How UK’s Reach is using AI to help produce more content faster

Over the past two years, Reach has focused on integrating AI tools into its newsroom processes, from content creation and editing to distribution and audience engagement, noted David Higgerson, Chief Digital Publisher, at our Newsroom Summitin Zurich this week.

Navigating revenue and relevance online

Unlike many publishers increasingly adopting subscription models, Reach remains committed to offering free access to its journalism. "We believe our journalism should be free to anybody who wants to read it; it's the only way forward," Higgerson said.

He pointed to Reuters data showing that only 9 percent of the people in the UK are willing to pay for news, as one of the reasons for keeping their journalism free. He also cited the presence of the BBC as a significant challenge in convincing readers to pay for content, since it is a publicly funded, but otherwise freely and widely available service.

However, the company does share some challenges with other publishers. In addition to revenue pressures, it faces a long-standing relevance issue.

"We need to stay connected with our readers throughout the day and constantly reassess what being digital-first really means," Higgerson said.

"Currently, this involves making data-driven decisions and rebuilding the real-world connections that once defined their brand, aiming to transcend platforms."

Exploring AI's role in free journalism

Reach has spent the past two years exploring how AI can support its commitment to free news.

They began by determining how AI could assist its journalists without replacing them, ensuring reporters retain editorial control.

"As our newsrooms have contracted, we've focused on allowing journalists to do what they do best - getting stories," Higgerson said.

And so, while the number of journalists has decreased during the past four years, Reach now has more reporters on the street gathering original stories, supported by AI.

Initially apprehensive about introducing AI, Higgerson found that journalists were already using a variety of AI tools independently. Some were adopted company-wide, while others were discontinued due to security concerns.

He noted that the build-or-buy decision on AI initiatives has been a critical one. While the product team often prefers to build in-house tools, the company is also exploring options from external partners and suppliers.

"Though much of the conversation has focused on generative AI, some of Reach's biggest successes have come from deploying other AI technologies," Higgerson added.

Key AI tools driving efficiency

AI tools are also helping the organisation streamline content distribution and amplification, which had become fragmented.

Reach adopted a "fail fast" approach, experimenting quickly across its 3,000 employees in 15 hubs in the UK, Ireland, and the US; collaboration across departments helped accelerate AI integration.

Two key tools have driven the company's AI efforts:

The first, Guten, was developed in collaboration with newsrooms, data science, and product teams to streamline tasks that slowed down their journalism.

The second, Mantis, largely built by the commercial team, is a brand safety and revenue generation tool with editorial uses. Now offered as a SaaS product, it has attracted publishers globally.

It streamlines editorial processes by transforming proprietary data sources, like public notices, into easily producible online content. While it automates much of the initial work, these items still require editorial review before publication, improving the traditionally slow and labour-intensive process for journalists.

Streamlining content production: From 9 minutes to 90 seconds

Guten has played a crucial role in streamlining Reach's content production. In response to newsroom feedback about the time it takes to processcontent from trusted sources, the publisher developed Guten to accelerate this workflow.

With 42 million monthly UK readers, Reach relies on high story output to drive advertising revenue, producing around 4,500 stories daily with a team of 1,800 journalists.

Guten helps by processing content from trusted sources, such as the Press Association and other publishers, and converting it into stories formatted in-house. These stories can then be quickly reviewed by editors and reporters for publication.

Guten has contributed significantly to Reach's output, generating an additional 1 billion pages across the network, and now supports around 25 percent of the output.

The speed of publishing breaking news has significantly improved, dropping from nine minutes to just 90 seconds with Guten after the shift to remote work post-pandemic.

The organisation is integrating AI into each step of its storytelling process - sourcing, creation, editing, publishing, distribution, amplification, and assessing story effectiveness.

Sourcing, a crucial area, is where the company aims to maximise journalists' time. To achieve this, Reach is exploring ways AI can enhance sourcing efforts, allowing reporters to focus more on their core journalistic duties.

Leveraging AI partnerships for competitive advantage

Since establishing its data journalism team in 2012, Reach has strengthened its data journalism by using AI to speed up analysis, helping its team handle larger data sets and uncover impactful stories.

The publisher uses AI-enabled solutions from brands like Data Miner and Futuri to enhance story sourcing. Futuri, an American company specialising in data-driven trend predictions. "This gives us a competitive advantage in the US market, where we are relatively new compared to other UK publishers. This partnership has already yielded significant success," Higgerson said.

Trusted data and information are then input into Guten to create reliable news content. "This process took time to refine, with efforts to incorporate local dialects into stories-a departure from traditional newspaper practices," he said.

Additionally, Reach uses Trint, an AI transcription tool, which saves sports journalists' time and boosts story output.

One key AI focus has been improving the content editing process.

"As we've shrunk our newsrooms and squeezed management, content editors have faced increased pressure to manage more tasks, similar to reporters," he said. "Guten has streamlined this process by handling stories that no longer require reporters' involvement," Higgerson said.

Optimising content creation and distribution

Reach is also collaborating with United Robots to enhance community engagement at a hyper-local level, "which has been challenging due to staff cuts over the years," he said.

An experiment on the MyLondon website involves using property postcode data to highlight average house prices. All content produced with United Robots is reviewed before publication

Many content editors have benefited from enhanced editing and publishing support following the merger of these functions. Guten plays a central role in this improvement.

Higgerson said newsrooms aim to stay relevant to their target audience, with Guten adapting one story into 15 versions for different local sites, avoiding rewrites.

"This approach attracts new visitors to the websites, who can then be directed to the local journalism that we prioritise," he said. "As a result, we've seen a massive reduction in content duplication, enabling journalists to produce more original stories."

Another collaboration is one with Echo Box to integrate AI into story selection for Meta and X, with plans to expand this into newsletters as well.

While content recommendation is not new, Reach is enhancing its "next best action" tool using AI to provide users with personalised content, rather than repetitive suggestions.

Higgerson said amplifying content is crucial for reaching a wider audience across various titles and retaining user engagement. And so, the organisation has partnered with Flora to utilise a more engaging and safe commenting community.

"In the assessment process, Reach faces the challenge of navigating an overwhelming amount of data, often leading to a focus on data points rather than actionable insights. While Chartbeat highlights the top stories, the company aims to identify trends in the middle tier of its 4,500 daily articles, such as the popularity of online death notices," he said.

Using tools like Gemini and insights from Chartbeat and Google Analytics, Reach is focusing on replicating successful middle-tier stories across its network.

Looking ahead, Higgerson said they are working on improving patience in managing new vendor inquiries and observing competitor strategies, recognising the need to focus on long-term improvements rather than immediate trends.

Additionally, the newsrooms see significant potential in audio and video content, though it is currently taking cautious steps in this area.

"Younger audiences, particularly on TikTok, are quick to criticise poor AI use, presenting an opportunity for us to enhance our offerings,"

"We aim to strengthen hyper-local partnerships and establish ourselves as the go-to source for local news," he said.

'A continuous journey': Lessons learned

In closing, Higgerson listed the key lessons they have learned in integrating AI in the Reach newsroom:

• While journalists are embracing AI's potential, the greater risk lies in its hasty implementation. The "build-or-buy" decision is crucial, with a dedicated product team working to strike the right balance in partnerships.
• Transparency is vital; readers expect to be informed about AI usage. "Reach has committed to ensuring that all content is human-approved before publication, clearly outlining its AI practices on its websites," he said.
• Despite advancements, Reach knows it is still at the beginning of its journey with AI.

"We emphasise the need for adaptability in our operations, acknowledging that the ability to pivot is crucial. This flexibility must align with our core mission of fostering informed communities while remaining true to our values," he said.

Neha Gupta

Multimedia Journalist

[email protected]