How Segwise got real-time P&L visibility across 100+ AI credit accounts

Real-time view

of Deferred and Recognized Revenue

Annual Cost Savings

CAC

Vsible on every trial account

Data Interportability

Credit Burn

Tracked across 100+ Accounts Automatically

Faster Pricing Iteration

Company

Segwise

Industry

Creative Analytics & AI-Powered Ad Generation

Size

Early-stage AI startup

Customers

Hunch, Traya

The Challenge

Segwise builds AI agents that analyze ad creatives across 15+ networks. Mobile gaming companies, DTC brands, and subscription apps use it to understand what makes an ad perform and generate new versions from that data. They have 100+ enterprise customers on credit-based contracts. Until Ferry, the P&L behind those accounts was largely invisible.

Credits sold, books closed on faith

"We were deciding pricing on the go for a lot of clients,"

said Kush Daga, Founding Engineer at Segwise.

"The first problem was getting clients. Once we got clients, then we were thinking about how to price them."

That worked until AI credits entered the picture.

When Segwise moved to credit packages for AI analysis, they sold annual bundles to enterprise customers before any system existed to track consumption. Cash came in at contract signing. Billing ran monthly. Trial credits went to new accounts on top.

The cash side and the revenue side were on different timelines, and nothing was reconciling them.

When Segwise sells an annual credit bundle, that cash lands on the books as deferred revenue under ASC 606. Revenue is earned as credits are consumed, ad analysis by ad analysis. Without real-time consumption tracking, nobody could produce the earned-versus-deferred split on demand. The P&L was a month-end reconstruction project.

The cost that had no number

Every new customer gets 1,000 credits upfront. Some convert. A lot do not. Before Ferry, there was no way to know which was which while Segwise absorbed the cost.

"We used to assume it's a big client, they'll probably convert later. But we didn't have any visibility on how much money we were burning on them."

— Kush D, Founding Engineer, Segwise

A trial account that churns is customer acquisition cost with no return. Without a number attached to each trial, Segwise could not separate accounts worth investing in from the ones burning through compute at a loss.

The burn was real. The number was not.

Company

Segwise

Industry

Creative Analytics & AI-Powered Ad Generation

Size

Early-stage AI startup

Customers

Hunch, Traya

“Our core product isn’t credits. We build ad analysis and ad generation technology, not billing or finance infrastructure, and that's where my focus needs to be."

— Kush Daga, Founding Engineer, Segwise

Scale your Revenue with Ferry AI

Get a Demo

Talk with Us

What Ferry's AI agent changed

Credit burn as a revenue signal

COGS before the decision

The outcomes

The Solution

What Ferry's AI agent changed

Ferry's AI agent tracks credit consumption for every Segwise customer as it happens. As a marketing team runs ad analyses, the credit balance moves. Deferred revenue becomes recognized revenue as credits are earned. The P&L reflects what is actually earned, updated continuously.

Annual bundles billed monthly are handled automatically. Ferry's AI agent keeps the billing schedule and the credit allocation separate.

Cash comes in monthly. Revenue is recognized as credits are consumed. Segwise does not reconstruct this at month-end. It is already current.

Free trial credits sit in their own allocation, tracked from first use. The cost of acquiring each customer has a number attached to it from day one.

Credit burn as a revenue signal

Ferry's AI agent sends Segwise a daily report to Slack showing the credit balance and burn rate for every account.

"We have connected a webhook to Ferry which sends us every day at the end of the day on our Slack how much credits are left for each org."

Green means the account has a healthy balance.
Yellow means under 200 credits: the renewal window is opening.
Red means gone. Before Ferry, Segwise found out about empty accounts at invoice time, the same moment the customer did.

"Now we know, okay, this is a big client, so we should get them on a call and talk to them. We should give them more credits on our own and not just let them use up all that infra without knowing."

Credit burn rate is now a live revenue forecasting input. Segwise can see which accounts are active, which are approaching renewal, and which are quietly going quiet.

COGS before the decision

Backfilling customer data is one of Segwise's costlier operations: running AI analysis on 30 days of historical ads. A single backfill can consume 1,000+ credits.

Before Ferry, the team ran them without knowing the credit cost relative to what the customer had left.

"Now our team knows: this would require a thousand credits. They only have 400. Do we want to do a backfill, or should we push back? Are they big enough for us to support this?"

Ferry's AI agent surfaces the cost before the decision. The subsidy is a choice now, not a line item that shows up later on the P&L.

The outcomes

Real-time view of deferred and recognized revenue across 100+ accounts.
No month-end reconstruction.
Trial account CAC tracked from first use.
Credit burn data feeding into revenue forecasts, so renewals happen before accounts lapse.
Backfill costs visible before the decision. 100+ enterprise accounts tracked automatically, zero manual credit management.

What Ferry's AI agent changed

Annual bundles billed monthly are handled automatically. Ferry's AI agent keeps the billing schedule and the credit allocation separate.

Cash comes in monthly. Revenue is recognized as credits are consumed. Segwise does not reconstruct this at month-end. It is already current.

Free trial credits sit in their own allocation, tracked from first use. The cost of acquiring each customer has a number attached to it from day one.

Credit burn as a revenue signal

Ferry's AI agent sends Segwise a daily report to Slack showing the credit balance and burn rate for every account.

"We have connected a webhook to Ferry which sends us every day at the end of the day on our Slack how much credits are left for each org."

Green means the account has a healthy balance.
Yellow means under 200 credits: the renewal window is opening.
Red means gone. Before Ferry, Segwise found out about empty accounts at invoice time, the same moment the customer did.

"Now we know, okay, this is a big client, so we should get them on a call and talk to them. We should give them more credits on our own and not just let them use up all that infra without knowing."

Credit burn rate is now a live revenue forecasting input. Segwise can see which accounts are active, which are approaching renewal, and which are quietly going quiet.

COGS before the decision

Backfilling customer data is one of Segwise's costlier operations: running AI analysis on 30 days of historical ads. A single backfill can consume 1,000+ credits.

Before Ferry, the team ran them without knowing the credit cost relative to what the customer had left.

"Now our team knows: this would require a thousand credits. They only have 400. Do we want to do a backfill, or should we push back? Are they big enough for us to support this?"

Ferry's AI agent surfaces the cost before the decision. The subsidy is a choice now, not a line item that shows up later on the P&L.

The outcomes

Real-time view of deferred and recognized revenue across 100+ accounts.
No month-end reconstruction.
Trial account CAC tracked from first use.
Credit burn data feeding into revenue forecasts, so renewals happen before accounts lapse.
Backfill costs visible before the decision. 100+ enterprise accounts tracked automatically, zero manual credit management.

What's Next in Billing For Segwise?

Segwise is building separate credit wallets per product line: one for creative tagging, one for ad generation, and one for competitor data. Each has a different cost curve. Ferry's AI agent will track each wallet independently, so customers can scale one product without drawing down another product's credit allocation.

The invoicing layer is next. As enterprise deals grow, Ferry's AI agent handles the full contract-to-cash cycle: credits tracked, invoices issued, revenue recognized.

What's Next in Billing For Segwise?

The invoicing layer is next. As enterprise deals grow, Ferry's AI agent handles the full contract-to-cash cycle: credits tracked, invoices issued, revenue recognized.

Ready to transform your revenue infrastructure?

If you're spending valuable engineering time on manual AR instead of your core product or if you need to scale complex month end close without overhead, FERRY AI can help.