Challenge

Retell AI is redefining customer service by building AI voice agents that are fast, intelligent and adaptable at scale, allowing businesses to automate phone calls for sales and customer support without losing the human touch. When it launched in San Francisco in 2024, the company knew its financial success would depend on accurately charging customers based on how they used the service – such as total call minutes per month, which LLM model and voice engine were used, and whether the customer used its own telephony technology or telephony supplied by Retell AI.
"Our business isn't like a typical SaaS company. It's very complicated and expensive, as every interaction uses a lot of GPU power and LLM tokens, so we needed the ability to look very closely at usage in terms of billing," said Todd Li, president at Retell AI.

The company tried manually calculating each customer's usage for billing purposes, but it quickly became too time-consuming for the 13-person company as its customer base expanded. In addition, the manual process lacked the transparency that customers needed when reviewing their billing statements, which led to complaints. A lack of auto-billing also contributed to a high rate of unpaid invoices, impacting Retell AI's path to profitability.

Retell needed a new solution that would automate its usage tracking and billing process. At the same time, the company wanted to improve other financial workflows, such as payment processing, fraud management and tax compliance, so it could continue growing without having to hire a large finance department.

Solution

Retell AI had launched its platform with Stripe Payments to process customer payments generated through its manual usage-based billing model. After several months of struggling with the challenges of its manual billing process, the company implemented Stripe Billing to automate its usage-based billing.

Retell AI uses Billing to track usage across multiple dimensions for each customer, including total call minutes, LLM model, voice engine and telephony options. Each factor significantly impacts Retell AI's infrastructure costs, which requires the company to track and bill customers precisely on what they use. With this data, Retell AI's billing system automatically creates detailed and accurate invoices that customers can verify against usage logs.

The new billing system also allowed the company to launch a new customer acquisition strategy: a credit grant model that preloads a new account with a limited number of free usage credits. Because Retell AI captures credit card information during sign-up, it can then automatically switch to the credit card for billing once customers deplete their free usage credits, reducing conversion friction for the user and eliminating the risk of unexpected service interruptions. This approach gives prospective customers the chance to explore the platform without exposing Retell AI to the risk of a customer over-using its service during a time-based trial.

Retell AI uses Payments' built-in optimisations to improve its payment performance. Features like Smart Retries automatically retry failed payments at the best time to increase the chance of a successful transaction.

Retell AI also used Stripe's Optimized Checkout Suite to quickly create a fast, flexible and secure checkout experience. To collect payment, Retell AI uses the Payment Element and the Hosted Invoice Page to create a Stripe-hosted customer portal where customers can enter a payment method for automatic billing. Having a payment method on file eliminated the need to manually send invoices each month. Adaptive Pricing automatically displays prices in local currencies across more than 150 markets to simplify the checkout experience.

To increase conversions, Retell AI implemented Link, Stripe's wallet that autofills payment details from saved cards and bank accounts – allowing customers to check out faster with their preferred payment methods. Today, Link powers 38% of Retell AI's payment volume, making it a key driver of growth.

As its platform reached new customers around the globe, Retell AI added Stripe Tax to calculate, collect and file taxes automatically through an integration with the tax compliance platform Taxually, helping the business avoid errors and file on time. This lets the company focus on international growth without needing to add headcount to its finance team.

Retell AI also added Stripe Radar for Fraud Teams to prevent fraud by creating rules to block high-risk transactions and payments from pre-paid credit cards, which pose a particular risk for a usage-based product like Retell AI's platform.

Results

Retell AI grows revenue 10x with usage-based billing

It took one Retell AI engineer a single month to integrate usage-based billing into the company's platform and its existing Payments account. Within a year of implementing Billing to automate its usage-based billing, Retell AI scaled from $1 million to $10+ million in annualised revenue and now processes more than 30 million calls per month.

"Our Stripe customer success manager was with us all the way and helped recommend a certified partner that helped us easily implement usage-based billing without pulling all our engineering resources away from product development," said Li.

Automated billing reduces unpaid balances by 87%

Collecting customer payment information through the Stripe customer portal reduced the number of customers with outstanding balances from 40% to less than 5%, an 87% improvement.

"It's much easier for us to collect payments instantly instead of emailing everyone to remind them to pay their invoices. Every dollar is processed automatically, without manual intervention. It saved us about 10 hours a week at least and allowed us to become profitable sooner," said Li.

Retell AI boosts its authorisation rate by 40%

Since launching with Stripe, Retell AI has increased its payment authorisation rate by 40%, allowing the company to capture $10+ million in sales while also achieving profitability in its first year.

"We are a 16-person team; managing billing and payments manually without Stripe would have required hiring a dedicated finance team, which would have impacted profitability. With Stripe, we've been able to achieve that," said Li.

Smart Retries boosts payment recovery

Retell AI recovers 60% of failed credit card payments using Smart Retries, significantly improving the company's subscription revenue retention.

Retell AI worked with Stripe to help shape the future of usage-based billing

Retell AI is working with Stripe as a pilot partner for a new credit burndown model that allows customers to pre-pay for credits, with credits automatically replenishing before the account runs down to zero to avoid any disruption in service. This model will help usage-based billing companies like Retell AI charge incrementally rather than wait for large balances to be accumulated, reducing the risk of failed payments while improving cash flow.

"This new model will reduce the risk of credit card failures as customers use our products more. Stripe is building this functionality based on our pricing model, so it's great to see how this will impact other usage-based billing companies," said Li.

Retell AI eliminates the need for a dedicated tax team

Retell AI uses Tax to calculate, collect and file taxes in more than 100 countries and 50 US states, supporting its international growth without adding tax-specific headcount.

"We were global from Day 1, so tax was something we knew we would need to automate. Stripe Tax does the equivalent of five full-time finance people focusing on our tax compliance," said Li.

Radar blocks $275K+ in fraud

Thanks to Radar's advanced ML fraud protection and Retell's custom Radar rules, Retell AI has prevented more than $275,000 in fraudulent charges, helping the company save money while staying in good standing with the card networks.

"Radar keeps us from getting a negative reputation with the card networks or being blocked as a merchant," said Li.