2023 State of the API Report

Postman State Of The API Report Postmanauts researching ontop of graphs. Illustration.

SOTA hex

APIs and generative AI

We asked survey-takers whether they were using generative AI tools, and if so, how. Sixty percent of respondents said they were employing these tools. The most common uses were finding mistakes in code, producing code, and coding instruction.

Senior API developers were much less likely to use AI tools than junior developers, except to generate documentation and comments. Here, their usage rates were similar.

Among industries, government and defense were least likely to use AI tools, with just 51% of respondents there adopting them. Some respondents cited security risks and corporate bans on sharing data with third-party AI tools.

Education had the highest rate of uptake, with 65% of survey takers there saying they used AI tools.

To find mistakes in my code: 51%
To generate code that a computer will understand: 35%
To teach me how to code: 34%
To generate comments on my code: 30%
To generate API documentation: 28%
To suggest product ideas for my code: 24%
To generate API design: 19%
To flag potential security vulnerabilities in my API: 18%
To generate prose or marketing copy: 14%
To generate a description of an API I've discovered: 14%
To generate Postman collections: 12%

Multiple choices allowed.

What apps are developers most excited about building over the next year? It may come as no surprise that generative AI-powered apps took the top spot. This sentiment was strongest in the education sector.

Internet-of-things apps placed second overall, except in manufacturing. There, IoT apps were viewed equally as exciting as AI apps.

Fintech apps placed third overall, except in financial services, where respondents deemed fintech apps twice as exciting to build as AI-powered ones.

Generative AI-powered apps: 37%
IoT apps: 16%
Fintech apps: 14%
Payment apps: 9%
Health apps: 9%
Crypto apps: 4%

Multiple choices allowed.

What sort of productivity benefit do API professionals expect tools like ChatGPT and Copilot will deliver to development over the next two years?

There was wide agreement that AI will deliver a boost in developer productivity, with the largest share of respondents forecasting improvements of 10% to 25%.

10% to 25%: 41%
26% to 50%: 30%
51% to 75%: 12%
More than 75%: 6%
Little to no productivity improvement: 11%

Due to rounding, percentages may not add up to 100%.

Postman v11 is here!

It's jam-packed with updates to help you collaborate on your APIs, augment yourself with AI, and more.

See what's inside v11 →