Figma Leverages OpenAI Athena: How Self-Service Analytics Gets Implemented

Figma is collaborating with OpenAI to embed self-service analytics into the daily workflows of product and marketing teams through its pilot program Athena. According to a video interview, this partnership has brought about an order-of-magnitude change in Figma's analytics delivery efficiency—work originally planned to take a year can now be advanced within two months.

What Is Athena? Why Did Figma Choose It

Athena is an OpenAI data capabilities pilot program, and Figma is participating as an early partner. From the interview, Athena's positioning is not that of a standalone data dashboard tool, but rather an analytics layer that can understand business context and deliver data insights directly to frontline staff's work interfaces.

Figma's core motivation for joining this pilot lies in the bottlenecks of the traditional analytics model: after business users raise data questions, they often must wait for data analysts to schedule time, pull data, and explain results—leading to long feedback cycles, with context easily lost in the intermediate steps. Athena attempts to compress this chain—allowing marketing analysts or email marketers to directly ask questions about business data without going through multiple layers of relaying.

Self-Service Analytics refers to a model that enables business users to query, explore, and interpret data on their own through intuitive tools or interfaces without relying on data analysts. Traditionally, data analytics has been a centralized service: business parties submit requests, analysts schedule data pulls and return reports. Self-service analytics aims to break this bottleneck by decentralizing data access to frontline staff. Typical implementations include BI dashboards and natural language query interfaces. Successful self-service analytics depends on good data governance, unified metric definitions, and sufficient business context; otherwise, non-technical users may easily misinterpret data or reach contradictory conclusions. Figma's practice with Athena is precisely a new form of this concept in the AI era.

How Self-Service Analytics Is Used Inside Figma

The specific scenario given in the interview is marketing campaign analysis. When a new marketing campaign goes live, Athena can understand the nature of the campaign, retrieve relevant information, and build the corresponding analytical context.

On this basis, marketing analysts or email marketers can directly ask the system questions: How is the market responding to this campaign so far? How effective is it in converting customers? Such questions previously required analysts to manually write queries, create visualizations, and then write reports; now they become a conversational interaction.

The key change is not whether the model can answer a single question, but whether the context is complete. Athena packages campaign information, business background, and analytical context together, enabling non-technical roles to obtain usable answers rather than a pile of raw numbers requiring secondary interpretation.

LLM-driven analytical interaction is fundamentally different from traditional SQL queries or BI drag-and-drop operations. LLMs can parse natural language questions, automatically associate relevant data sources, and generate explanatory answers. However, to achieve reliable business analysis, the model needs sufficient context—including campaign definitions, metric definitions, historical data, and more. This is typically achieved through techniques such as Retrieval-Augmented Generation (RAG) or Function Calling. What Athena demonstrates—"understanding the nature of a campaign and building analytical context"—is precisely the result of combining LLMs' semantic understanding capabilities with structured data queries. This approach lowers the barrier for non-technical users, but also imposes higher requirements on underlying data quality and the accuracy of context injection.

What Efficiency Gains Really Mean

At the beginning of the interview, there is a rather blunt statement: things originally planned to be completed within the next year can actually be done within the next two months. This comparison does not come with a specific project list, but it points to a change in tempo—the response cycle for analytical requests compressed from quarterly to weekly or even shorter.

For a company like Figma with fast product iteration and frequent marketing campaigns, the value of this compression is not in how much labor is saved, but in the shortening of the decision chain. When conversion feedback can be obtained immediately after a campaign goes live, teams can make adjustments within the campaign cycle rather than discovering problems only during retrospective reviews.

As a side note, this efficiency description is a unilateral statement from Figma's side; the interview did not provide quantitative metrics such as specific project counts, labor savings ratios, or analytics request processing volumes. Therefore, a more prudent interpretation is: the direction of efficiency improvement is clear, but the magnitude varies by team and scenario.

Three Prerequisites for Implementing Self-Service Analytics

From Figma's practice, it is clear that for self-service analytics to truly take off, at least three conditions must hold simultaneously.

Automatic acquisition of business context. The reason Athena enables marketers to ask questions directly is that the system can automatically understand "what this campaign is." If every question requires manual supplementation of background information, self-service analytics degenerates into just another query interface.

Role-adapted answer presentation. For the same data question, analysts and email marketers need different forms of answers. The system needs to adjust the depth and expression of output based on the questioner's role—this was explicitly mentioned in the interview as one of Athena's capabilities.

Alignment with existing workflows. What Figma emphasizes as "delivering self-service analytics" is not about getting everyone to learn SQL, but about embedding analytical capabilities into the processes that marketing, product, and other roles already use. Tools that do not change work habits are the ones that can be sustained.

Reference Value for SaaS Teams

Figma's case demonstrates that the value of large models in data analytics is shifting from "assisting analysts in writing queries" to "directly serving business roles." This shift has two implications for SaaS companies.

First, the positioning of analytics teams may change. When routine questions can be answered by business users on their own, the value of analysts is more reflected in metric system construction, attribution methodology, and deep decomposition of complex problems, rather than routine data pulling.

Second, the return on investment in data infrastructure is being redefined. Self-service analytics has far higher requirements for data quality and metric consistency than traditional reporting models, because errors are directly exposed to non-technical users. Figma's ability to be the first to make this model work is not unrelated to its own data governance foundation.

OpenAI's Athena is still in the pilot phase, and Figma is one of the early participants. Whether this partnership can be extended to a broader customer base and how it performs in companies with different levels of data maturity remains to be seen. But at least from Figma's feedback, the distance from concept to daily tool for self-service analytics is closer than most people expect.

One-line Summary

Figma leverages the OpenAI Athena pilot to embed self-service analytics into workflows, dramatically compressing analytics response cycles.