In category after category, buyers are asking their first questions in AI tools instead of traditional search engines or brand websites. For marketers, that shift changes where discovery happens and which channels actually influence revenue.
Mortgage might be the clearest stress test. It is a long, intrusive decision with high stakes and low trust. Yet that is exactly where Dawn Dawson, CMO at Mortgage Solutions Financial, has started using ChatGPT as both an ad channel and an education layer to overhaul how her team finds and qualifies demand.
Dawson’s early results, shared at HubSpot’s Unbound conference this week, point to a bigger question. If AI tools are becoming the starting point for research, what does it look like to treat them as core media, not just creative helpers?
AI Search Shifts the Top of the Funnel
In the Unbound session, AI: A New Marketing Channel in a Changing Consumer Journey, Lorry Destainville from OpenAI’s Monetization Partnerships team claimed that over half of ChatGPT’s billion weekly users lean on AI platforms to navigate major decisions, according to internal research across 15 countries. Consumer usage is even higher among Gen Z and Gen Alpha.
That behavior changes how and where marketers can show up. ChatGPT currently serves ads to free and lower-tier users by placing an image, short copy and a call to action directly under the AI’s answer. Advertisers can target those ads by location, first‑party audiences or the types of conversations they want to be associated with.
A Mortgage Funnel That Is too Long and Awkward
For Dawson, the trigger to experiment was simple. Organic search traffic had started to slip, and the lead‑gen engines she relied on felt increasingly hungry with diminishing returns. At the same time, the product she sells is one of the most uncomfortable to talk about in public.
A mortgage application forces people to reveal missed payments, credit issues and other sensitive details. Many prospects are not ready for that level of scrutiny when they are still wondering whether they can buy a home at all. Dawson said that the gap between early curiosity and hard qualification produces what every CMO knows well — unproductive calls, frustrated sales teams and wasted media spend.
Dawson wanted to move further up the journey and meet people where they felt safer asking basic questions.
“We are a very intrusive purchase, so I want people to be able to talk through their situation before they ever have to face a salesperson,” Dawson said, describing the early research phase she is now targeting in ChatGPT.
Using ChatGPT as an Education and Trust Layer
Instead of pushing prospects to the company’s website and hoping they convert, Dawson now uses ChatGPT to educate and filter. When someone asks the model about home loans or specific programs, an ad for Mortgage Solutions Financial can appear under the answer with content designed to help people understand their options, not rush into an application.
The strategy is less about catching broad keywords and more about matching the language people actually use when they are uncertain. Mortgage Solutions Financial offers more than 1,100 loan programs, from VA loans for veterans to FHA, USDA and down‑payment assistance products. That variety is only apparent to consumers when the right product shows up for the right question.
“Being part of a natural chat with someone who just wants education lets us focus on trust, not just leads,” Dawson said. “If we can serve them in that upper funnel, they are more ready for the conversation later.”
This approach has also helped with qualification. As users reveal more context in their prompts, Dawson’s team can narrow who sees which offers and feed sales teams conversations that are further along in intent.
Sponsored Agents and the Feedback Loop
Dawson is also testing “sponsored agents,” which extend that model one step further. Announced at the show this week, the new brand‑built agents can sit inside the platform, answer category questions and hand off outcomes like lead creation into tools such as HubSpot. In this format, an ad can lead into an agent trained on company data that guides users toward actions like checking eligibility or booking time with a human.
Early results have surprised her team. Impressions and clicks ramped quickly, but the more important learning has been qualitative.
“We are seeing what customers actually need in their own words, and some of it is different than what we expected,” Dawson said. “The platform makes it fast and easy to tweak based on that feedback.”
Because the OpenAI integration flows into HubSpot — another announcement from this week — her team can see how those conversations translate into leads and opportunities instead of just upper‑funnel activity. Such reporting is critical when she sits down with a CFO or finance leader who wants to understand whether AI advertising belongs in the channel mix.
What CMOs Can Do Now
Most brands are not ready to rebuild their full funnel inside AI platforms. But they don’t need to be, according to the panelists. Dawson’s example suggests a more pragmatic playbook marketers can adapt:
Start with one friction point: Pick a painful moment in your journey — unqualified demos, awkward first calls, stalled deals — and ask how AI‑based conversations might ease it.
Map the real questions, not just keywords: Talk with sales, service and frontline staff to collect the exact phrases customers use when they are confused or hesitant. Those become the conversation types you target with AI ads and agents.
Lead with education before conversion: Design ad creative and agent scripts that help people understand tradeoffs, timelines and readiness. A shorter list of better prepared prospects usually beats a larger list of indifferent ones.
Wire AI into your revenue systems: Connect AI ad pilots to your CRM and reporting tools so you can see how impressions and chats become pipeline and revenue. This type of data will matter in the next budget cycle.
Claiming Space in AI‑driven journeys
AI advertising inside platforms like ChatGPT is still early, and the interfaces will keep evolving. What seems less likely to change is consumers’ AI usage. If younger buyers are already using AI tools to research big decisions, they will expect brands to meet them there with real help, not just slogans.
“The opportunity is to support someone across their entire journey, even if that means telling them they are not ready yet,” Dawson said. That mindset, and the willingness to experiment where customers are already searching, may be what separates the next wave of AI‑savvy CMOs from those still treating AI as a side project.
