Q3: Steering Group Meeting Notes

Published: 02 Dec 2025

Ad Ops

• The meeting opened with a call for a new co-chair for the group going forwards – if you are interested in finding out more, please get in touch.
• The group shared updates on their current focuses, including internal system improvements, global standardisation efforts following recent acquisitions, ongoing testing of brand safety tools and commercial AI transformation projects aimed at improving campaign performance and operational efficiency.
• AI adoption across the group continues to grow, with teams using integrated chatbots to streamline onboarding processes and answer routine staff queries more efficiently. Microsoft Copilot and SharePoint are being utilised to build comprehensive internal training and information-sharing platforms. Scribe was noted as a tool that automatically records screen interactions to create step-by-step guides without compromising data confidentiality. Early trials of AI audience-building tools are being conducted, with a focus on those that allow for cross-campaign data analysis and more granular audience modelling, helping teams to refine targeting and drive better commercial outcomes.
• Data privacy and security remain top priorities, particularly regarding the use of generative AI platforms where concerns exist around the exposure of proprietary data. The group discussed exploring proprietary solutions such as Vertex, alongside developing internal guidelines to clarify appropriate data usage. While some members of the group are piloting data clean room solutions, inconsistent adoption across the industry is creating operational friction for agencies and publishers alike.
• Challenges related to IVT were raised, particularly the impact of bots and crawlers causing inflated traffic metrics and potential revenue losses. Members of the group highlighted the need for clearer industry standards and more effective detection technologies to prevent unnecessary bot activity from undermining campaign performance, noting that generative AI discussions increasingly cite these issues as a concern for ad effectiveness and transparency.
• Broader regulatory developments were reviewed, including expected government exemptions for low-risk advertising technologies scheduled for early next year. The group debated how these changes might influence industry practices and technology platforms, with some optimism about potential relief from restrictive brand safety regulations, though a cautious stance remains regarding the speed and clarity of forthcoming guidance from regulators.
• Members also shared insights about newly launched ad targeting solutions in the market that leverages AI to optimise audience reach and engagement, with ongoing monitoring of how these innovations integrate with existing workflows and data governance requirements. This link was shared. There was an emphasis on maintaining a balance between innovation and compliance, as well as ensuring that data security and brand safety remain central to all AI-enabled advertising activities.

Audience Development

• The group discussed the evolving impact of Google's recent core and spam updates, with highly varied results across publishers. Some reported strong traffic gains driven by local or niche content, particularly through Discover, while others experienced sharp declines, including significant losses of Discover visibility. Many noted a lack of clarity around what’s driving performance shifts, with speculation that Google has reversed its previous emphasis on ‘helpful content’ and may now be favouring hyper-specific or spam sites. Concerns were raised about the spread of low-quality content in Discover feeds, while others highlighted the growing dominance of platforms like YouTube in both search and AI summaries. There was consensus that visibility has become harder to predict and measure, particularly across mobile vs desktop and regional markets.
• The introduction of Google’s preferred sources feature in the U.S. was met with confusion and scepticism. This LinkedIn post was shared. While some publishers are planning to promote it to American audiences, others questioned the feature’s purpose and potential implications for traffic. Measurement remains a challenge, with little data currently available on its impact. The group discussed whether the feature might reflect deeper strategic moves from Google, including a shift toward favouring known or commercial partners in AI environments. Ther were concerns raised about the feature being used selectively or opaquely. There was also discussion of the CMA’s upcoming decision on whether to designate Google as having Strategic Market Status, but expectations for significant change were low, based on previous regulatory efforts globally.
• Several publishers shared updates on platform monetisation and performance across Meta, TikTok, and emerging newsletter strategies. Some have seen revenue success from Meta’s Facebook monetisation scheme, despite suspicions around its intent and the lack of traffic impact. TikTok’s link pilot was reported to be growing steadily, with speed and volume cited as key factors in success. Others noted that testing on platforms like Reels yielded little actionable insight, while experimentation with Substack was limited. Some publishers are focusing on launching owned newsletter products as a more sustainable alternative to third-party platforms. Overall, publishers remain cautious but curious about platform-based revenue streams and are reallocating resources selectively.
• The perceived usefulness of AI tools for content analysis and visibility remains limited, with many publishers reporting underwhelming results. Tools like Profound were described as expensive and only moderately useful – good for spotting old or unexpected content showing up in AI summaries but lacking depth in actionable insights. Some teams are still exploring alternatives or conducting manual audits instead. There was agreement that these tools may only be worthwhile for larger operations or in high-stakes verticals like health or finance, where AI citations can drive reputation and traffic. Others expressed concern that the tools are being used as motivational crutches for work that should already be best practice.
• AI overviews and AI Mode continue to affect content visibility and user journeys in subtle but important ways. While most publishers reported that direct traffic from AI summaries remains low, several noticed changes in branded query performance and a decoupling of impressions and clicks, particularly for evergreen or travel-related content. There is a growing belief that AI experiences will continue to expand, even if they currently offer limited conversion value. Demographic data shared by the group indicated that younger users are more likely to trust AI summaries, suggesting a longer-term shift in how content is consumed and attributed.
• Concerns were raised about leadership changes at Google and what they mean for publisher communication. Danny Sullivan’s role has changed so there is uncertainty around who the new point of contact will be. The group shared anecdotal reports that Google may be hiring a new communications lead, but the structure remains unclear. Many publishers remain frustrated by the lack of transparency and the broader implications of Google consolidating control over how news content is surfaced, ranked, and monetised. Several group members expressed interest in pushing for more direct lines of communication and clearer guidance, especially as AI continues to reshape discovery.

B2B

• Group members discussed planning for 2026 with a focus on product development, AI integration, and operational efficiency. Several are investing in infrastructure upgrades such as data lakes and new delivery platforms and others are rebuilding tech teams and adopting new marketing platforms. Confidence in certain sectors (e.g. insurance, accountancy) remains high, supporting stable digital and events growth forecasts. Some are converting editorial sites into data products, implementing AI-assisted tools to free up editorial resource, and planning international expansion. Event businesses are being rebuilt post-COVID, with a mix of cautious scaling and renewed ambition.
• AI experimentation is widespread, with varying levels of maturity. Use cases include editorial support, content summarisation, proprietary data processing, AI-assisted translations, and audience personalisation. Some are working with specialist vendors to improve AI-powered search or recommendation tools, while others are training LLMs to replicate specific editorial tones. Search usage remains low in some cases, while others report improved content engagement from AI-driven discovery. There’s a general move towards empowering editorial teams with opt-in AI tools, alongside caution around trust, transparency, and data protection when engaging with external AI platforms.
• No significant traffic impact from AI search has been reported across the group, though concern exists about longer-term risks around user acquisition and content security. Most B2B publishers have not experienced the same volatility as consumer or affiliate-driven sites, though individual cases note threats as search behaviours shift. Some interpret recent platform updates as favourable to niche content providers, while others remain wary of increasing reliance on volatile referral channels.
• Geo-expansion remains a viable growth strategy for events, with multiple publishers adapting successful formats into new regions. Some are testing markets like the Middle East, Asia, and Australia, focusing on scalable topics such as water, energy, and finance. Success depends heavily on local market knowledge and resourcing. In some cases, localisation is aided by AI translation followed by human review. Others are prioritising replicable event models to strengthen positioning for future acquisition.
• On the topic of AI resourcing, members of the group are not hiring dedicated AI specialists at this stage. Approaches include upskilling internally, working with consultants, or maintaining a cautious stance while monitoring developments. There's consensus on the value of continued training and strategic experimentation, rather than large-scale hiring. As AI becomes more embedded in operational workflows, teams are balancing innovation with editorial standards and audience trust.

Commercial: Programmatic

• We are looking for a new co-chair for the group going forwards – if you are interested in finding out more, please get in touch.
• Members of the group shared their Q4 priorities and early planning for 2026, with a strong focus on video, sponsorship, and PMP strategies. Budgeting for 2026 is already underway for some, with Q4 efforts concentrating on external sales. Several publishers reported strong Q3 performance and a positive outlook for the remainder of the year. Video remains a key area of investment, particularly across news, audio, and branded formats. PMP performance is mixed; some have successfully scaled by integrating it into programmatic teams and securing global deals. Curated marketplaces were highlighted as gaining traction, and one-to-one PMP deals remained strong over the summer. Podcast-led video is emerging as a growth area, with content often published to third-party platforms and branded sponsorships becoming more common.
• The future of programmatic trading, especially with the integration of AI-driven solutions, was met with scepticism. The group noted that many current implementations appear overhyped, with limited practical impact at this stage. Several likened the discussion to earlier waves of programmatic, conceptually promising, but still dependent on foundational factors like quality inventory, clean signals, and technical hygiene. Publishers are continuing to explore in-house solutions, particularly for video, but are cautious about such innovations unless they bring clear commercial benefit.
• There was an extended discussion around cookie compliance and regulatory changes, particularly related to Limited Ads and new consent frameworks. AOP shared updates on discussions, noting that proposed amendments to reduce consent requirements for low-risk advertising may be tabled in early 2026. Whilst the current stance appears lenient around first-party data used to enhance on-site user experience, publishers raised concerns over lack of visibility on cookie behaviours from certain partners, particularly when users select reject all. Some have paused direct campaigns under these circumstances due to compliance risk and lack of clarity.
• Verification tools and measurement platforms continue to pose challenges, both in terms of cost and effectiveness. Some publishers are under internal pressure to reduce spend on tools that don’t generate direct revenue. Others highlighted issues with transparency, as some platforms may still be dropping cookies despite user opt-outs. The group agreed that brand safety & quality assurance remain important, and it was suggested that clearer validation and more consistent enforcement by vendors is required.
• Google’s Limited Ads were described as difficult to manage, particularly for direct campaigns. Publishers cited unclear guidance around frequency capping, cross-region delivery, and the exact nature of fraud prevention cookies. Some have attempted direct campaigns in limited mode but found performance lacking, particularly for multi-geo delivery. Requests for clearer documentation from platform partners have not been successful to date, however there was broad consensus that if legislative exemptions are introduced, the industry has the tools and processes to adapt without major disruption.
• First-party audience strategies were discussed in the context of consent and compliance. There was general agreement that audience curation for direct sales is permissible, though technically still subject to consent under current legislation. Publishers would like further clarity on the use of third-party wrappers and verification layers in such scenarios. AOP confirmed ongoing dialogue to help secure a more stable and practical framework.

HR

• The group discussed HR priorities, including new employee benefits benchmarking, DEI programs, and redundancy support strategies. Several publishers shared progress on internal networks and inclusion initiatives, particularly in supporting neurodiverse employees and refining recruitment practices to ensure accessibility and transparency. There was strong interest in maintaining open communication and job sharing across the network to support employees affected by redundancies.
• AOP's HR Resources Table was highlighted, which was created by this group for peer comparison and to create an industry benchmark. AOP encouraged those who have not yet completed both the benefits and providers tables to do so – link to table here: https://docs.google.com/spreadsheets/d/1BkgU_eTWA8vI8EqYXxL993aee6XUvzBzUcHVhE4hRRw/edit?usp=sharing
• Recent AOP Employer Excellence Award winners, Haymarket, highlighted their recognition as a certified Disability Confident Employer and reported significant improvements in candidate diversity and hiring outcomes as a result of targeted internal programs and networks. Several others discussed their own DEI frameworks, including employee-led groups and updates to onboarding processes.
• Redundancy planning was a key topic, with participants sharing varied experiences, ranging from voluntary schemes to compulsory programs. There was consensus on the importance of handling the process with transparency and empathy, noting a shift in stigma around redundancy. Challenges included large-scale team exits, timing of announcements, and support for employees from print-focused roles. The group considered how to better support displaced talent, including through informal job-sharing networks.
• AI usage across HR and editorial functions was discussed, with one member sharing a disciplinary case involving AI-generated content. AI is increasingly used in editorial summaries, onboarding prompts, recruitment support, and self-service HR platforms. However, there was a shared view that AI must be supervised, particularly in sensitive scenarios such as legal disputes, performance management, and redundancies.
• Workflow platforms are being used by multiple teams to reduce administrative burden and increase consistency in handling employee queries. Platforms like ServiceNow are being configured to allocate tasks, generate analytics, and support digital-native employee expectations. Members of the group emphasised the need to balance efficiency with a human touch to avoid transactional interactions within HR.
• The limitations of AI in legal and disciplinary contexts were highlighted. While tools can assist with meeting transcription and documentation, there was caution against over-reliance due to the risk of misinterpretation, lack of nuance, and oversimplification. There was discussion of using AI in notetaking and interview processes, with mixed feedback on accuracy and usability.
• Recruitment and talent sharing were discussed, including challenges in high-volume application processing. Members expressed a preference for referrals over cold applications and discussed using shared platforms or informal groups (e.g. Slack) to distribute job opportunities within this group.

JAB

• For this JAB meeting, we were joined by Professor Charlie Beckett, founding director of Polis at the LSE and who leads JournalismAI, a global initiative helping news organisations adopt artificial intelligence responsibly, to discuss all things AI.
• AI is becoming a critical focus in journalism, not just as a tool for automation but as a means of enhancing reporting, editorial workflows, and newsroom efficiency. There is a growing shift from fear and resistance toward experimentation and practical integration, particularly among smaller and independent newsrooms that are often more agile. Training initiatives, fellowships, and collaborative projects are helping journalists and technologists develop useful tools, such as AI agents for monitoring news sources or tackling specific challenges like churn reduction. There’s increasing emphasis on using AI to return time to human journalists for original reporting, with the most successful outcomes coming from teams that combine AI tools with deep editorial knowledge.
• News brands are recognising the need to adapt formats to audience preferences - especially younger audiences who favour digestible, authentic, and personality-driven content over traditional longform articles. Informal, podcast-style tones and trusted personalities are proving more effective at engaging users than institutional voices. As platforms evolve, publishers must explore new distribution models that integrate with users’ daily routines.
• There is concern that generalist publishers may struggle more with AI disruption, while specialist and B2B publishers, if properly equipped, stand to benefit significantly. However, many B2B teams risk underestimating how AI can support them. More advanced, task-oriented AI agents are demonstrating capabilities in research and content development that may require journalists to shift roles, from creators to editors of AI output. Successful adoption depends not on scale, but on mindset, workflow design, and willingness to adapt editorial practices.
• AI is currently being used to support summary creation, topic research, and content repackaging, but creative storytelling and high-quality longform journalism are still seen as human strengths. Nonetheless, journalism is acknowledged to be formulaic in many areas, and journalists may overrate their creative uniqueness. There's a call to rethink the journalist’s role, not as a creative genius, but as someone who curates, verifies, and enhances the output of intelligent systems.
• Trust and authenticity remain essential in a fragmented information landscape. Audiences, especially younger ones, want to feel that news reflects their world and values. They’re more likely to trust outsiders, such as influencers, than traditional publishers, partly due to disillusionment with political systems and perceived institutional alignment. Successful news brands are those that feel owned by the audience. Informality, personality, and consistency are valued, with platforms like podcasts excelling by blending warmth and content.
• While AI summaries may not significantly boost traffic, they have been shown to improve engagement and readthrough. The key is not to attract the most people, but to build loyalty among those who do engage. Experiments show that honest, helpful content increases user satisfaction, reinforcing the value of transparency and tone. AI’s best use here is in augmenting the human capacity to connect, not replace it.
• Looking ahead, news verification and trust will become premium features. Publishers that can demonstrate accountability, particularly at the local level, will thrive. Larger institutions face credibility challenges, while niche or community-oriented publishers may hold a trust advantage. There’s a risk, however, that younger audiences become more apathetic toward truth, not due to disinterest, but because they feel disconnected from traditional sources of power and representation. Publishers need to rebuild connection through relevance and shared identity.

Product Development

• The group discussed the growing impact of AI features such as Google’s AI Overviews and AI Mode on publisher traffic and visibility. Several publishers reported significant declines in traditional search traffic, particularly to high-performing content like reviews, with AI-generated summaries often replacing or burying original links. While some have seen minor gains through Discover, these have not compensated for the overall loss in traffic or revenue. There was widespread concern about the lack of clarity from Google on how to optimise for AI surfaces, leaving publishers uncertain about where to focus their efforts. Some have begun monitoring visibility across top-performing queries, with early data showing that less than a third appear consistently in AI Overviews. Others are experimenting with open source indexing projects or considering separate strategies for AI discoverability, though the long-term payoff remains unclear.
• Publishers are shifting strategic focus from top-funnel traffic growth to user retention, engagement, and conversion. As organic reach continues to drop, teams are concentrating on making their platforms “stickier” through improved UX, interactive features, and encouraging users to sign in. Some have moved towards product-led growth models, using login features and content personalisation to boost known user rates. Others are trialling new tools that add value for users but cannot be easily scraped by AI systems. There is a broad recognition that traditional SEO alone is no longer enough, and a more holistic approach, combining content quality, product design, and audience development, is now required to future-proof against further changes.
• Despite SEO still playing a critical role, publishers are finding it increasingly difficult to defend SEO investment internally due to falling returns. Some teams have maintained a strong technical foundation, focusing on page performance, core web vitals, and legacy CMS upgrades, but are facing challenges getting buy-in from development or leadership teams. Google's opaque algorithms and evolving features such as “preferred sources” are contributing to frustration, with many noting that even optimised content no longer ranks as reliably. The ongoing spam update and related volatility have only added to the uncertainty, making it harder for teams to plan long-term. As a result, SEO is becoming more reactive, with fewer resources being allocated to experimental work or strategic growth.
• AI tool usage across editorial and product teams is increasing, though most publishers remain in the early stages of integration. Use cases include generating headlines, creating summaries with brand-aligned tone, drafting product briefs, and assisting with ticketing or workflow management. Some have built internal tools or experimented with platforms like Workato and GPT-based assistants to support editorial tasks. While these tools are seen as productivity enhancers rather than replacements, there are concerns around the accuracy and reliability of AI-generated outputs, particularly for summaries and recommendations. Verification and prompt training have become essential to avoid publishing errors or misleading content. Most AI efforts remain in testing or pilot phases, with full-scale production integration still some way off.
• A number of publishers are investing in more advanced infrastructure to support AI integration, including the development of vector-based content databases. These systems aim to allow editorial teams to quickly search, retrieve, and reuse legacy content using semantic or image-based queries. The goal is to improve efficiency and content discoverability at scale, but return on investment is still being evaluated. Some publishers are also experimenting with summarisation tools that adapt tone based on brand prompts or conducting competitor analysis using AI to identify coverage gaps. Despite promising use cases, securing internal funding and leadership support continues to be a major hurdle, with several teams relying on informal experimentation while broader business cases are developed.
• There was a shared sense that industry-wide guidance on optimising content for AI experiences is lacking, and publishers are largely operating in the dark. Several participants expressed concern about losing brand attribution in AI summaries – especially where bylines or publication names are omitted. Smaller or niche brands may be particularly vulnerable if visibility and recognition erode further. Meanwhile, internal teams are having to adapt quickly, often with limited resources and cross-functional misalignment across global markets. Education around AI prompt design, tone control, and editorial usage is beginning to roll out within some organisations, but there is still a long way to go. Many see current efforts as necessary groundwork for a future where AI-led search and content delivery become the norm.

Research

• We are looking for a new co-chair for the group going forwards – if you are interested in finding out more, please get in touch.
• The group received an early overview of the AOP’s upcoming Generative AI Impact Study, which aims to provide publishers with actionable insights into how content will perform in an AI-driven media landscape. The study comprises three key components: identifying which content types are likely to thrive in an AI context, quantifying the reduced value of AI-generated content when disconnected from original human content, and benchmarking the performance of different content types using consumer behaviour insights. Results are expected in late October and will provide evidence to support arguments around remuneration and copyright protection.
• AOP presented a new research proposal, exploring the evolving influence of content creators and influencers in digital media, and the implications for publishers. The group acknowledged that influencer marketing is a growing revenue stream with increasing advertiser demand and trust. Participants discussed the challenges of competing with influencers and questioned whether publishers could integrate influencer-like strategies under their brand umbrella to retain credibility and value while providing marketers with scalable social reach.
• Challenges remain in measuring the effectiveness of influencer-led campaigns. Members of the group highlighted the lack of standardised metrics across platforms and the difficulty in attributing incremental value from influencer activity. While many report on high-level outcomes such as views and recall, most do not currently provide a detailed breakdown of impact per channel or influencer. The group discussed the potential of new benchmarking tools such as CreatorIQ and suggested a future research brief could compare publisher-led branded content to influencer-led executions using matched campaign data.
• Some members of the group expressed interest in exploring how influencer marketing could be reframed to demonstrate accountability and alignment with traditional journalistic values. There was agreement that publishers can offer unique value in this space by combining trusted editorial environments with influencer voice, thereby preserving integrity while meeting advertiser expectations for reach and authenticity.
• Alongside the influencer discussion, the group revisited current research and measurement tools, particularly around Brand Metrics. Experiences varied, with some publishers finding the tool cost-effective for display campaign measurement, while others flagged concerns about small sample sizes, pop-up fatigue, and mixed client feedback, especially when evaluating new product launches. Alternatives are being explored, and members were encouraged to consult AOP’s Digital Directory (link here) for supplier recommendations.
• The topic of CMPs also came up, with one member currently reviewing replacements for OneTrust. Others shared experiences using vendors such as SourcePoint, which was noted for its robust approach to privacy and compliance. AOP offered to facilitate introductions between members exploring similar transitions and encouraged affected teams to join the AOP Ad Ops Steering Group for further peer support.