Professor Hannah Fry’s pigeons probably weren’t on everyone’s bingo card for a marketing and PR conference, but they turned out to be a surprisingly good place to start.
On October 1, marketing, PR, communications and insights professionals gathered for Meltwater Summit on Tour London, the UK leg of our popular annual event series. Across keynotes, panels and breakout sessions, speakers explored how to build brands, protect reputations and make better decisions as AI becomes part of everyday life.
There was plenty of new technology to get excited about, along with some useful reminders about the skills that make that technology valuable: asking good questions, understanding people and turning information into action.
Here are seven takeaways to bring back to your team.
Contents
Human judgment becomes more valuable as AI gets better
AI recommendations are becoming part of the buying journey
Brand visibility needs accuracy and a consistent story
Creator partnerships work through cultural relevance
Start with the decision your stakeholders need to make
Make reliable insight easy to access
Turn faster insight into more time for useful work
Keep asking why and put one idea into practice
Human judgment becomes more valuable as AI gets better
Hannah Fry’s keynote took us through pigeons identifying cancer in medical images, elephants communicating with one another and machines interpreting human instructions in spectacularly unhelpful ways. It was a lively reminder that performing a task well and understanding its wider purpose are different abilities.
The pigeons offered a particularly memorable lesson, reminding us that being good at spotting patterns doesn’t automatically give you all the expertise needed to do a job. The same distinction helps us think more clearly about AI.
For marketing and PR professionals, that means looking beyond how polished an AI answer sounds and investigating the underlying quality of the generated content. Does it use the right evidence? Does it understand the context? Would acting on it produce the result you actually want?
That theme resurfaced in the live Data Malarkey podcast, hosted by Sam Knowles with Canon EMEA’s Gareth Crew and Golin Ketchum’s John Brasington. Brasington described finding AI especially useful for work he already understood, where he could judge the quality of its output. Venturing into unfamiliar territory made those checks much harder.
It’s a useful principle for adoption: establish who can validate the work, give them time to do it and make their responsibility clear. A convincing report still needs someone who can spot what’s missing.
AI recommendations are becoming part of the buying journey
Gener8 founder Sam Jones brought the customer perspective into sharp focus, beginning with a personal story about how he used AI to plan a family holiday and choose a stroller, then buying the first stroller it recommended.
His presentation drew on an analysis of more than 100 million AI conversations from over 200,000 people across the US and UK. The research explored how people use AI and what they do after receiving an answer.
One finding gave brands a clear reason to pay attention: in the situations Gener8 compared, people were 3.6 times as likely to visit a brand’s website when that brand appeared in a commercially relevant AI response, compared with a response that featured a competitor and omitted the brand.
Jones also shared survey results in which 46% of UK adult generative AI users said they would be comfortable letting AI research, select and buy a product costing under £50 (such as an electric toothbrush, for example) without a final review.
These findings give AI visibility a clear place on the PR comms agenda. Understanding what people ask before buying can reveal which product benefits, use cases and sources deserve more attention.
Identify the questions people ask when comparing your category, choosing between products or solving a problem your brand can help with, then investigate which brands appear in the answers, what the recommendations say and which sources support them.
Brand visibility needs accuracy and a consistent story
The generative AI panel, hosted by Meltwater’s Jenny Force, brought together Laura Thomas of Unilever, Alicia Edmonds-Smith of Snap Inc. and Chloe Singleton of eight&four. Their discussion put an important detail at the center of AI visibility: appearing in an answer is only useful if the answer represents your brand in a relevant, accurate way.
Thomas discussed the challenge of managing brands across markets, where names and product associations can vary. Her approach focused on the product attributes the brand needs to be known for, and ensuring they come through consistently in the content people encounter.
Edmonds-Smith emphasized a strong narrative and a consistent story across communications, alongside the growing importance of factual accuracy. Information published today can become part of an AI answer later, giving errors another route to travel.
Singleton shared work for Sandals Resorts that began with detailed customer questions and community discussions, then used GenAI Lens to benchmark AI responses and identify content gaps. The lesson for content teams was straightforward: find out what people actually need to know, then make sure your content answers it.
Force also shared Meltwater’s own experience of rewriting key website pages in plain language around questions customers were already asking, and highlighted a 73% increase in AI citations across the pages measured within six weeks of the website launch.
The Data Malarkey discussion extended this beyond customer acquisition. AI may also help employees, investors and other stakeholders research your organization, so reputation teams have good reason to understand those answers too.
A useful first step is a shared audit across PR, content, social and search teams. Review the facts, messages and customer questions that should remain consistent, then look for gaps across owned content, earned coverage and AI responses.
Creator partnerships work through cultural relevance
Deloitte Digital partner Nick Purdon made a compelling case for brands to understand their place in the conversations around them. His presentation focused on joining the narrative, recognizing that communities and creators already shape it.
Creators can give brands a credible way to participate, but that credibility depends on understanding the community and contributing something people value.
Purdon encouraged teams to give greater weight to cultural relevance, emotional meaning and lasting behavior change when evaluating creator partnerships. Follower counts alone can’t tell you whether a partnership will build a meaningful connection.
His Vaseline example illustrated how a familiar product can find fresh relevance through ideas that originate with consumers. Brands can support and amplify those ideas, helping creators build on something their audiences already care about.
He also described the need to serve both the rational and emotional sides of discovery. AI systems need clear product information: what something does, its ingredients or specifications and the evidence behind its claims. People also respond to humor, taste, identity and lived experience, which is where AI is less helpful.
For your next creator brief, define the audience, the community context and the behavior you hope to encourage. Give creators room to contribute their knowledge of that audience. Then test the content and learn from how people respond.
Start with the decision your stakeholders need to make
Anyone who has produced a substantial report and wondered whether anyone read it will have recognized the challenge in Heineken’s session in which Narek Garit, Global Commercial Analytics Manager at Heineken, described how fragmented data, separate teams and different interpretations can make decisions harder. Social intelligence becomes more useful when teams connect it with relevant media, research and internal information.
His recommended starting point was the use case: Understand what a stakeholder needs to achieve, then work backward to the data, analysis and delivery method that will help.
That changes the brief considerably, because a brand manager planning a campaign and a communications leader assessing an emerging issue may need different answers from the same underlying data. A standard report can leave both of them searching for the most useful page.
Garit shared an experiment applying this approach to media reputation analysis, creating a process that could deliver insights more frequently than a retrospective annual report. He also stressed starting small, proving the value and designing with future scale in mind.
The difficult part, he explained, was agreeing on the use case with stakeholders, as the technology comes into focus much more easily once people agree on the problem.
The connection between insight and action also ran through the agenda’s Mars Pet Nutrition and Weber Shandwick session on culture, conversation and growth. In her closing remarks, Alexandra Bjertnæs highlighted Mars’s use of local market insights to guide decisions and activity.
Before commissioning your next dashboard, ask the intended users what decision it should help them make. Agree on that first, then build the reporting around it.
Make reliable insight easy to access
Meltwater’s Antony Cousins introduced a problem he called “shadow insights”: people using general AI assistants to answer business questions without the data or methodological checks those questions require.
The appeal is easy to understand; You ask a question, get a confident answer and move on, whereas an insights team might first ask about the audience, geography, time period and evidence needed to answer it properly. While those clarifications make the work more reliable, they can also feel like friction to someone in a hurry.
Cousins described a case where an apparently useful answer about consumers’ attitudes toward electric vehicles relied on a single, outdated blog post. A busy stakeholder could easily mistake that response for a much broader body of research.
His session explored how Meltwater’s MCP connector can bring relevant data and existing searches into AI workflows. He also discussed using reusable skills to capture expert methods, so self-service analysis can draw on the way specialists approach a question.
This opens up a valuable role for insights professionals: designing the methods, definitions and checks that colleagues use, alongside answering the harder questions themselves.
Start with a question your team receives frequently, then define the appropriate sources, how to analyze them and the limits of the answer. Making that process accessible can help colleagues get useful results while preserving the expertise behind them.
Turn faster insight into more time for useful work
Meltwater CEO John Box and Chief Product Officer Chris Hackney opened the day with a vision of intelligence that connects monitoring, discovery and action.
GenAI Lens showed how teams can investigate the way AI represents a brand and examine the sources behind those answers, while Trends Center demonstrated how teams can explore emerging conversations. An MCP demonstration connected an AI assistant with Meltwater to build searches, analyze results and develop content suggestions within a conversational workflow.
Hackney also outlined plans for richer analysis of visual content and future trend playbooks, pointing toward more context and clearer routes from discovery to action.
The opportunity for teams is to spend less time assembling the ingredients of their work. Gathering coverage, structuring searches and preparing recurring reports can absorb hours before the strategic conversation even starts.
The keynote included an example from Havas Red, where custom Mira prompts helped accelerate recurring reporting and monitoring, creating more room for strategic and creative work for clients.
Deciding how to use that time is part of the job. Put it toward understanding a new audience, developing a stronger campaign idea, examining an emerging risk or helping leadership make a difficult decision. Faster production becomes valuable when it improves what your team can contribute.
Keep asking why and put one idea into practice
In her closing remarks, Meltwater Chief Strategy Officer Alexandra Bjertnæs emphasized connected intelligence, action and the value of curious, adaptable professionals. She encouraged people to keep asking why, going beyond the first answer to understand what sits underneath it.
That was a fitting thread to bring the day together. Better questions help teams choose relevant data, better context helps people judge AI output, and better understanding helps brands contribute to communities and respond to changing customer behavior.
There are plenty of lessons we can take from Summit, but one well-chosen experiment is a good place to begin. Pick a business question, involve the people who need the answer, agree on what useful evidence looks like and decide how you’ll evaluate the result.
And if Hannah Fry’s pigeons make an appearance in the team debrief, we suspect they’ll help people remember the point.
FAQ Meltwater Summit on Tour: London
What was Meltwater Summit on Tour: London about?
Meltwater Summit on Tour London explored how marketing, PR, communications and insights teams can use AI, connected intelligence and cultural understanding to make better decisions and build stronger brands.
How is AI changing brand visibility?
AI is becoming part of how people research products, compare brands and make purchasing decisions. Teams need to understand how their brand appears in AI-generated answers, whether the information is accurate and which sources influence those responses.
Why does human judgment still matter when using AI?
AI can identify patterns and produce useful recommendations, but it does not automatically understand context, business goals or potential consequences. Human expertise is needed to validate outputs, interpret information and make responsible decisions.
What is connected intelligence?
Connected intelligence brings together relevant media, social, research and business data so teams can answer specific questions and act on useful insight. The process should begin with the decision a stakeholder needs to make, rather than with the technology or data available.
How can marketing and PR teams apply the lessons from the event?
Teams can begin with one important business question, identify the evidence needed to answer it and agree how success will be measured. They should also review their AI visibility, clarify key brand messages, assess creator partnerships through cultural relevance and make reliable insight easier for colleagues to access.
