Impact
- 50+ custom apps built to automate company operations
- Tens of thousands of dollars saved by building AI tools in-house instead of buying multiple third-party solutions
- Hundreds of hours of manual work automated per week with AI tools for customer support and operations teams
"Our entire company runs on Retool—whether it's our custom CRM, sales automation platform, or the AI apps that make life so much easier for our employees." — Nathan Sann, Senior Business Operations Manager, Descript
Summary: Automation at scale
Descript↗ makes creating beautifully designed videos as easy as editing documents. As the company grew, scaling operations became a challenge. Instead of solving the challenges with more headcount, Senior Business Operations Manager, Nathan Sann, was part of the team that built everything from their core CRM to AI evaluation platforms with Retool as its app layer. Descript’s team now operates with the efficiency of a company twice its size.
Moving forward, the question became: how to maintain this efficiency while continuing to scale?
Company overview
- Size: 150+ employees
- Industry: Technology software
- Regions: North America
- Retool capabilities used: Apps and Workflows
- Use cases: CRM, user data enrichment, support tools, AI evaluation, outbound automation
- Data sources: Production database, BigQuery, Salesforce, Zendesk, Stripe, APIs (Clearbit, Apollo)
Challenge: Fragmented dashboards slowing teams down
To address the company’s expanding operational needs, Nathan needed to figure out how to provide better operational tooling to help teams do more with less. When Nathan joined Descript's Business Operations team, the company relied on a patchwork of ad hoc dashboards engineers had built to solve immediate problems. These tools had inconsistent interfaces and access patterns, making them inefficient to use at scale.
“There were all sorts of dashboards with different purposes, buttons, and designs,” Nathan explains. “It was a giant overwhelming series of buttons and tables.”
Descript needed a centralized system and better tools to make work easier to get done. However, standard solutions couldn’t accommodate the company’s unique data architecture. And building each tool from scratch would have consumed more engineering resources better allocated to its core product.
Solution: Building a custom platform to power company operations
Nathan used Retool to consolidate the disparate dashboards into a central operating system without needing to rely on engineering resources. The team built a custom CRM and suite of tools that connect to Descript’s production database, BigQuery instance, Zendesk, and Stripe, providing a unified view of customer data, projects, and system configurations. Instead of building custom interfaces for each AI project, they could use Retool to drop in components, connect to any API, and start iterating immediately without a huge investment.
“Retool is our secret weapon,” says Nathan. “It allows us to remain flexible and build things that scale very quickly without taking on heavy risks.”

The team identified two areas where automation could have the biggest impact.
Contact enrichment workflows for sales and marketing
The team also built AI-powered enrichment workflows that connect to databases and APIs like Clearbit and Apollo. These workflows validate emails, check against domain blacklists, and use AI to add structured data from other third-party sources. The enriched data flows into their BigQuery instance and downstream tools, creating a single source of truth that routes valuable information to sales and marketing teams.
This system has improved lead scoring and qualification while enabling more targeted marketing. With cleaner data, Nathan’s team can make better decisions about who to target and can deliver more personalized messages. “This has saved us tens of thousands of dollars in enrichment costs alone since we can cross-reference data from our internal systems first to avoid duplicating enrichment costs,” says Nathan. With improved contact data, Descript’s reps save 3-4 hours weekly from having to sift through invalid leads.
Beyond sales and marketing automation, Nathan recognized another critical area where custom tools could save both time and money.

Evaluation tools to validate AI features
Before buying a separate evaluation tool, Nathan developed an AI evaluation platform to test and improve language translation models for Descript’s product. The platform focuses on matching syllable counts across languages to enhance lip-syncing in videos.
The solution connects to Descript’s production database and incorporates various language libraries. It tests translations to ensure syllable counts align properly between source and target languages. The platform programmatically evaluates AI model outputs and stores all test results in a Retool Database↗ table, creating a comprehensive evaluation repository. The quality of AI translations in Descript’s product has also improved with immediate feedback on model performance.
“I built an evaluation platform that’s custom-fit to the ways we work,” says Nathan. “It has all the same functionality, but it’s a tiny fraction of the cost of an evaluation vendor.”

The cumulative impact of these custom solutions has transformed how Descript operates.
Results: Organization-wide automation
With Retool, Descript operates with unprecedented agility, turning AI experiments into production apps in hours instead of weeks, and allows them to keep up with rapid industry innovation.
“Our entire company runs on Retool,” Nathan says.
Descript has automated hundreds of hours across teams through over 50 custom Retool apps that are used daily, delivering unique solutions that accelerate the company's operations. These solutions unify disparate processes by establishing direct connections to Descript's data sources and tools.
“Retool is a force multiplier—it lets us drop in AI-powered tools rather than people to fill gaps,” explains Nathan. “And that’s a huge step for us in a company our size as we're growing—you can’t just hire your way out of problems all the time. The ROI isn’t just in dollars saved, but in our ability to move quickly and solve complex problems.”
Looking ahead, Nathan sees even greater potential for automation.
What’s next
Nathan plans to build agents and more AI apps. “I’m finding ways to automate hundreds of hours of manual work using Retool as our app layer for AI. As models get better, Retool is my go-to platform for building custom tools and agents on top of them. The possibilities are endless.”
Ready to see Retool in action? Get a demo↗ with a Retool expert or start building for free↗ today.
Related stories
- ClickUp↗ automates GTM operations with custom AI apps and workflows
- University of Texas Medical Branch↗ screens 10x more patients with AI-powered Retool apps.
- Power Digital↗ transforms contract management with Retool as their app layer for AI