After devastating floods claimed over 100 lives in Kerr County, Texas, our newsroom started investigating: could this tragedy have been prevented? The team narrowed in on public records of meetings by the Kerry County commissioners. The sheer volume of transcripts and a legacy filing system made a manual search impractical. But we quickly realized that AI could help. Using in-house built AI tools, we were able to uncover critical pieces of information central to the creation of  a landmark piece, published within days of the disaster.

To be clear, this was not a story written by AI. However, it was a story made possible by it. And it’s a perfect example of how we’re deliberately approaching AI at The Wall Street Journal – always to enhance journalism, not replace it. We spent the summer honing that deliberate approach with the goal of building knowledge, solutions, and a vision for AI’s newsroom role. 

⇒ Takeaway: Given generative AI’s transformative impact and the industry’s rapid change, newsroom leaders must address short-, mid- and long-term strategies all at once. 

NOW: Building a strong AI foundation

In the short term, we’re focused on ensuring our journalists know how to leverage AI and can do so in alignment with our newsroom AI guidelines. 

We rolled out enterprise AI tools and provided the standard company-wide “AI 101” training earlier this year. Yet even after months of effort, many reporters remained unsure how to use these tools effectively. I frequently heard questions like “How should I apply this in my job?” and “Where do I even start?” 

To address those concerns, we launched a summer series of lunch and learn sessions with the Newsroom AI task force to demonstrate how reporters and editors are already using AI in their work. Reporters walked away with role-specific tip sheets and more visibility on who is the AI evangelist in each department. Better yet, with the editor in chief in attendance, the sessions delivered a clear endorsement that the newsroom can and should be using these tools. 

 ⇒ Takeaway: We found that reporters need clear examples that demonstrate the return on investment for learning this new technology, especially as caution and uncertainty remain high. 

NEXT: Identify opportunities and embrace transformation

We scheduled brainstorming sessions with teams to identify elements of typical workflows that are rote and repetitive, more chore than creative. The key question we wanted to address was how can we free up time for journalists to focus on stuff only they can do? 

Concrete takeaways from these meetings are a prioritized list of workflow challenges. We identify places where new tooling or AI agents might help. We’ve now got a list of projects to prototype in the newsroom and to collaborate on with our tech and engineering colleagues.

Zooming even further out, we hosted a series of sessions with each coverage area that focused on how our newsroom can evolve to anticipate changes in reader behavior and expectations. The goal was to think big: Given what ChatGPT, Gemini and Claude can already do, how do our workflows evolve? How should our coverage change?

The industry is at a massive inflection point in the way audiences discover, digest and manage information – and that’s our business. While we won’t always get it exactly right, we do need to anticipate audience behavioral changes to stay ahead. Those conversations have never felt more urgent. Our future-proofing sessions clarified the need to focus on direct relationships with readers and creating unique, human-centered storytelling experiences that AI can’t replicate. 

⇒ Takeaway: Executive-led sessions reinforced that AI is a top priority – and every journalist should treat it with the same level of urgency. 

Our AI roadmap

As we shape our strategy and roadmap, we’re actively harnessing AI to deliver impactful journalism. AI continues to be a force multiplier for our newsroom, increasing speed and discoverability in areas and formats that have historically been hard to cover by a single reporter. 

Our data journalism team built a tool that does just that. Nicknamed “Orca,” this tool helps reporters digest and summarize conversations happening in podcasts. Where previously a reporter would have to listen to every episode, tallying hundreds of hours on a particular topic or show, now Orca leverages AI to transcribe conversations into text that can be searched, summarized or quantified. Orca was the engine behind this piece, which analyzed how conservative podcasters have discussed the Jeffrey Epstein case. Orca listened to more than 22,100 episodes from 148 podcasts, obviously a feat previously unfathomable by even a large team of reporters. 

⇒ Takeaway: Roadmaps and cross team alignment can take time. Let’s not keep that from holding back areas where AI can move at speed and make a real impact— our journalism.  

Embracing AI wisely means moving with both urgency and care – deploying it where it can supercharge reporting today, while building the culture and strategy to guide its long-term role. At The Wall Street Journal, our guiding principle is clear: using AI to enhance our journalism, not to replace it. By investing in training, rethinking workflows and pushing ourselves to imagine what’s next, we’re ensuring that our distinctive, exclusive and human-led journalism remains at the heart of every story we publish.