CUSTOMER STORIES
Tipalti Streamlines $75B in Annual Payments with Democratized Access to AI-Powered Insights
By enabling LLM data exploration in Snowflake’s AI Data Cloud, Tipalti accelerates insights, saves time and money and empowers key teams with actionable insights.
5xMore queries run simultaneously, allowing faster access to key financial insights
$75bIn payments managed yearly
Industry
Technology, Financial ServicesLocation
California, USA, IsraelMaking finance work smarter — not harder — for thousands of companies
From mom and pop shops to Fortune 500 companies, every business depends on its payment systems to keep the lights on. Yet outdated, time-sensitive and disconnected solutions mean businesses often move faster than payments. When one missed transaction can have significant consequences, overcoming this challenge is key to future success — or in some cases, survival.
Headquartered in Foster City, California, Tipalti provides a global payables automation platform that helps hundreds of high-growth businesses save their time, energy and expertise by streamlining and scaling financial operations — to the tune of $75 billion a year.
The company’s goal is to completely transform the way finance teams work. To deliver on this promise, Tipalti has an innovation strategy that places data and intelligence at its core, designed to provide insights and simplify decision-making across its internal business units.
Today, it uses Snowflake’s AI Data Cloud, running on Amazon Web Services (AWS), to support that strategy. With Snowflake and Cortex AI, Tipalti delivers LLM-powered data exploration to key internal teams, enabling employees to use pre-selected prompts to easily generate vital insights — all without the need for specialist data skills.
Story highlights
Centralized access to LLM prompts democratizes insights and data exploration: With pre-defined LLM prompts, teams can achieve greater consistency of insight, drive seamless collaboration and explore data without technical know-how.
Reduced complexity removes operational overhead: Through Snowflake and the inbuilt functionality of Cortex AI, Tipalti has removed any need for external AI infrastructure, reducing huge amounts of complexity, effort and cost.
- Seamless scalability can bring faster time to value, organizationwide: Tipalti’s sales and marketing teams are already benefiting from easy access to AI insights, but Snowflake’s elastic compute means this solution can be scaled to meet the needs of all internal teams.
Improving consistency and business collaboration with on-demand insights
Tipalti’s innovation strategy is aimed at spreading the value of data across the organization, which means providing key internal teams with access to timely insights.
To meet this need, the company built an “AI-prompt store” on Cortex AI; a large library containing pre-defined LLM questions to help teams like sales and marketing dive into data without coding capabilities.
The solution enables Tipalti to integrate AI insights into the platforms its teams use every day, such as Salesforce, so users have the insights they need exactly where they need them.
“The biggest benefit is that the insights are now always there for our people,” says Taito. “If they want to analyze a certain opportunity or group of accounts, they can just log in, make a query, and it’s all there. They don’t have to think about how to build a prompt or manipulate data, they just get answers.”
The overall result has been improved internal collaboration and consistency of insight between marketing and sales teams. And Snowflake’s seamless scalability means this solution can easily be rolled out to other areas of the organization. Already, more than ten new features have been deployed to over 50 users since the initiative began. The result has been more than 600 ad-hoc analyses that significantly streamline sales and interaction evaluations.
Supporting better decisions with risk-free data exploration
Tipalti initially chose Snowflake to support its innovation strategy because of numerous capabilities that aligned with its own ambitions. The company had three major objectives: running large-scale prompt executions efficiently and securely. Maintaining scalability and consistency across multiple use cases. And eliminating the need for any external inference services or complex orchestration layers.
“We had a clear strategic vision for our data architecture,” says Yair Taito, Director of Data Engineering at Tipalti. “We wanted maximum scalability with minimum infrastructure overheads, and that’s why we built our data warehouse on Snowflake.”
Snowflake also had numerous advantages for increasing time to delivery and value of insights. Most importantly, it allowed Tipalti to apply AI to its data without incurring privacy and security risks, a common issue when moving data between various systems. Instead, the company could simply leave data where it resides, and rely on the security and governance features already present in the data cloud.
“Snowflake has enabled us to bring AI capabilities to our data, rather than moving data to the AI,” says Michael Naor, Senior Data Engineer at Tipalti. “This shift allows our teams to experiment and deploy AI-driven insights faster, while maintaining the vital security and governance of our data platform.”
Delivering AI capabilities at speed and scale with native Cortex features
Cortex AI contributed several features that allowed Tipalti to achieve its goal of delivering instant and widespread access to insights.
AI_COMPLETE integrates LLM capabilities from Cortex directly into SQL workflows without needing additional infrastructure or maintainance. The response_format feature removes potential hallucinations and errors and allows developers to ensure consistently structured outputs from LLM queries. And TRY_COMPLETE ensures that any single failure won’t cause an entire batch process to fail. This saves countless hours identifying and investigating problematic rows.
Tipalti’s data team also used Snowpark to develop a custom function so multiple queries can run in parallel. This increases throughput speed and responsiveness by up to five times so more insights can reach more people in less time.
We really wanted to leverage the capabilities of Cortex and Snowflake. It meant we didn’t have to move our data and we knew we could make sure the answers were good, scale the solution at will, and add new features quickly.”
Yair Taito
Director of Data Engineering at Tipalti
Through building this functionality in Snowflake, on AWS, Tipalti has seen more reliable and predictable parsing thanks to the robust features within Cortex. And not having to rely on external AI solutions, but instead doing everything within Snowflake itself, has made managing, maintaining and developing functionality far easier.
“It would have been much more complicated to build this without Snowflake,” says Taito. “Everything from security and maintenance would have taken a lot more effort. But with Snowflake we don’t have to worry about resources or external infrastructure and LLMs — we don’t need anything.”
In the future, Tipalti expects to evolve the framework into a fully autonomous, context-rich AI assistant layer, using features within Snowflake. Its next steps will be to integrate Cortex Semantics Models, enable retrieval-augmented generation (RAG) using Cortex Search with embeddings, and continue to improve its LLM responses to enable deeper data exploration.
These new additions will further simplify access to insights, allowing any and all business teams to explore data at will. Armed with that insight, Tipalti can better meet its goal of transforming the face of finance, and helping everyone from mom and pop shops to major enterprises work smarter, not harder.
Our goal is to enable self-service exploration for as many end users as we can, giving them the ability to interact with data and get the insights they need. We’re going to use all the Snowflake intelligence products available to do that.”
Yair Taito
Director of Data Engineering at Tipalti
