As 2025 draws to a close, Jodie Hopperton, lead of INMA’s Product & Tech Initiative, offered her take on what the industry has learned in 2025 and where it is headed.

During this week’s virtual Product & Tech Town Hall, Hopperton distilled a year of experimentation, disruption, and recalibration into key learnings.

10 key product and tech takeaways from 2025

1. AI is now core infrastructure.

Hopperton began with what she called the most obvious yet most transformative shift: AI is no longer an experiment but a foundational business system.

“We are well out of the experimentation phrase, and we are finding that the one thing executives regret is moving too slowly,” she said. “They’ve really got to get this across the organisation.”

AI is now viewed as part of the infrastructure underpinning operations, strategy, and product development. 

2. Answers are replacing search.

One significant change during 2025 was the change in how users get information: “It’s really a change in behaviour, and it’s not just in search,” Hopperton said, explaining that consumers increasingly turn to answer engines rather than traditional search.

“I use ChatGPT way more than Google now,” she said. This change demands news publishers rethink their strategies: Should they now optimise for answer engines rather than SEO?

The rise of “answer engine optimisation” looms large over the coming year and promises to raise more questions.

3. Search traffic decline is structural.

Search traffic decline, Hopperton emphasised, is not cyclical but structural. As users get answers directly from AI interfaces, they no longer click through to external links.

“If people are going to these answer engines, they’re getting answers, they’re not getting links to go elsewhere. So we need to start thinking about Google Zero,” she reminded.

It is imperative news companies find ways to shift away from dependency on search referrals.

4. Agentic Internet will reshape distribution.

Hopperton highlighted the emergence of agentic AI, which moves beyond providing answers to executing actions. She illustrated this with examples: asking for an address (search), receiving directions (answer), and booking a ride (action).

“It’s a new layer of platform dependency, it’s a new way of thinking about consumer behaviour, and that’s something that's going to be very big for us,” she predicted. “We’re seeing the beginnings of it in 2025, but this is really going to come through in 2026.”

5. Content protection and licensing need governance.

Scraping and licensing have reemerged as critical issues. Tools like Cloudflare’s bot management are being deployed to monitor and block scrapers.

Hopperton urged publishers to establish bot governance frameworks to protect content and manage who accesses it: “That’s the one thing you absolutely need to do today because otherwise all of your content is being scraped and being put elsewhere.”

6. Workflow is the real disruption.

The real disruption, Hopperton argued, is not in CMS replacement but in workflows. AI augments rather than replaces systems, reshaping how people work.

“It’s partly about people, and partly we’re adding an extra layer,” she explained. “We are seeing automation [and] new tools come in … but it’s really about people.”

7. Distribution will happen through AI interfaces.

Distribution continues to fragment. Beyond Web sites, apps, social media, and newsletters, Hopperton pointed to answer engines as new distribution channels.

ChatGPT’s app integrations with brands like Zillow and Canva show how AI platforms are becoming gateways for consumers using a visual interface, and Hopperton said that’s something publishers can’t ignore: “How can we start thinking about [if this is] a useful platform for us for distribution?”

8. Personalisation is the default expectation.

Personalisation is now the default expectation, Hopperton said, but she cautioned against overuse.

Whilst personalised content often outperforms static experiences, shared experiences remain vital. She cited the success of Wordle at The New York Times, which fosters communal engagement.

News, too, requires shared facts. The challenge is balancing personalisation with collective experiences that build community.

9. AI creates continuous feedback loops.

Hopperton described the acceleration of feedback cycles.

“When we started the Product Initiative five years ago, we looked at that real product process, how you come up with an idea,” she recalled. “You test your hypothesis, you prototype again, you test it, and then you might launch a beta version, and then you iterate.”

However, with AI, news publishers have entered an era of fast, continuous feedback that enables iterative refinement. Strategies must be iterative because inputs, products, and consumer behaviours change rapidly, making agility essential.

10. Publishers compete with AI products.

News companies are no longer just competing with one another; they’re competing with smart chat and AI products, Hopperton said.

“So we are seeing user loyalty shift from brands to interfaces and formats because they serve needs better,” she said. “So again, there’s some ways we can look at that branding and staying front of mind, but that’s something we’ve really got to think about.”

Media Tech & AI Week takeaways

Hopperton also shared her experience at INMA’s Media Tech & AI Week in San Francisco. The event combined a three-day study tour with a two-day conference, bringing together major tech companies, startups, and media practitioners.

For Hopperton, the week brought one significant realisation: “The Internet is effectively forking at the moment,” she said. “We have the human Internet, the Internet that we know we produce for humans, but there’s also this underlying agentic Internet.”

Even in the Internet built for human consumption, AI-generated content has overtaken human-created content for the first time. Hopperton saw this as both a challenge and an opportunity.

Verified journalism, brand-forward storytelling, and truth-based reporting could become more valuable assets in a sea of synthetic content. Building direct audience relationships, fostering habits, and prioritising loyalty over traffic are essential strategies for thriving in the human internet.

Beneath the surface, Hopperton described an “agentic Internet” populated by bots and agents. Scrapers feed foundational models, and AI systems increasingly interact with content directly.

Whether publishers like it or not, they are serving bots as well as humans. Preparing for this requires structuring data differently, capturing information at granular levels, and considering monetisation opportunities in agentic ecosystems.

One of the lessons Hopperton emphasised was to begin with small agentic workflows and practical automation. Improving onsite search or streamlining processes are manageable entry points. Handing entire businesses to AI is premature; incremental steps are safer and more effective.

The pace of change was palpable during the study tour; Hopperton recounted visiting OpenAI just as it launched an AI browser, underscoring how difficult it is to keep up with updates.

“We see these AI companies working so hard right now for that land grab,” she said. “They want to be the best product out there.” The pace is only going to escalate, and she warned publishers that they need to be aware of how rapidly the environment changes.

The week opened Hopperton’s eyes and left her with a sense of urgency. The forked Internet — human and agentic — requires publishers to rethink strategies, structure data, and prepare for agentic AI.

Looking forward, Hopperton predicted 2026 will be the year the industry fully confronts its place in the AI-driven ecosystem. Fragmented distribution, evolving product formats, and the rise of agentic AI will demand new strategies.

“We are really going to have to prioritise customer experience [and] direct people-first relationships over chasing traffic while simultaneously really getting to grips with AI,” she said.

News companies will experiment with agents and new monetisation models, and for INMA’s Product & Tech Initiative, the focus will be on helping members prepare for this shift whilst embracing AI as both competitor and collaborator.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .