrinks. …”
5
Bill pay inbox surfaces flagged items
Routing
Lauren's inbox surfaces only what needs human review, with each item pre-flagged with a reason.
▾ source quote
“Lauren's inbox surfaces only what needs human review. Each item comes pre-flagged with a reason.”
6
Dynamic approval routing for large invoices
Routing
For large invoices, dynamic approval routing sends them directly to the right executive without finance playing scheduler.
▾ source quote
“For large invoices, dynamic approval routing sends them directly to the right executive without finance playing scheduler”
7
Month-end review output
Output
Month-end becomes a review rather than a scramble because all transactions have been automatically addressed throughout the month in real time.
▾ source quote
“Gone are the days of scrambling to chase receipts, manually coding really large data sets. All of that has been automatically addressed throughout the month in real time.”
Reported outcome

Ramp's Accounting Agent auto-codes over 97% of card transactions and automates an average of 163 hours of finance work per month, including roughly 115 hours from month-end close compression, with Ramp generating over $5M in cumulative savings for Perplexity.

Reported metrics
Cumulative savingsover $5M
Card transaction auto-coding rate>97%
Card transactions coded automatically without manual touch97%
Monthly card transaction volume7,000–9,000
Show all 10 reported metrics
cumulative savingsover $5M
card transaction auto-coding rate>97%
card transactions coded automatically without manual touch97%
monthly card transaction volume7,000–9,000
card transactions automatically codedNearly 100%
total finance work automated per month163 hours per month
monthly close time compressedroughly 115 hours per month
GL coding time saved per month from Accounting Agentapproximately 28 hours
monthly card administration time eliminatedroughly 16 hours
bill pay hours saved per month4+ hours
Reported stack
RampAccounting Agent
◆ Does this fit your context?

Compare to your context

Tell us your scale, team, and constraints. We'll show what changes at your size, what fails at your scale, and whether this case is a fit, needs adaptation, or won't scale to you. Free demo, no signup.

Compare to your context →
~30 seconds · free
Source
https://ramp.com/customers/perplexity
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

Ramp's Accounting Agent auto-codes over 97% of card transactions and automates an average of 163 hours of finance work per month, including roughly 115 hours from month-end close compression, with Ramp generating over…

What tools did this team use?

Ramp, Accounting Agent.

What results were reported?

Cumulative savings: over $5M; Card transaction auto-coding rate: >97%; Card transactions coded automatically without manual touch: 97%; Monthly card transaction volume: 7,000–9,000 (source-reported, not independently verified).

What failed first in this deployment?

General-purpose credit cards distributed without structure created cascading fraud exposure across dozens of vendor relationships, and AP software could process bills but could not surface anomalous charges, requiring…

How is this finance ops AI workflow structured?

Vendor-specific card issuance → Accounting Agent auto-codes transactions → Agent surfaces anomalies with reasoning → Human correction and model update → Bill pay inbox surfaces flagged items → Dynamic approval routing for large invoices → Month-end review output.

Related finance ops cases

WHAT TO DO WITH THIS

Now compare it to your context

This case is one data point. Whether its pattern fits you depends on your volumes, your stack, and your exception load — that comparison is the step no case study can do for you.

Browse the Finance ops cases →