For the better part of two decades, HubSpot had built its business on a single, reliable premise: create content that people find helpful, and they will find you, trust you and buy from you. The company did not merely advocate this approach. It wrote the book on it – literally – when co-founders Brian Halligan and Dharmesh Shah coined the term “inbound marketing” and published a guide to the methodology in the mid-2000s. The strategy powered HubSpot to a US$2 billion valuation and helped more than 250,000 customers grow using the same principles.

Then, in 2024, the engine stalled.

Kat Warboys, HubSpot’s Senior Marketing Director for the Asia-Pacific region, described the moment the company realised its demand generation model was in trouble. What had begun as a dip in inbound leads turned into a downward trend that could not be dismissed as a reporting anomaly or another Google algorithm update. At the height of HubSpot’s inbound era, marketing-generated leads had accounted for 80% of the sales team’s revenue. When those leads began to decline, the consequences were potentially dire for the business.

“We used to joke that marketing would go on holiday, the engine would keep running, and we’d still hit our goals. And it wasn’t really a joke. It was kind of true,” said Ms Warboys, who recently delivered a keynote presentation at the UNSW Marketing Analytics Symposium Sydney (MASS) 2026. “So, you can imagine, it didn’t take long then for this downward trend to really trickle down to the sales team and to revenue. So we did what many of you, I’m sure, would do – we issued a code red, which is really internal language for everything is on fire, all hands on deck. We need to figure this out now.”

What HubSpot figured out, Ms Warboys told the audience, was that its code red was not a temporary performance dip. It was the beginning of AI’s disruption of the marketing playbook itself – from search to demand generation to the way buyers make purchasing decisions. The company’s response over the past two years has involved a wholesale rethinking of how marketing teams should operate in an era when artificial intelligence is reshaping the rules of engagement.

Why search engine optimisation stopped delivering results

The first and most consequential shift, Ms Warboys explained, was in search. HubSpot had built its inbound model on content designed for search engine optimisation (SEO). The logic was straightforward: rank on Google, win the click, convert the visitor. But AI overviews – the summaries generated by search engines and AI tools that now appear above traditional search results – had begun intercepting that traffic. Ms Warboys cited data showing that six in 10 searches now never result in a click through to a website, because users receive answers directly from AI-generated summaries.

“Today’s biggest discovery engine, the biggest emerging discovery engine that we’re seeing, doesn’t actually send people to websites anymore,” Ms Warboys said. She pointed to HubSpot’s internal forecast that by 2028, the majority of website traffic will come from ChatGPT rather than Google – a projection she described as a fundamental undoing of the assumptions that had shaped marketing strategy for two decades.

Photo gallery: The UNSW Marketing Analytics Symposium Sydney (MASS) 2026

When inbound leads declined, HubSpot turned to paid advertising to fill the gap – the next available lever. But costs were climbing across platforms, with TikTok CPMs rising nearly 16% year-on-year by 2025, and performance was not keeping pace with the increased spend. The company found itself paying more to stay in the same place. Ms Warboys noted that the rising cost of paid media was not unique to HubSpot. It was a market-wide trend, compounded by the fact that the channels marketers had relied on for predictable returns were becoming less effective even as they grew more expensive.

How people-first content replaced the brand-led playbook

Behind the decline in search traffic, a bigger change was underway in how buyers consume and respond to content. In 2022, at the height of HubSpot’s inbound era, educational content – blog posts and resources – drove 61% of the company’s leads. By 2025, that figure had dropped to 28%. In its place, people-first content – material fronted by individuals rather than brand channels – was generating 55% of inbound leads.

Ms Warboys said the trend reflected a shift in trust. Buyers no longer place the same level of confidence in brand-produced content as they once did. Instead, they trusted individuals: peers, subject matter experts and people they considered authorities in a particular domain. She described how this played out in her own team’s campaigns.

Learn more: Marketing in the AI agentic enterprise: Lessons from Salesforce

“Pretty much every time we launch a campaign, now they’re coming to me halfway through, saying we want to shift the spend from our brand channel, you know, HubSpot LinkedIn, HubSpot Meta channels, to the people-first content we have running,” said Ms Warboys, who takes a lead role in delivering this content for HubSpot. “This is me promoting our webinar on LinkedIn. We have lots of people at HubSpot that do this throughout the campaign, and we boost that ad through LinkedIn, and it is outperforming the brand channel ads.”

The effect of these changes on the customer journey was significant. The predictable, linear pathways that marketers had relied upon no longer existed. Buyers were researching AI tools, building trust through peer networks and community platforms such as Reddit, and arriving at brands already informed, educated, and, in many cases, more qualified than leads generated through traditional channels. HubSpot found that visitors arriving from these newer sources were converting at three times the rate of those from traditional channels, converting faster and generating more revenue per customer.

Ms Warboys acknowledged a paradox in the data: while overall lead volume from established channels was declining, the quality of demand was rising. This left the company in a position where the old playbook was losing relevance, but the emerging patterns offered grounds for cautious optimism. As Ms Warboys framed it, awareness that had once been centralised – owned or purchased through brand channels, Google or Meta – was now scattered across platforms, creators and communities. Traffic, once the widest part of the marketing funnel, was shrinking in favour of AI summaries. Conversions, once the narrowest part, were becoming faster and AI-assisted. The challenge for marketing teams was that much of this activity was invisible to traditional analytics.

The loop marketing model: a four-stage AI framework for growth

In response, HubSpot developed what it calls the “loop marketing model” – a four-stage framework designed to combine human intuition, judgment and expertise with the efficiency and analytical capacity of AI. The four stages are: express, tailor, amplify and evolve. Each addresses a different part of the marketing challenge the company identified during its code red period.

The first stage, express, focuses on brand voice. Ms Warboys argued that one of the most common questions marketers ask about AI is how to use it to create content that remains distinctive to their brand and does not simply replicate the generic output that AI tools produce by default. The answer, she said, lies in feeding AI the right context.

“AI scales what you feed it, and if you don’t define your voice, the algorithm will,” Ms Warboys said. She noted that most organisations have brand guidelines, but many have not been reviewed in years and do not reflect current market conditions. HubSpot’s approach involves feeding AI a continuous stream of real-time inputs – customer surveys, call transcripts, support tickets and leadership insights – to create what Ms Warboys described as a living, breathing brand guideline that updates in real time.

This context is then made available to every content creator at HubSpot, either through an AI-powered “copy chief” that reviews work against brand standards or by embedding the guidelines directly into the editing tools teams use. The AI can also adjust its output depending on the channel: blog content is more informative, knowledge base articles are more instructive, and social media posts are less formal. It can further tailor output by persona, taking into account the likely challenges facing a particular audience segment.

Subscribe to BusinessThink for the latest research, analysis and insights from UNSW Business School

The second stage, tailor, addresses personalisation at scale. Ms Warboys contrasted the company’s previous approach – automated emails with first-name personalisation and behaviour-based triggers – with its current AI-powered method. Today, HubSpot’s email engine examines the individual recipient’s website, industry, role, pain points drawn from CRM data and call transcripts, and even competitor information, to produce emails where the value proposition is directly tied to the recipient’s circumstances.

“This is not about replacing teams, which is another question we get a lot,” Ms Warboys said. “Our HubSpot humans are still steering for strategy. They’re approving the approaches. They have final sign-off. They own the relationships. What AI is doing here is blending in real-time CRM insights and intent data to write one-on-one emails in seconds.” She reported that this approach had increased email conversion rates by 45% and generated approximately 10,000 meetings booked by HubSpot’s sales team each quarter.

To illustrate the approach beyond HubSpot’s own operations, Ms Warboys shared a case study from Hungry Hungry, a HubSpot customer. Liz, a marketer working across three brands with a limited budget, had been sending plain-text emails to existing customers to promote premium features. The emails were generating negligible engagement. After adopting HubSpot’s personalisation framework, Liz combined AI-generated personalised video – featuring an AI avatar of the company’s chief AI officer, augmented by CRM context – with targeted email. The result was a jump from a 3% click-through rate to 29%, a nine-fold increase. Recipients did not merely watch the video; they proceeded to the website and trialled the product.

From SEO to answer engine optimisation: building brand visibility in an AI world

The third stage of the loop model, amplify, deals with meeting customers where they actually are – which is no longer primarily on brand websites. Ms Warboys described how HubSpot’s content strategy had evolved from a blog-centric model to what she called a media network. In January 2023, HubSpot’s chief marketing officer made the decision to invest in people-first content by acquiring The Hustle, a newsletter with a substantial following in the United States. Since then, the company has built out a portfolio that includes podcast networks (featuring shows such as My First Million and Marketing Against the Grain), additional newsletters, including Mindstream, a YouTube channel running 11 shows, and a creator network. The combined media network now generates 33 million monthly engagements, and demand from these channels has grown 85% year-on-year.

On the question of search specifically, Ms Warboys outlined HubSpot’s shift from SEO to what the industry is calling answer engine optimisation (AEO) – the practice of structuring content so that AI tools cite it as a source when generating answers. She noted that the principles share similarities with SEO but differ in emphasis.

Where SEO rewarded broad, high-ranking pages on topics such as “how to do content marketing”, AEO requires content that answers specific sub-questions – such as “how to do content marketing in a law firm” – because AI tools seek precise, authoritative responses rather than general overviews. Structure matters as well: clear headers, data, expert quotes and FAQ formats make it easier for AI systems to scan and extract content. But the most significant shift, Ms Warboys said, involves brand authority. Under the SEO model, authority was measured in backlinks. Under AEO, it is measured in mentions on the platforms AI trusts – Reddit, LinkedIn, review sites and discussion forums.

Learn more: AI-powered marketing: From the classroom to the boardroom

“When we were optimising this for Google, this really meant backlinks,” Ms Warboys said. “Today, this means mentions on the sites that AI trusts. You want to ask yourself: 'Is my brand getting mentions on Reddit, on LinkedIn? Are we predominant on review sites? Can you encourage your team to engage in these forums and discussions? Can you encourage your best brand advocates to put the good word out there for you?'” She added that businesses should also consider approaching creators to review their products and share ratings with their audiences, and that internal subject matter experts – not just external influencers – could be valuable in building authority.

Since adopting this strategy, HubSpot has become the most cited CRM in answer engines, Ms Warboys reported, outperforming its competitors on visibility metrics. The traffic driven through these channels also converted at higher rates and generated more revenue per visitor.

Key takeaways: practical lessons for marketing teams in an AI era

The final stage of the loop model, evolve, reflects what Ms Warboys called a mindset shift in how marketing teams approach measurement and optimisation. The traditional cycle of launching a campaign, waiting for it to conclude and then reporting on results over six or 12 months is, in her view, no longer tenable.

Instead, every channel should feed live learnings back into the system for continuous optimisation – when a landing page converts at a higher rate, the team is notified, and the data set is updated; when an email outperforms, future messages are adjusted automatically; when content succeeds in one format, AI tools can reformat it for other channels in seconds. HubSpot uses the term “remixing” internally: a single piece of content that performs well can be transformed into a video, an email, a landing page or a social post within seconds, ensuring that successful material reaches audiences across every relevant channel.

Ms Warboys identified two capabilities that she considered the most significant unlocks for marketing teams. The first is access to data without delays. In larger organisations, obtaining performance insights has traditionally required logging a request with a data analyst and waiting days or weeks for a response. AI now allows marketers to access, analyse and compare performance data in real time, reducing decision-making cycles from weeks to hours.

The second is the speed of experimentation. Ms Warboys said HubSpot now encourages what it calls an “explorer’s mindset” among its marketing teams – a departure from the previous approach where teams could rely on established formulas to produce predictable outcomes. In a landscape where the rules are changing, the ability to test, learn and iterate quickly is more valuable than executing a static plan.

“When this is stuck in people’s heads, that knowledge – what we tested, what worked, what didn’t – goes out the door the minute they leave,” Ms Warboys said of the importance of feeding learnings back into AI systems rather than allowing them to remain as institutional knowledge held by individual team members. “And that’s really the power of evolve – when you’re not waiting for the end of the campaign, and you are really learning in real time.”

For business professionals and marketing teams navigating the same shifts, Ms Warboys’ account of HubSpot’s experience offers several practical lessons. First, declining organic traffic is not a temporary disruption – it reflects a structural change in how buyers discover and evaluate products, and organisations need to plan accordingly. 

Learn more: How businesses can leverage AI agents (without crossing the line)

Second, brand authority in the age of AI is built through mentions, reviews and community engagement on the platforms AI systems trust, not through backlinks alone. Third, personalisation must move beyond inserting a first name into an email template; AI tools can now draw on CRM data, call transcripts and intent signals to make every communication specific to the recipient. Fourth, people-first content – material fronted by individuals with expertise and credibility – is outperforming brand-channel content, and organisations should invest in building internal and external voices. 

Finally, the organisations that will adapt most successfully are those that treat every campaign as a learning engine, feeding results back into AI systems in real time rather than waiting for post-campaign reviews.

“If I can leave you one takeaway, it is this,” Ms Warboys told the audience. “When AI has context, it can level up your brand in authentic ways. It can treat every customer as if it knows them intimately. It can show up and drive high-converting demand across the channels your customers are actually on, and it can help you learn fast, feeding every action back into the organisation to iterate and optimise.”