Last week we attended the Marketing Association's Digital Day Out, with sessions from Google, Orange Labs, Trade Me, Dentsu, and Mastercard, among others. AI was the thread running through almost every talk, but the more interesting conversation wasn't about the technology itself; it was about what it means for how we build brands, trust, and teams. Here are our top takeaways.
1. Search is becoming a conversation
Google's session covered how search is adapting to two fundamental consumer needs: the desire to learn and the drive to get things done. Traditionally that's been served through keywords, but the new model is a back-and-forth conversation with search, powered by AI Overviews and AI Mode.
A few stats stood out: AI Mode searches run about 3x longer than standard searches, and people using AI Mode are making decisions faster. The takeaway for brands is to make sure you're "the right answer" when someone is searching in AI Mode, and critically, AI only works if people trust the source. Ultimately, people are looking for a business they can trust, not just an answer.
2. Brand visibility now depends on being understood by machines, not just people
Orange Labs made a sharp point: if your brand isn't statistically common in the data AI models are trained on, AI is going to struggle to represent you accurately. Their advice for showing up well in an AI-driven world:
Be understood - invest in technical SEO, accessibility, and structured data so AI can actually parse your content (a great reminder: AI models are often taking screenshots of pages rather than reading structured data cleanly so make sure your on-page content is actually relevant).
Be trusted - brand mentions are becoming the new backlinks.
Be discussed - visibility in communities and social conversations matters more than ever.
Their closing line was a good gut-check for any marketing team: this isn't an awareness problem in marketing teams; it's an execution problem. Have the courage to execute imperfectly rather than waiting for the perfect opportunity.
3. AI adoption is a culture shift, not a tech rollout
This came up in a few different sessions, but Trade Me put it most plainly: rolling out AI is a culture change, not a technology change. It needs progress over perfection, and clear, measurable goals, not just tool access.
Their framing was "copilot, not autopilot". Success isn't just about giving people access to AI tools; it's about training and support alongside it. Practical ideas raised included dedicated weekly AI time (e.g. a 30-minute check-in) and the reminder that leaders need to walk the walk, not just mandate adoption from the top.
Dentsu added a sharper edge to this: technology lifts the floor, but teams have to lift the ceiling themselves. Simply knowing what you're doing isn't enough anymore. Pragmatism, not perfection, is what leads to great work.
4. Start with the problem, not with the AI
A really practical, tactical takeaway from Orange Labs was this: don't start with "how can we use AI here?" Start with the actual problem. Their suggested process was simple: Ideate → Gather → Validate → Scale, with the goal of just getting to an 'ugly V1' rather than a perfect solution first time.
Rather than trying to tackle something broad like "improve all our content," pick a low-risk, clearly measurable use case first. One concrete example: generating a weekly piece of content for your business based on trending industry topics; small in scope, but with a clear, testable outcome.
5. Community and trust are the real differentiators
Across several sessions, a common theme kept resurfacing: as AI makes content and search more automated, community and human trust become more valuable, not less. Google’s session pointed to the psychological need for belonging; having a community builds trust in a way that content alone can't. Mastercard's talk on agentic commerce raised similar questions about how AI reshapes the funnel, and where privacy fits into that.
On the more practical side, there was a good example of using AI agents to support (not replace) community building - from a social media coordinator agent that drafts content based on trends for weekly review and approval, to a community manager agent that monitors and helps respond to comments across channels. A Reddit-specific idea was notable too: identifying where your customers actually spend their time, building genuine community there (like AMAs), and using engaged organic posts to inform paid campaigns.
The common thread
If there's one line that ties all of this together, it's something Dentsu's speaker said: AI has a branding problem - it's being pushed on people whether they want it or not. The sessions that resonated most weren't about chasing the technology for its own sake, but about starting with real problems, building genuine trust, and treating AI adoption as a people and culture challenge as much as a technical one.