How Airtable designed the support experience its customers actually need.

Airtable has always been a composable platform where business users could build and work. What began as a no-code tool for creating custom apps has grown into a shared workflow and data layer for humans and agents. Now, humans and agents can collaborate across interfaces on verified, relational business data in a centralized system with clear governance and guardrails.

More than 500,000 organizations, including 80% of the Fortune 100, build on Airtable. As Amanda Harris, VP of Customer Support at Airtable, puts it, "Our customers don't just use the platform. They build mission-critical operations on it. And when your business runs on what you've built, you have high expectations for how support should show up."

As Airtable innovates and evolves, the support model has to evolve with it. A growing product surface, an expanding enterprise customer base, and rising complexity per interaction meant the old approach couldn't keep pace.

Agents execute in Airtable

Agents are a part of Airtable’s DNA. Airtable’s infrastructure enables agents to be built into its foundation. Agents now have the same range of action in Airtable as humans but with additional governance they need to produce consistent, reliable outcomes.

Agents have quickly taken over some of the most tedious tasks from humans. Now with conversational app building, humans can simply describe the database, outcome, or workflow they want and Airtable’s Omni will build it. With embedded Field Agents and autonomous agents, customers can streamline execution.

It’s not just Airtable’s technology that has transformed. Internally, the company has restructured its support organization around an AI-native operating model, with dedicated teams for AI-powered support, human support, and support optimization working in concert. Live support representatives are equipped with AI-powered tooling that helps them research issues, draft responses, grade interactions, and escalate with precision.

When Airtable looked to uplevel its AI Support platform it faced the classic question –- build or buy? Airtable's AI research and development goes toward what only Airtable can build: a system where agents and teams across product, marketing, sales, and operations get the relational data, composable logic, and flexible interfaces they need to execute work. Partnering with Sierra let Airtable focus on its product differentiators while still delivering best in class conversational AI support experiences for thousands of customers.

The limits of the old approach

Airtable's support org serves solo entrepreneurs troubleshooting a CRM build all the way to Fortune 100 companies running high-stakes, cross-functional workflows. The bar is the same in both cases: fast, accurate, contextual support that actually resolves the problem.

Their legacy chat experience wasn't clearing that bar. When a customer asked a question, the AI agent typically responded with a help center link and asked them to click through, rather than walking them toward a resolution. Most customers had already tried that path before reaching out. They weren't looking for a pointer to documentation. They were looking for an answer.

The team was also managing a fragmented omnichannel experience. Chat, email, and even marketing touchpoints all had slightly different experiences that hindered consistent outcomes. There was no unified entry point and no consistent interaction model across them.

The problem had two sides: the quality of each individual response needed to improve and the channels needed to have a single, coherent support experience.

A single entry point for everything

Airtable partnered with Sierra with a clear ambition: build one intelligent experience for customer interaction that could scale while maintaining consistent, high-quality outcomes. "What really stood out," Amanda says, "was that if we could design the journey, Sierra could actually bring it to life." That confidence unlocked speed.

Within four weeks, Airtable’s Sierra agent was live via chat, and email followed in just seven days. The rollout moved fast, but it was grounded in testing, simulations, and a shared urgency to deliver something meaningfully better for customers.

For customers, the difference was immediate. Instead of being directed to help center articles, they get answers directly in the conversation, step by step, contextual, and tailored to what they're trying to accomplish. When conversations shift topics, the experience holds. Customers move between questions and the agent maintains context in ways previous systems couldn't.

A system that scales itself

The impact showed up quickly in the data.

Airtable's AI resolution rate hit 80%, a significant jump from its previous platform. Sierra is now carrying the equivalent workload of roughly 95 full-time support agents. That capacity isn't replacing the human team. It's redefining what the human team does. The work that reaches human agents is more complex, more interesting, and better aligned with their expertise, while AI handles the volume at a level of quality customers trust.

Email, traditionally one of the hardest channels to automate, is also showing results. Sierra is resolving multi-turn email conversations end to end, including longer exchanges that would have previously required human involvement.

And this is still early. With deeper agentic capabilities and broader integrations ahead, Airtable expects continued gains in both resolution rates and efficiency as the system matures.

Designing for the long term

Behind the scenes, Airtable is building toward something more expansive. "We want a single entry point that can support customers consistently," says Sean Simons, Senior Manager of AI Support at Airtable, "and extend across other parts of the business, not just support." That foundation is enabling more advanced capabilities, including agentic workflows where Sierra can autonomously handle tasks a support agent would otherwise do manually, and that represent unnecessary wait time for the customer.

To sustain and evolve this system, Airtable is hiring for a new role: AI Agent Architect. This person will own conversation and journey design, integration planning, QA, and the ongoing evolution of Airtable's customer-facing AI agent posture across every touchpoint. It's a signal of how seriously Airtable takes the craft of AI-led customer experience, as a discipline that requires dedicated, long-term ownership.

That said, agent quality won't be siloed in one role. Every person on Airtable's support team is building the AI experience. They review interactions, surface gaps, and feed improvements back into the system daily. The AI Agent Architect owns the architectural direction. The rest of the team is the flywheel that keeps making it better.

What this looks like going forward

Airtable isn't treating its partnership with Sierra as a point solution. It's the customer-facing conversational layer in a broader AI support architecture that the company is designing deliberately, one where the system is learning, the team is iterating, and the architecture is built to evolve.

For a company whose customers are builders, that's exactly the kind of support experience they expect.