There’s no value for marketers in answer engines, but AI agents look promising

“It’s going to get weirder” according to an article I was sent on generative AI and how searching the web is changing. From traditional search to answer engines to agentic AI, the future of search looks bleak for publishers and content creators.

Earlier this week, one of In Marketing We Trust’s search specialists sent me an MIT technology review article by Mat Honan titled ‘AI Means the End of Internet Search as We’ve Known it’. It’s a fascinating read and touches on many themes I’ve been reflecting on. The shifts in search and the role of generative AI have been on my radar since OpenAI launched.

The evolution from traditional search engines to answer engines and agentic AI is disrupting the balance of the value exchange for publishers. While answer engines provide faster, more comprehensive results for users, they fail publishers and marketers, eroding the content ecosystem. Agents, as the next phase, hold promise to restore this balance by creating tangible value for commercial publishers (marketers) through bookings and commercial transactions.

New Value Exchange for Publishers

The value exchange underpinning traditional search has historically worked well. Publishers (including businesses) created content, users accessed high-quality information, and search engines drove trac while monetising through ads. That ecosystem, while not awless, ensured everyone benefitted in a balanced way.

However, the rise of answer engines disrupts this model entirely. While these engines deliver incredible value to users (faster answers, more comprehensive responses), they leave publishers out in the cold, stripping away trac and visibility without offering adequate value in return.

Answer engines, like ChatGPT, are leveraging the work you put into creating content and answering user queries without giving you credit or traffic.

“The AI was hoovering up their premium content, repackaging it, and promoting it to its audience in a way that didn’t really leave any reason to click through to the original.” – Mat Horan, ‘AI Means the End of Internet Search as We’ve Known it’, MIT Technology Review.

While there is some overlap, I think about this transformation in three models:

1. Search engines (Google, Bing, etc.)

Traditional search engines provide ranked lists of web pages based on keyword queries, requiring users to click through and information themselves.

2. Answer engines (OpenAI, Perplexity, Claude, plus a load of new startups)

Answer engines are AI-powered platforms that directly provide synthesised, conversational responses to queries without requiring users to visit external websites.

3. Agents (TBC, but emerging as the next phase, I think Google will nail this)

Advanced AI systems that combine planning, reasoning, and tool-using capabilities to autonomously execute complex multi-step tasks and engage in ongoing contextual interactions.

Note: Your input for answer engines as well as agents is multimodal and the output is multimodal.

Value Exchange Models

Here’s how the value exchange works for each:

Search Engine Value Exchange: Works

• Publisher > produces content = gets traffic on site to generate revenue
• User > searches = gets information they need
• Search Engine > indexes and returns publisher results = generates ad revenue on search results

Answer Engine Value Exchange: Broken  Publisher > produces content = no traffic

• User > searches = gets information they need (faster than ever, likely zero-click)
• Answer Engine > creates answers from publisher information = generates revenue (various models used and this will change over the next 18 months as investors look for returns)

Agent Value Exchange: Potential to Work

• Publisher > produces content = gets a booking or transaction
• User > sets criteria, discusses options = service booked or product bought (faster than ever)
• Agent > searches, researches, books/buys = generates revenue (various models, mostly likely advertisers pay-to-play like an extension of PPC)

The New Value Exchange for Publishers is Broken

The answer engine model is where the cracks are most visible. If publishers do not see value in continuing to create content, we risk losing the very foundation that supports these systems.

Solutions

Some potential solutions include introducing new commercial models, like royalties for zero-click answers, where publishers are compensated based on usage (a nightmare to implement and even harder to audit).

ChatGPT for example has already struck financial deals with publishers, providing payment for the information that gets rolled into its results. However, for OpenAI to work the way it wants to—providing current up-to-the-minute answers—it has to provide information from all kinds of publishers that it doesn’t have deals with.

Alternatively, adopting a brand display model, where publishers receive prominent branding and referencing in answers, could at least provide some visibility and indirect value, though it likely will not drive significant traffic. This will change the way marketers see this part of the funnel. It will be more akin to brand awareness advertising through display/programmatic
advertising.

While there are significant challenges to implementing these solutions they are necessary for the ecosystem to survive.

Greater Value in AI Agents for Publishers

For agents, I think the future looks a lot more promising. If they result in bookings or transactions publishers will see a clear and tangible return. That said, I imagine agents will likely lean toward pay-to-play models. Google, for instance, could favour paid advertisers to maintain its commercial viability, especially when integrating booking and purchasing into these systems.

“Let’s say you have a trip coming up in a few weeks. An agent that can get data from the internet in real time can book your flights and hotel rooms, make dinner reservations, and more, based on what it knows about you and your upcoming travel—all without your having to guide it.” – Mat Horan, ‘AI Means the End of Internet Search as We’ve Known it’, MIT Technology Review.

The key will be to focus on understanding how consumer behaviours shift across these models and to ensure our clients are in the right “consideration sets”, as this is where the battle for visibility will intensify.

Many businesses treat search as a bottom-of-the-funnel activity, to capture demand. We are at a pivotal moment in search where consumers have more choice than ever. Ensuring clients are in the consideration sets customers shift to in this new era is paramount.

Google Still Dominates

Google remains the dominant player representing 91% of search trac worldwide. The greatest competitive advantage for Google is its ability to leverage its own data (the largest structured dataset on the planet).

My hunch is Google can address these challenges before losing too much ground to the newer entrants. As Google has been in the commerce game well before ChatGPT and other gen AI tools came along, they are uniquely positioned to evolve into agents and have a clear line to restoring the value exchange for publishers, marketers and users, albeit, with a different model.

Adapt or Decline

While there is no definitive answer yet, these shifts underline the need to stay ahead of trends and adapt strategies to find the best ways to capitalise on these changes. The search ecosystem is facing rapid change and businesses need to adapt to maintain visibility across platforms.

The role of search experts is now more critical than ever. Having a specialist search team will be pivotal to helping brands optimise their content and strategies in this evolving search landscape.

In Marketing We Trust is tailoring our recommendations to clients to address immediate concerns around the shift in the search landscape by diversifying content strategies, optimising for answer engines and preparing for agent-based platforms.

It’s a fascinating and critical moment for search, and I’m excited to continue exploring how we can best position our clients to thrive in this evolving ecosystem.

Paul Hewett is the CEO of In Marketing We Trust

Got a tip?

topics ai ChatGPT In Marketing We Trust Mat Horan MIT Paul Hewett search search engine value exchange

Share Tweet Share