19/05/2026
Full recap: Ahrefs Evolve Singapore 2026. 11 speakers.
Every session. Every insight.
1️⃣Tim Soulo — CMO, Ahrefs: How to Make Your Brand Discoverable in the Era of AI
People discover brands through three main channels: Search, Notice and Talk. The problem is not that people are searching less. Clicks on Google are disappearing because search is spreading across more platforms. Google has lost around 5% of its traffic share over 10 months, but it still leads ChatGPT by a significant margin.
♦️Do not just chase AI traffic. The real goal is to get your brand mentioned by AI.
♦️LLMs pull from multiple sources and personalise answers based on who is asking. Every person gets a different response.
♦️To get AI to mention your brand, you need to accumulate brand mentions from as many places as possible: YouTube, listicles, press, forums and third-party sites.
♦️Start with your own properties first. Keep them up to date, well structured and information dense.
♦️A low DR does not mean AI will not cite you.
♦️To track AI search performance properly, look at AI referral traffic, AI bot visits and AI visibility scores. Avoid using real user prompts as benchmarks as they are too long and personalised to be useful. Use self-attribution instead.
♦️Ahrefs has just launched Agent AI, a new tool that handles all of this, from tracking brand mentions to monitoring AI visibility.
2️⃣Dan Petrovic — Managing Director, DEJAN: I Probed Gemini and GPT 1,000,000 Times
After prompting LLMs to name brands randomly over a million times, a clear pattern emerged. LLMs have their own brand network with a fixed set of top-of-mind names: Apple, Samsung, Nike, Google and Toyota. Brands that are associated with these names get pulled into responses too. The higher the frequency of association, the more visible a brand becomes.
♦️You need to know which network your brand sits in. That determines your relevance and citation potential.
♦️Brand networks work both ways. They can help or damage your brand depending on the associations.
♦️Most of the data LLMs draw from comes from the US and Japan. Chinese brands are largely absent from Western LLMs.
♦️Content that performs well with AI uses specific numbers, verbs over adjectives, precise geographic references and detailed customer segments. For example, changing "Trusted by professionals" to "Used by surgeons at Mayo Clinic and Johns Hopkins" makes a significant difference in how AI picks it up.
3️⃣Robert Lai — CEO, Kaliber Growth: Happens Before Search Intent
Real buying behaviour does not follow a neat funnel. Here is an example using an LED face mask purchase:
→ 120 days before purchase: wife mentions LED masks, leads to first marketplace search → 10 days before purchase: walks past one in a mall, searches brand specs and reviews → 7 days before purchase: searches for "scientific value" to confirm the decision → Purchase
♦️Keywords are the receipt, not the journey. They show what someone typed, not what shaped the question.
♦️People start with a situation, then form a question, then reach for a keyword. Not the other way around.
♦️Sometimes it goes situation to memory to purchase, with no search involved at all.
♦️Everyone is competing at the obvious query level: Ads, Map pack, Directories, Reviews and SEO. Winning the question before the keyword matters more.
♦️In the AI era, you can influence an answer without getting the credit. The citation may not appear but your explanation still gets used.
♦️Brands need to map the buyer before the keyword and be clear on the difference between Demand Forming (branded search, direct traffic, share of search) and Demand Converting (leads, purchases).
4️⃣James Norquay — Founder, Prosperity Media: Digital PR and GEO Growth Strategies to Dominate APAC in 2026
People have been saying SEO is dead for 29 years. It never was. What has changed is that Digital PR is now a key driver of LLM visibility, not just traditional media coverage.
♦️Being featured in news and press directly improves your chances of being mentioned in ChatGPT.
♦️Monitor trending topics through Google Trends, Glimpse and Exploding Topics, then adapt them into your own content.
♦️Data journalism is underused. Find interesting datasets and build stories around them. Tools like Pudding.cool and Visual Capitalist are good references.
♦️Award pages and Reddit are two places LLMs draw from heavily.
♦️List your business on Foursquare. It is a direct partner with OpenAI and feeds location data into ChatGPT Search.
♦️If some AI platforms cannot read your site, having an LLM.txt file helps them understand it.
♦️Ahrefs Brand Radar and Ahrefs MCP, which works alongside Claude, are useful for tracking mentions consistently.
5️⃣Viola Eva — Founder, Flow Agency The Future of B2B Search Marketing
Visibility does not equal accuracy. A brand can appear in AI responses with incorrect information. Both problems need fixing, not just one.
♦️You need a layered brand narrative that defines what each product stands for.
♦️In B2B buying, every stage from research to final decision uses different channels, and LLMs are increasingly one of them.
♦️The five steps are: Articulate your brand narrative clearly, Audit and Refresh starting with your own website since it is the fastest to control, Update and Align by checking that socials, review sites, media appearances and partner integrations all say the same thing, Scale into new listicles and earned media, and Monitor continuously.
♦️Accuracy of what AI says about your brand matters just as much as being seen in the first place.
6️⃣Loki Yan — Co-Founder, First Optimise Rebuilding for the Machine: A Day 0 Framework
Three things nobody is talking about. Every AI engine sees your website differently. Everyone focuses on Schema but forgets that HTML is the actual foundation. And everyone optimises individual pages while forgetting that the entire site is one entity.
♦️Good traditional SEO is already AI SEO. A developer writing good code is already optimising for AI without realising it.
♦️AI recommends entities, not just pages. Knowledge graphs still matter enormously.
♦️AI avoids hallucination by defaulting to official brand pages. Your About Us page needs to exist from Day 1, not Day 100.
♦️Watch out for cross-language entity disconnect, especially in APAC where brand names often differ across languages and AI cannot connect them.
♦️Chinese AI systems such as DeepSeek, Doubao and Qwen operate in a separate bubble entirely, drawing from WeChat, Taobao and Zhihu with almost no reference to English-language content.
♦️Four decisions to make from Day 0: use SSG or Streaming SSR only, no pure client-side rendering; Semantic HTML first with Schema as annotation only; build your About Us page on Day 1; and maintain one responsive domain with cross-language entity consistency.
7️⃣Ted Chong — Digital Marketing Director, Ice Cube Marketing: How to Create Mind-Reading PPC Campaigns
The Curse of Knowledge is damaging conversions. We understand our own products so well that we forget customers do not. Good insights matter more than good copy.
♦️Read Google Reviews for your own brand and your competitors. Rank the objections from most to least mentioned, then use them to write ad copy that addresses those fears before people even ask. If reviews show that people are afraid of hard selling, write an ad that says "No Packages, No Hidden Fees."
♦️The best sources of insight are CRM data, comments, forums, competitor reviews and market trend data. AI can help you analyse all of these much faster.
♦️In AI-powered Google Ads, creative assets now matter more than keywords and bidding.
♦️Facebook and Instagram Ads remain the strongest platform for high-ticket sales and lead generation due to AI-driven targeting. TikTok works well for low-ticket e-commerce and virality but converts poorly for high-ticket products.
8️⃣Alan Chan — Head of GTM, Manus AI: How Manus AI Uses Manus for GTM
Most people think AI agents are just smarter chatbots. In practice, they can handle research, writing, coding, data analysis and full workflow automation.
♦️For SEO, Manus connects with Ahrefs to run site audits and backlink analysis.
♦️For influencer outreach, it handles bulk messaging automatically with task chasers and post streak trackers.
♦️For performance marketing, it runs an Ads Evaluation Bot and an Ads Performance Tracker through Slack.
♦️It can analyse viral videos and extract scripts and storyboards in multiple languages.
♦️The core principles: context is king, automate at scale, and if you can explain it, you can build it.
9️⃣Ilman Akbar — Founder and CEO, DailySEO ID: How to Talk So the C-Suite Will Listen
Before investing in SEO, ask three questions first. Are customers searching? Are competitors visible? Is the business model profitable? If there is no demand, SEO is not the answer.
♦️Connect SEO to business objectives by working top down: Business Objectives to Marketing Objectives to SEO Objectives and KPIs to SEO Key Levers to Deliverables.
♦️Use the Business Model Canvas to understand the business before anything else.
♦️What a C-Suite wants to hear in five minutes: what you need and why it ties to a business objective, where you currently stand against KPI, and what you are going to do next.
♦️Offensive SEO for market challengers focuses on brand marketing, organic traffic and AI visibility. Defensive SEO for market leaders focuses on performance marketing and reputation management.
🔟Panel Discussion — Jin Ma, Shane Yuen, Ariff Kamal: The Hard Calls We Are Making in AI and Search Right Now
♦️SEO foundations must come first. Content needs to work for both people and AI.
♦️Sales can drop even for well-known brands when AI recommends a competitor instead. People trust what AI says more than existing reputation.
♦️Measuring LLM impact is difficult for enterprise clients. Focus on fixing what can be measured and impacted first.
♦️LLM optimisation requires real testing and iteration. Doing things "for AI" without a feedback loop does not work.
♦️If certain AI platforms cannot read your website, an LLM.txt file provides a fallback.
♦️Measuring GEO performance properly means going beyond citation and visibility to include freshness, accuracy and brand mention count.
1️⃣1️⃣Charlotte Ang — Managing Partner, Traffic Bees: How We Built a $1M+ E-Commerce Brand from Scratch in the AI Era
AI means anyone can produce content in seconds. The only way to stand out now is to build a real brand. Branding is no longer optional. The framework they used is called V.E.T.S.
♦️Video is the new top of funnel. Film once and distribute everywhere. One video can reach 121K views on TikTok, 87K on Instagram, 20.5K on YouTube and 6K on Facebook. You need three formats: evergreen for content that stays relevant long term, authoritative to build thought leadership with a recognisable face, and trendy which takes the least effort and tends to go the most viral.
♦️Events are the moat AI cannot replicate. They create real emotional connection with your audience. Every major event they ran produced a brand search spike, a traffic surge and backlinks from major publications.
♦️Trust is built through Google Reviews since LLMs pick up real reviews and use them to support brand visibility, through marketplace presence since in Southeast Asia most product searches start on a marketplace, and through consistent livestreaming to build a loyal community.
♦️Search is no longer a top-of-funnel channel. It is now a conversion layer. V, E and T all feed back into and strengthen SEO performance.
1️⃣2️⃣Sasha Gusain — Canva Scaling Canva to Unicorn Status Through Product-Led SEO
Canva has grown 120x in 10 years with over 100 million monthly visits, 71,000 pages across multiple markets and support for over 30 languages. Organic growth compounds when the path to value is obvious.
♦️Good SEO research goes well beyond keywords. You need to look at community questions, creator signals, visual patterns, AI discovery opportunities, information gaps and emerging intent clusters.
♦️Start from a keyword, generate a large number of use cases, validate them, then scale what works.
♦️Localisation needs to sit between two failure modes: too centralised means everything bottlenecks at the centre, too decentralised means quality fragments across markets. The approach that worked was to codify the thinking and make the system the strategy.
♦️Continuous research is essential: what does great local UX look like, which products solve a real intent, what is the business opportunity, and how healthy is the funnel at each stage.
====
Is your brand invisible in AI answers?👉
http://audit.prml.com/
====
