The race to appear in artificial intelligence search results will not be won by brands that simply publish more content, according to platform executives who say large language models are increasingly rewarding useful community discussion, identifiable expertise and accurate product information. Executives from Reddit, LinkedIn and Carsales on an IAB Australia panel said brands need to improve the quality of the material AI systems can find across official websites, professional networks, marketplaces and consumer communities. Their advice points to a broader shift in marketing, with AI discoverability emerging alongside brand and performance as a discipline requiring its own strategy, investment and measurement. The platforms approached the issue from different directions. Reddit emphasised authentic conversation and peer validation. LinkedIn argued that professional authority is increasingly attached to individuals rather than corporate pages. Carsales focused on comprehensive, machine-readable product information that helps AI systems make reliable comparisons. But there was also a consistency to their advice; the practical challenge is to build visibility without flooding those environments with synthetic material designed to manipulate models. Reddit exec's five step guide Reddit senior agency development lead, Tom Tilney said AI systems value the platform because users share detailed and often unfiltered experiences that can help models form a consensus. According to Tilney's research conducted with AI visibility company, Profound,  five reasons the platform’s content was being surfaced across large language models were identified. “Broadly, it’s because on Reddit you have real humans coming to the platform and sharing unfiltered experiences at scale,” Tilney said. The first factor was the conversational nature of Reddit content and the role it played in helping models assemble an apparent consensus. “When these models were looking to form a consensus, it was pairing the conversational content on Reddit alongside non-conversational destinations such as an official brand website, a Carsales or even a Wikipedia,” Tilney said. The research also found that question-and-answer content was frequently cited. Tilney said three to five specific subreddits would often appear repeatedly in response to questions on a particular topic. Purchase-oriented queries were especially important. Discussions built around questions such as “What car should I buy?” were being treated by models as authoritative, peer-driven reviews, he said. “They were often being surfaced above official brand websites for purchase intent-based queries,” Tilney said. Popularity alone did not appear to determine which posts were selected. Although Reddit users can upvote content, Tilney said the models were more interested in whether a contribution was useful. “It's more so about helpfulness,” he said. “Helpfulness coming across in a clear, human and conversational direct tone that had a balance of both positive, constructive and negative sentiment.” Tilney's next lesson for brands was to avoid treating AI visibility as a short-term publishing tactic. Some Reddit conversations cited by models had remained influential for as long as a year. “This visibility effort is not something where we should be chasing virality and trying to do it quickly,” he said. “It’s all about building a long and enduring strategy.” This can replace the opening portion of the existing Reddit section before the discussion of listening to communities and avoiding inauthentic participation. His final recommendation was direct. “Go to Reddit.com, listen to the communities, learn from what they’re saying, and then find a way to launch into that specific community in a way that is going to add value and utility,” Tilney said. LinkedIn: Invest in people, not only pages LinkedIn director of marketing solutions, Louise Wilson said professional expertise was becoming an important signal for AI systems, particularly during business-to-business research. “Expertise is becoming the currency that AI rewards,” Wilson told attendees. “LLMs are looking for trust signals, and those trust signals sit in platforms where there are verified identities, real people with real reputations, contributing to real conversation that is surrounded by quality conversation.” Wilson said brands should stop treating LinkedIn solely as a campaign distribution channel and invest more heavily in executives, employees, customers and subject-matter experts. “What we’re seeing from all LinkedIn citations is that 75 per cent of those are coming from individuals rather than LinkedIn company pages,” she said. “If you’re not investing in your thought leaders, in your customers, in your employees who are speaking about your brand, you’re missing a big opportunity there.” Wilson recommended structuring content in formats that are easy for both people and AI systems to understand, including articles, longer-form posts, bullet points and how-to guides. She cautioned that structuring content for AI did not mean asking AI to manufacture generic material. LinkedIn is also attempting to reduce the distribution of content that appears automated or lacks genuine professional perspective, Wilson said.  “We are reducing the distribution of content that appears to be AI created or lacks perspective and expertise,” she said. “We are really using measures to detect and limit the number of automated comments.” Brands should also encourage meaningful discussion around expert content and consider amplifying strong employee or customer posts through thought-leadership advertising. Wilson pointed to Sydney legal software company, Smokeball, which uses customer testimonials and a 10-week video series by its chief executive. "They have their customers talking about their brand video, and they promote that and problems that Smokeball solve." “What they do a great job of is not selling, but teaching and being an expert,” she said. Carsales: Give AI reliable product facts Carsales GM of data, Matt Doherty said AI visibility depended on brands making product information trusted, comprehensive and machine-readable. “The way thatwe’ve strategised and thought about this at Carsales is really to dig into what are the drivers for us to support that space,” Doherty said. “Three key pillars are really around having a trusted marketplace with industry-reputable information behind it.” That includes detailed vehicle specifications, real-time listings, prices, availability, buyer guides, comparison reviews and research articles. “That’s really all of the high-confidence signal-type information which LLMs really look to use and understand and speak to the authority in that specific category or vertical,” he said. Carsales research found AI use was concentrated in the research and consideration stages, with consumers moving between the marketplace and companion LLMs while learning about vehicles and categories. Doherty said brands should clearly describe how products differ at the brand, model and variant level, then test how those differences appear in AI answers. “Really look at where you’re going to be able to be competitive or put yourselves above your category in terms of how people understand what the value and benefit is,” he said. “Actually run the exercise, go out there and test it, see where you’re showing up.” When combined, the panel's message was that AI discoverability requires more than optimised copy. Brands need reliable product data, credible human expertise and useful participation in the communities where buyers test claims against real experience. The bottom line consensus: Brands that provide the clearest and most trustworthy evidence will have the best chance of shaping what AI systems tell their customers.