Velora co-founders Peter Stuart and Danny Bellion
A former editor of Cycling News has launched a new largely automated sports news brand.
Velora Cycling and the company behind it, Velora Digital, have been launched by Peter Stuart, who edited Future -owned Cycling News from March 2022 until last month, and Danny Bellion, who was previously head of AI at fintech company Capital on Tap.
Stuart told Press Gazette Velora Cycling, which is now live in beta , would be a trial run of the concept. If it works, he and his co-founder plan to launch sites covering other sports and could even white-label the platform for use by other publishers.
Stuart begins as the only member of editorial staff, producing coverage of competitive cycling and the latest gear on his own.
He is assisted by the platform’s AI systems on everything from researching stories and writing them to uploading and publishing online.
However, Stuart insisted that human editorial expertise will remain an essential part of how articles are produced, even as more of the process is automated.
“You don’t want it to be totally writing end to end,” he said. “You need to have some human editor interaction.”
‘Taking ownership’ of inevitable AI changes
While many companies are already introducing AI into the newsroom, recent Reuters Institute research suggested the promise that the technology will reduce low-level tasks has yet to be fulfilled.
Nevertheless, Stuart believes businesses will inevitably need to adapt.
“I can’t see a future where journalists ten years from now are doing the exact same content in the exact same way,” Stuart said.
“I think it’s going to have to change. So it’s about hopefully taking ownership of it and putting it in a decent, forward-facing trajectory.”
As a digital editor, Stuart witnessed the rise of AI content farms that are now hoovering up traffic with low-quality output, often repackaging existing articles .
Stuart began to think about how newsrooms could use the technology to produce more and better stories, and fight back against the slop sites.
“I think the best immediate thing we can do is to be much better about automation,” he said.
How Velora works
An obvious first candidate for that automation is the content management systems (CMS) that most publishers use to get stories onto websites.
Stuart argues that the repetitive tasks involved with uploading stories, such as selecting an image, adding alt text and inserting links, can drain time.
With Velora Digital’s platform, writers can instead send copy via Slack or email and the system will make it ready for publication.
But the system can be used in all parts of the news production process, including flagging trends, sharpening angles and even writing drafts.
A typical workflow for Stuart, as the solo editor of Velora Cycling, might start with an idea selected from the system’s identified trends. He can then feed the concept into the system’s features pipeline, which fleshes it out with background research and source validation, a process he said he and Bellion have fine-tuned to be in line with the desired output.
From there, the platform can draft a full article, but Stuart can refine it, using both his knowledge of the subject and editorial nous to shape the story based on the assembled research.
“I know that ultimately there’s a huge amount of cynicism, quite rightly, about the idea that AI is now automating stuff to really low quality,” he said.
“I think the most important thing is the written content that’s produced by this is strong content that adds something rather than leeching off what someone else has written.”
Velora’s founders see an opportunity to push new boundaries in how stories are presented, harnessing AI to create interactive features.
In the test case of cycling news, that could include visual elements that incorporate live data about events.
“If it came out and like, two or three months from now, it was identical to what is on the market, I would feel quite disappointed by that,” Stuart said.
He added: “I’m only really interested if it genuinely offers something beyond what publishers are able to do.”
Launching fast, learning fast
Both Stuart and Bellion have only been working full-time on the project since the start of November but they have built all the tools themselves.
They used models from all the major AI labs and tested them against each other to find the best fit for the various tasks.
The features have been built “super quickly, very, very low cost” Stuart said. The business hasn’t yet sought any external investors.
Potential revenue streams are programmatic advertising and affiliate links, as well as the anticipated future play of selling the tech to other publishers, but these will not be part of the initial launch.
Stuart emphasised that the project was still a testing ground. He also recognised that any form of AI in journalism will face scepticism.
“The key thing is that journalists are the ones that take ownership of how automation happens, rather than having publishers dictate [it],” he said.
Email [email protected] to point out mistakes, provide story tips or send in a letter for publication on our “Letters Page” blog