Sometimes building in public means asking people what they think of what you’ve made and sometimes it means listening closely enough to change it.
In late July, we brought journalists, editors and donors from across the US together for a day of conversation about AI, journalism and the 2026 midterms. We wanted to talk openly about what is working, what is not, and what newsrooms actually need from AI tools at a moment when trust, accuracy and access to reliable information matter enormously.
The conversation was not exactly a love letter to AI.
There was excitement about what these systems might make possible, especially tools that help newsrooms make use of their own reporting, but there was also skepticism, informed by plenty of experiments (and a few disasters) about what happens when AI systems are deployed without enough transparency, oversight or connection to the communities they are meant to serve.
At Meedan, we know the tension well. We have spent more than a decade working with journalists and civil society organizations on information integrity. One thing we have learned is that the problem is rarely a complete absence of good information. More often, valuable information already exists inside local newsrooms and organizations but is difficult to surface at the moment somebody needs it.
Elections make that problem particularly visible: Where do I vote? What is the registration deadline in my county? Who is running in my district? These questions have highly local answers and getting them wrong can have real consequences.
Journalists at our convening told us they were already seeing general-purpose AI systems struggle with answering precisely these types of questions correctly, including in some of the most heavily covered parts of the country (such as New York). Meanwhile, local newsrooms may already have the correct information, but are faced with the challenge of getting it into people’s hands.
That led us to another question: what does meeting people where they are actually look like in the United States?
The channel is part of the product
Suwali was never intended to be a chatbot that lives in one place: we designed it as infrastructure that allows organizations to connect their own trusted information to the channels their audiences already use, which varies not only from country to country, but also depends on demographic, social and other types of audience characteristics, as well as their habits. A newsroom serving readers in São Paulo may make a different choice from one serving voters in Ohio, while a health organization may need a different channel again.
Suwali already works over WhatsApp and Telegram, channels that make sense for many of the organizations and communities we work with. But as we started talking to newsrooms in the US, we heard a consistent piece of feedback: those channels alone would not be enough to reach the audiences they had in mind.
According to data compiled by the Pew Research Center, only 32% of US adults use WhatsApp, whereas SMS is one of the most ubiquitous channels for reaching a broad domestic audience. New research from YouGov shows that standard text messaging has a 2% non-use rate among Americans.
There is also ample evidence that US publishers are already using SMS as a way to reach their core audiences in a “more personal, urgent way on their most important devices, calling it an “untapped” means of distribution for pushing great, scoopy, exclusive reporting.
For election information, that simplicity is especially compelling: a reader can text a newsroom’s number, ask a question and receive an answer grounded in information the newsroom trusts. They can subscribe to updates, receive alerts and follow links back to the reporting behind the response.
There is no new app between the newsroom and the reader and no algorithm deciding whether the information appears in someone’s feed.
So what does this mean for Suwali?
Ahead of the 2026 midterms, newsroom partners in the US will be able to offer readers a phone number they can text to ask questions, receive alerts and access reliable election information grounded in the newsroom’s own reporting and sources.
But SMS is also part of a broader experiment: we want to build Suwali alongside the organizations using it, rather than deciding in advance what every newsroom or every audience needs. That means putting prototypes in front of people, listening when something does not fit, and treating distribution as part of the design problem rather than something to solve after the technology is finished.
The takeaway from our US conversations was straightforward: if you want to build tools that serve different publics, you also have to build for the different ways those publics communicate.
If your newsroom is thinking about how to make reliable election information easier for people to reach ahead of the 2026 midterms, we’d like to build with you.
Contact us or sign up for a Suwali trial to test SMS with your newsroom.
