Case Study: Transforming Workflows with AI at Zamaneh Media


Published September 20, 2024

AI in the Newsroom is a series of case studies from the Online News Association (ONA) that highlight specific ways journalists are building and using AI tools.

Zamaneh Media, a small Dutch-based newsroom focusing on Persian-language content, embraced AI to overcome challenges in news production and translation. The newsroom developed two AI-driven tools—Newsletter Hero and Samurai—that significantly streamlined their workflows. Despite a small team of just two people with limited technical backgrounds, they managed to innovate and improve the newsroom’s efficiency by reducing the time spent on routine tasks like newsletter creation and translating long Persian articles into English.

Opportunity

With a team of 13-15 journalists, Zamaneh faced significant challenges in content creation due to limited resources. The weekly task of producing newsletters was particularly burdensome, leading to frustration among the staff. Additionally, as a Persian-language publication with an English counterpart, Zamaneh struggled to efficiently translate and adapt its content for a broader audience. The existing translation tools, such as Google Translate, often produced inaccurate and misleading results, underscoring the need for a more reliable solution.

Two solutions

1. Newsletter Hero: The idea for Newsletter Hero emerged during Zamaneh’s participation in AIJC, Open Society Foundation’s AI in Journalism Challenge. Given the team’s limited technical expertise, they opted for a no-code approach, focusing on prompt engineering rather than heavy coding. The tool was designed to simplify the newsletter creation process by automating the generation of content such as summaries, Q&A sections, subject lines, and preview texts. The primary focus was on creating prompts that could be fed into GPT-4 to generate the desired outputs efficiently.

A view of Zamaneh’s Newsletter Hero content generation tool.

2. Summarization Tool (Samurai): Following the success and learnings from Newsletter Hero, Zamaneh turned their attention to a more pressing challenge: efficiently translating and summarizing long-form Persian content for their English-language audience. The summarization tool, nicknamed “Samurai,” was developed as a more streamlined and focused AI tool. It aimed to automatically generate concise summaries of Persian articles, which could then be translated into English and published on Zamaneh’s English website.

A view of Zamaneh’s Samurai translation tool.

Outcomes

Newsletter Hero proved to be a significant time-saver, reducing the newsletter creation time from almost a full day to just over an hour. However, its development was eventually paused. The team found that integrating the tool into their existing workflows required more manual effort than initially anticipated, which led them to focus on a higher-priority project. Nonetheless, the experience and knowledge gained from this project laid the groundwork for the development of Samurai.

On the other hand, Samurai significantly improved Zamaneh’s ability to produce English-language content. The time to publish translated articles was reduced from days to less than an hour per article. This efficiency allowed Zamaneh to update their English website more frequently, increasing engagement and expanding their reach to non-Persian-speaking audiences. The summarization tool also showed potential as a marketable product, with Zamaneh considering building it into a tool to offer other small media organizations facing similar challenges.

Lessons learned

  • AI Adoption Requires Patience and Collaboration: Initially, there was skepticism among Zamaneh’s journalists regarding AI’s role in their work. By involving the newsroom in the AI development process and demonstrating tangible benefits, the team gradually built trust and acceptance.

  • Workflow Innovation is Key: Zamaneh’s success was not rooted in cutting-edge AI technology but in their ability to innovate within their existing workflows. By focusing on practical, no-code solutions, they were able to create tools that met their specific needs without requiring extensive technical expertise.

  • Continuous Prompt Refinement: Regularly updating and refining AI prompts is essential to maintain the quality and relevance of AI-generated content. Zamaneh’s experience highlights the need for ongoing maintenance and a systematic approach to improving AI performance.

  • Language-Specific Challenges: Working with AI in less commonly supported languages like Persian presented unique challenges. The team’s ability to customize and refine AI outputs was crucial in overcoming these obstacles and producing high-quality content.

Zamaneh’s experience demonstrates that even small teams could leverage AI to solve specific, real-world problems, achieving significant impact with limited resources.

Hear more from the team

We’re hosting the Zamaneh Media team for a “how we built it” learning event on Oct. 9. Attendees had the opportunity to discuss these tools, ask questions and hear what they’re up to next. Explore the archives for ONA’s AI in Journalism Initiative trainings to watch this recording and access dozens more resources. (Note: You must be logged in to your ONA member account to access the page.)

This resource is part of the AI in the Newsroom series. Read other case studies you might have missed:

Related Topics
Author
Nikita Roy
ICFJ Knight Fellow and Newsroom Robots Podcast Host