AI Strategy Index

Duolingo AI Strategy: An AI-First Memo and 148 AI-Built Courses

How Duolingo uses AI on its own work: an AI-first memo, AI in hiring and reviews, and 148 courses built with AI in under a year. Updated November 2026.

Company
Duolingo
Industry
Technology
Headquarters
Pittsburgh, United States
Region
US
AI strategy
Published AI strategy
Index group
AI Pacesetter (not counted)

Language-learning app, also teaching math, music and chess, with a subscription and advertising model

Where the AI strategy is set out

Duolingo CEO Luis von Ahn published his AI-first memo in April 2025 and clarified it in May 2025.

Read the company source

Where Duolingo Is Going with AI

Duolingo declared itself AI-first in April 2025. CEO Luis von Ahn’s memo said Duolingo would “gradually stop using contractors to do work that AI can handle”, according to Entrepreneur.

The memo tied AI to people decisions. AI use counts in hiring and in performance reviews, and teams can add people only if they cannot automate the work, MarketBeat reported.

Von Ahn published the memo, so Duolingo’s approach is on the record. Duolingo sits outside the S&P 100 and STOXX Europe 50, so it is profiled as an AI Pacesetter and is not counted in index figures.

The focus has three parts.

  • Content: generative AI that creates and checks course material
  • Headcount: new hires only where teams cannot automate the work
  • Everyday work: AI use rewarded in reviews
Duolingo’s green owl mascot surrounded by tiles showing the flags of Japan, Korea, China, Germany, France, Spain and Italy
Duolingo’s image for its April 2025 launch of 148 new language courses, which it said were built with generative AI. Image: Duolingo

What It Takes: Generative AI Content, Expert Review and Clear Communication

Generative AI multiplied course output. Duolingo launched 148 new language courses in April 2025, built in under a year, while its first 100 courses took 12 years, according to Duolingo.

Experts focus where they matter most. Jessie Becker, Senior Director of Learning Design, said in April 2025: “Now, by using generative AI to create and validate content, we're able to focus our expertise where it's most impactful.”

The memo drew a backlash. Von Ahn later wrote: “When I released my AI memo a few weeks ago, I didn't do that well.” He said Duolingo was continuing to hire at the same speed, a lesson in AI change management.

Quality needs checking. Von Ahn said about 20% of 1,000 AI-generated learning stories could be “pure slop”, according to IBTimes in October 2026.

Headcount rose 50% from 2022 to 2025, to about 900 full-time employees. The figures exclude contractors, the group the memo targeted.

AI fluency, coded against the Lead with AI Fluency Matrix (September 2026 edition):

  • Fluency floor: Level 1 (Chatting), by expectation. AI use counts in hiring and in performance reviews.
  • Highest level evidenced: Level 3 (Integrating), company-claimed. Generative AI creates and checks course content in production.
  • Portfolio design: Duolingo puts AI at the core of content production and keeps experts on quality.

How Duolingo Measures Results

Duolingo reports output gains. The figures below are company-claimed:

  • Courses: 148 new courses built in under a year, against 12 years for the first 100
  • Hiring: von Ahn said Duolingo was hiring at the same speed after the memo

Gaps remain. Duolingo reports volume and says little about quality or cost, and its CEO flagged weak AI output himself. Reporting both is the step from AI adoption to AI value.

Next markers: contractor numbers, quality measures for AI-built content, results in reviews and hiring, and headcount for 2026.

Change log

  • April 2025: Von Ahn declares Duolingo AI-first and launches 148 AI-built courses.
  • May 2025: Von Ahn says the memo lacked clarity and that hiring continues.
  • October 2026: Von Ahn flags weak AI-generated stories.

The Bottom Line

  • Disclosure: 4 of 4. Duolingo answers all four questions, the last in part, with output figures and no value figure.
  • AI-first since April 2025, with AI in hiring and reviews.
  • 148 courses built in under a year.
  • Level 3 is evidenced in content production, on company-claimed figures.
  • The say-do gap is people. The memo targeted contractors, and Duolingo has published no figures on them.
  • Watch quality. The test is whether Duolingo reports how good AI-built content is.

Revenue, Headcount and Revenue per Employee

What the numbers show: Revenue rose 181% from 2022 to 2025 while headcount rose 50%, so revenue per employee rose 87%, to about $1.15M. These figures do not show that AI caused any change.

Revenue, headcount and revenue per employee, FY2022 to FY2025

  • FY2022: $370m revenue, 600 employees, $616k per employee
  • FY2023: $531m revenue, 720 employees, $738k per employee
  • FY2024: $748m revenue, 830 employees, $901k per employee
  • FY2025: $1.04bn revenue, 900 employees, $1.15M per employee

Change, FY2022 to FY2025: revenue up 181% (2.8x), headcount up 50%, revenue per employee up 87%.

Notes: Revenue is as reported. Headcount is full-time employees at year-end and excludes contractors.

These figures give context on scale and productivity. They do not show that AI caused any change. Sources: Duolingo 10-K filings and S&P Global data.

How Duolingo Compares with Internet, Media and Telecom Peers

Duolingo is an AI Pacesetter: an AI-first company outside the index, profiled for contrast and not counted. Here is how it compares with internet, media and telecom companies in the index.

CompanyDisclosureAI ownerFluency floorGovernanceAI value reportedRevenue per employee
Alphabet4 of 4CEO; no internal AI lead namedNot published; AI training open to all staffAI Principles, rewritten in 2025Engineering results, no total$2.11M
American Tower3 of 4CEO; network operations leads predictive workNot publishedControls and oversight stated; no published frameworkNone published$2.19M
AT&T2 of 4Chief Data and AI Officer, reporting to the CTOLevel 1, training required before accessReview board for high-risk use casesFolded into total cost savings$945k
Comcast4 of 4Co-CEO and CFO lead the cost transformationNot publishedOrchestration layer and traceability for AI agentsEarly care results, no total$691k
Disney3 of 4CEONot published; internal AI policy and approved toolsInternal AI policy site; usage dashboardNone published$409k
Meta3 of 4CEO; Chief People Officer runs AI in reviewsLevel 2, set through performance reviewsFrontier AI Framework for modelsNone published for internal use$2.55M
Netflix2 of 4Chief Product and Technology OfficerNot published; AI fluency an aspiration for all staffGenerative AI guidelines for production partnersProduction examples, no total$2.82M
T-Mobile4 of 4CEO; Director of AI EngineeringNot published; AI tools for frontline staffNot publishedResults by area, no total$1.18M
Verizon4 of 4CEO; CTO leads network and engineering AINot published; Claude Code for 33,000 technology staffHumans own agents; token allowances by roleResults by area, no total$1.54M
Duolingo
AI Pacesetter
4 of 4CEO set the AI-first directionLevel 1 by expectation: AI use counts in hiring and reviewsNot published148 courses built in under a year$1.15M
Fiverr
AI Pacesetter
3 of 4CEO leads the AI-first resetNot publishedNot publishedNo: a smaller team, no AI figures$816k

Disclosure counts how many of the four strategy questions (where it is going, where it will focus, what it will take and how it will know it is working) a company answers in public. Fluency floor is the lowest level of AI fluency a company expects of all its people, on the five levels of the Lead with AI Fluency Matrix: Avoiding (0), Chatting (1), Collaborating (2), Integrating (3) and Transforming (4). The other columns summarize each company’s profile. Revenue per employee is for each company’s latest fiscal year, in US dollars. Rows marked AI Pacesetter or Notable are companies outside the S&P 100 and STOXX Europe 50. They are shown for contrast and are not counted in any index figure. See all companies in the AI Strategy Index.

Frequently Asked Questions

What is Duolingo’s AI strategy for its own work?

Duolingo is AI-first: it builds content with generative AI, counts AI use in hiring and reviews and adds people only where work cannot be automated. CEO Luis von Ahn set this out in April 2025.

‍

What did Duolingo’s AI-first memo say?

The memo said Duolingo would gradually stop using contractors for work AI can handle. It also tied AI use to hiring and performance reviews.

‍

Who leads AI at Duolingo?

CEO Luis von Ahn set the AI-first direction. The sources reviewed name no separate AI chief.

‍

Did Duolingo replace employees with AI?

Von Ahn said he does not see AI replacing what Duolingo’s employees do. The memo targeted contractors, and full-time headcount rose to about 900 in 2025.

‍

How is this profile put together?

Lead with AI profiles use public sources only, such as annual reports, earnings materials, filings and on-record interviews, dated at the time of writing. AI fluency is coded against the Lead with AI Fluency Matrix, and company figures are marked as company-claimed unless they have been independently verified.

About the AI Strategy Index

The Lead with AI AI Strategy Index profiles how the largest companies in the US and Europe approach AI, starting with the S&P 100 and the STOXX Europe 50. Each profile answers four questions: where the company is going with AI, where it will focus, what it will take and how it will know it is working. AI Pacesetters and notable companies outside the two indexes are profiled for contrast and are not counted. Profiles use public sources only, and AI fluency is coded against the Lead with AI Fluency Matrix.

See all company profiles