11 March 2026

Tech talk: How Bright Sites is helping to make newsrooms more efficient through AI solutions

AI solutions developer Bright Sites was, aptly enough, born in the beating heart of a news publisher. Creating bespoke platforms from a desk in the middle of newsrooms in the UK, founder Brian Alford became enamoured with the buzz of his surroundings.

“I really like the energy in newsrooms – they always want things yesterday or something happens and they suddenly want a bit of new functionality, like a map or a poll or just a different way of doing things,” he says.

Alford’s newsroom experience prompted him to pivot from building out CMSs to developing SaaS products that allow publishers to grow revenue and audience in sustainable ways through innovation, automation and data-driven workflows.

The products developed by Bright Sites include Storycue – AI-driven recommendations optimised by audience experts that help journalists pick the best stories to cover – and Flow, which is designed to reduce friction in content creation through AI-integrated features that make recommendations based on what works for each publication.

We sat down with Alford to talk about how Bright Sites products are changing the face of newsrooms.

What are the key elements to successfully integrating AI into a newsroom?

I think the technology has to earn its place in the newsroom. The game has changed and we can add features and buttons that do things really quickly now. But you actually end up spending a lot of time not doing that and instead really understanding the value of what we can do. For instance, we found great value in headline optimisation, but we kept on challenging ourselves. Like, at the first stage, you’re getting some results back because you’re putting it through some sort of LLM. Then we decided, actually, we want to be privacy first, so we have our own AI systems, which are only related to that publisher. And then we use all the analytics and start to put revenue data in there as well. So, it’s not just asking an LLM, which anyone can do. This has come from working with publishers. The more data is used in the whole system that’s unique to that publication, the better it is and you can fine tune it to be what you want. I think that’s how we learned to ship AI features that are safe for the publication, but also at scale.

To what extent is developing products a collaboration with publishers?

The way we’ve always worked is a close partnership. We’ve not distanced ourselves through a ticketing system where that’s the only communications. We still spend a lot of time in newsrooms – sitting with the editorial team and seeing things in real time. Sometimes it’s to gather ways of improving our products, but also to get a deep understanding of the urgency in the newsroom and exactly what they need. It’s about not just finding the solution, but finding a solution that’s reusable in an easy, a simplistic way.

What were the major obstacles for you as you pivoted from bespoke platforms to SaaS?

I think we were only limited by the time we had with our clients. We spend a lot of time not just adding features but actually debating the fastest editorial workflow way of doing things and finding the friction points. And, so, I think one of the challenges was just around finding time with publishers. You’ve got people working absolutely 150% of their time on writing articles, updating them etc, so trying to get time for us to ask our questions – to book a day or two-day workshop – it’s not necessarily easy. That’s where we took the initiative to actually go and sit by them. And I think that’s where we’ve gained an advantage – living the experience of the journalists and all the chaos within the newsroom.

---

---

How long did it take to earn the trust of publishers?

The honest answer is it takes a bit of time to be able to show what you can do and deliver the results. I think we’re very open about what we can do and what we’re not good at. We don’t sit behind shiny PowerPoint presentations. I’d much rather go and talk to people about how things are working, how things are not working. Where some of the trust comes from is that we are at the centre of what newsrooms are doing in terms of where the content is created

How does your relationship with leading tech brands help you?

We do a lot of integrations with third parties, whether it’s on the video front, or the image front. Obviously, there’s some AI integrations at the moment. And we also work with some of the big tech people, whether it’s through Facebook or Google. Last year we did two projects with Google, and we were featured at Google I/O, the developer conference. We partner with them and we are investing that time to give them the best opportunity get ahead of things like AI monetisation and subscriptions.

Could you tell me a bit more about Flow and Storycue?

Flow’s our flagship product. It’s been used for about 10 years or so. It’s a mature platform. It’s stable, but highly innovative and also highly customised to individual newsrooms ways of working. Storycue is an opportunity where it’s a low-barrier to entry tool for newsrooms. It opened up a whole range of new clients, which is interesting, not just from a commercial point of view, but also the fact that we are working with more newsrooms and facing different challenges. It’s really exciting. It’s being used by the likes of the Daily Mail, for example. It’s adding value and they see that. It’s basically uncovering opportunities for stories to write about, which they may have missed. It allows them to track various sources, whether it’s YouTube, Reddit, Google, whether it’s their own particular topics and niches they want to cover. And it processes all of that. This was really about being able to provide a stable audience assistance into the newsroom, whether it’s on the actual audience team or whether it’s given to every single journalist to use as an assistant.

What about the concerns that new tech will ultimately replace journalists in the newsroom?

Well, Storycue is not there to take jobs away from people. It automates what someone will be spending a lot of their time doing manually – so that they are then free to review and add additional value elsewhere. It’s helping them make smarter data-informed choices. One person can’t monitor every single source on the internet the whole time. So, this is about giving them the controls as well to be able to do that. Obviously, we are uncovering new areas which is recommended to them. They can control this and fine tune it to their particular needs and wants. And there’s a whole feedback loop within the platform. So, if they don’t want to cover a particular story and they don’t think this particular story matches when they download that, it gets fed back into the system and refines what’s produced next. We are using the client’s data that is shaped on their editorial style guide in terms of tone, and we deliberately don’t create the stories. I mean, that as a developer is the most obvious, easy thing, that’s the last step, you give them the story. But we deliberately don’t. That’s based on feedback from journalists because this isn’t an AI wires copy integration generative tool. It’s a newsroom efficiency tool, basically. So, if you don’t have an audience team, this gives you that support for every journalist. If you do, this can help support the audience team. Another important part was integration. We know how difficult it is to get people to open another tab, have another tool open, you’ve got their analytics in one place, they’ve got CMS, they’ve got previews. While it is a standalone tool, we do integrate with a variety of CMSs. So hopefully that integration makes it a bit more seamless as part of their workflow as opposed to another tool sitting outside of their platform.

How important is it to get feedback from journalists?

It’s absolutely crucial. That was actually one of the first features built into the platform – being able to get feedback in real time. So, how the platform is used, but also they can provide feedback on every single opportunity we recommend to them. And they can discuss whether it was either a great recommendation or not a great recommendation. So that feedback is really important, but that’s only on one level. The one that’s interesting to me as well is the actual direct face-to-face time we can have, sometimes just actually sitting down with publishers and understanding how it can help. Because again, this comes down to us enjoying being in the newsroom in that whole environment. Sometimes the issues that are being described or can be solved in multiple ways – Storycue can help with that.

What does the future of AI integration in newsrooms look like?

We’ve always taken the stance of being cautious and having tight controls around anything generated with AI. The last six months, the world’s evolved, and there are areas and methodologies where you can actually put more checks in with AI, it’s not just one-off process. We’ve got ways of building in fact-checking and reducing hallucinations. The latest stuff we’re working on is massive human control, but I think as we get confident with some of the parts, we are able to automate more or provide more information basically. It’s evolving. The next six months will be really interesting. I do think we have got to a place where, not everything, but certain bits can be trusted more than a year or so ago. And course, we’re monitoring everything that’s produced, we are putting that back into the feedback loop and everyone benefits from that.