Your buyers now do much of their deciding inside answer engines. In 6sense's 2025 Buyer Experience Report, a survey of nearly 4,000 B2B buyers found that 94% use LLMs in their buying process, and most of that use happens mid-journey as they compare options and build a shortlist. The same report found that buyers choose a vendor from their day-one shortlist 95% of the time. Consumers are on the same path. Adobe found that visitors who reach US retail sites from AI tools convert at a 54% higher rate than other traffic and bring in 53% more revenue per visit.

That makes AI search a demand channel, and chances are you’re already tracking it. Many companies now have an AEO platform, often run by the SEO or AEO team, that shows how often the engines recommend them, for which questions, and next to which competitors.

What still takes time is everything that comes after the dashboard. You size a new topic from the data, work out where buyers drop you, build a campaign around what they ask, and pull together what happened afterward. That’s where AEO platforms are heading next. The best among them are becoming agentic, running agents that take on that work, from sizing demand to drafting the campaign readout.

Profound runs its Agents on the data it collects: more than 2 billion real user prompts, the pages each engine cites in reply, and how they describe you and your competitors. You can build your own agents in a drag-and-drop builder or start from Profound’s library of agent templates.

Below, we walk through five jobs most growth and demand gen marketers own, and the agents that help with each.

Size the demand before you commit budget

Demand sizing usually starts from search volume, paid search data, and your own pipeline history. All three mostly count people who already know what to look for, or who already found you. Buyers asking AI are often earlier than that. They describe the problem before they know which category solves it, and the questions they ask don't always look like the keywords you bid on.

Prompt Volumes shows how often people ask answer engines about a topic, how that's trending, and the intent behind the questions. It's built from conversations Profound licenses from opt-in panels of real answer engine users, and you can break volume down by platform, age range, and income bracket across ten countries. It also shows the prompts where a competitor's content is cited and yours isn't, which is a short list of demand you're currently handing to someone else.

Two agents turn that into a plan for a specific topic:

  • The Pull SEO + AEO Keyword Volume agent reports search demand and answer-engine demand for a topic side by side, so you can see where the two diverge before you put budget behind either.
  • The AEO + SEO Research Report agent shows which keywords have search volume, what ranks for them on Google, and how the answer engines respond to the same questions, including the sources they cite.

If you run paid search, Profound's Google Ads integration pulls your campaign, keyword, and conversion data into Profound Agents and maps it against your performance in Google's AI Mode and AI Overviews. An agent built on that data can flag topics where you're paying for clicks while AI Overviews already recommend you, and topics where you're missing from both.

Jordan Digital Marketing, an agency that runs organic and paid programs for high-growth companies, used Prompt Volumes to find the questions expecting parents ask AI engines for a nursery furniture client. The team built short-form content around those questions, and the brand became a top recommendation in ChatGPT.

Find where buyers drop you while they're still deciding

Buyers can drop your brand at any point in their research, and where they drop you matters. Missing from the first list of options is a different problem from making the shortlist and losing the comparison, and each one calls for a different solution.

One way to find the point is to track different kinds of prompts separately. Chapter three of our AEO Guide describes six types, each testing something different. For example, discovery prompts check whether you get recommended when someone is exploring options, while competitor prompts check who the engine picks when it’s narrowed to two named brands. If you tag the prompts you track by type, Answer Engine Insights shows your visibility for each tag, and the gaps between them point to where buyers drop you. A brand that shows up in 40% of discovery answers but 10% of competitor answers is making shortlists and losing the comparison.

Once you know the stage, two tools show you why:

  • Query Fan-Out shows the searches an engine runs behind each prompt, which of those searches most often decide the sources it cites, and which words it adds when it rewrites the prompt. At the comparison stage, those searches tell you what the decision hinges on, like pricing, integrations, or support.
  • The Citation-to-Content Engine agent looks at which pages the engines cite across a broad topic, works out the patterns behind your competitors' citations, and returns the specific prompts where a page from you could compete.

1840 & Co., a remote staffing company, started at 0% visibility, missing from the discovery answers where Toptal and Upwork dominated. Profound's citation data showed the engines favored short comparison lists, direct statements about why to choose a company, and FAQ sections, so the team published one post built that way. Its visibility reached 6% within two weeks and 11% by month's end, putting it among the top five brands in AI answers on remote staffing.

Check that you're reaching the right buyer

Many products sell to more than one kind of buyer, and each one describes the problem in their own terms. A CFO might ask about cost and payback, while an operations manager asks about setup and workflows. When a question carries that kind of context, whether in its wording or in what the engine already knows about the user, the answer can change, and so can the shortlist. The people who convert tell you which buyers you reached, but they can’t tell you whether the engines recommend you as readily to everyone else.

You can test that directly instead of assuming it:

  • Personas in Answer Engine Insights. A persona is an audience profile, like “small business owner” or “enterprise IT buyer,” that Profound adds as context when it runs a prompt. Profound runs the base prompt and the persona version separately, so you can compare the answers and see whether your visibility holds for each buyer or drops for one of them. Each persona adds a run for every prompt it’s assigned to, so start with the two or three that matter most.
  • Query Fan-Out by persona. Query Fan-Out shows the searches an engine runs behind a prompt, and lets you compare them across personas. If the CFO version of a prompt sets off searches about pricing and the operations version sets off searches about integrations, those are the sub-questions each buyer’s page needs to answer.
  • The Customer Question Miner agent. It pulls recent AI answers for one of your Profound topics, filtered by persona if you choose, along with the searches the engines ran to write them. It then writes a report on the top questions, the themes and intent behind them, the questions nobody answers well, and the content angles worth pursuing. Run it once per persona, and you have a brief for each segment’s content and campaigns.

Segments aren’t always job titles. Arizona College of Nursing recruits prospective students in 24 local markets, and each market has its own competitors and its own questions. The team built a library of more than 300 market-specific prompts, organized with custom tags, and tracks, for each market, how often the college is cited, which competitors outrank it, and how that varies by content type.

Its agents are mapped to campus tags and local page formats, so the team runs the same agent for each market and acts on what it finds, from tuition comparisons to local FAQs about admissions. In 90 days, AI-referred enrollment inquiries rose 51%, driven by content that surfaced in program comparison and cost questions, and the college became the most cited domain for its priority prompts.

Build the nurture sequence for a segment

The questions a segment asks AI are useful raw material for a nurture sequence. They show what buyers in that segment still want to know, in their own words, and because Profound tracks them over time, you can see when those questions change and update the sequence to match.

This is a good job for an agent you build yourself. For example:

  • Start with a step that pulls the questions one persona asks in your category from Answer Engine Insights.
  • Add an LLM step that drafts a sequence answering them in the order buyers tend to ask, from the broad problem to the comparison questions, and give it your messaging and approved claims. For an accounts payable product, a finance manager's questions might run from how to stop paying invoices late, to whether a tool integrates with NetSuite, to how two vendors compare for a 200-person company, and each one becomes an email.
  • To cover several segments at once, run the same agent in a spreadsheet with one row per persona. If you run email in Klaviyo or Customer.io, Profound integrates with both for email campaigns and workflows. Nothing goes out until you approve it.

The same questions make good briefs for landing pages and ad copy, since buyers already use that phrasing when describing the problem.

Find out what happened after the campaign

A campaign readout for AI search has to answer a few questions that web analytics can’t answer alone. Did the engines start recommending you more often after the campaign, and for which questions? What did your competitors do in the same window? And did any of it bring people to your site?

Profound helps you find the answer for each:

  • In Answer Engine Insights, set a date range for the campaign and compare it with the period before, against the competitors you track.
  • Agent Analytics shows when AI crawlers visit your pages and how many visitors who arrive from AI search go on to convert. Connect your GA4 property through the Google Analytics integration and Agent Analytics can import your conversion and revenue metrics, so AI activity sits next to business outcomes. Profound has an Adobe Analytics integration as well.
  • If you advertise in ChatGPT, the OpenAI Ads integration reads your campaigns, ad groups, and ads, so you can report ChatGPT ads performance alongside your organic AI search performance.

Comparing against competitors also shows how much of a change to credit the campaign for. AI visibility can shift for reasons unrelated to you, like an engine changing how it answers questions about your entire category. If your visibility rose 5 points during the campaign and most of your competitors rose by about the same amount, the campaign probably isn’t the main reason. If you rose while they held flat, you have a much stronger case that it was.

CRS Credit API connected Profound's data to Google Analytics and Looker to follow buyers from AI answers through to MQLs and closed deals. It has seen 15% growth in pipeline attributed to AI search traffic.

Work with your team's AI Marketer in Profound

Each agent above does one job when you start it. The harder call is which job deserves your week, and in Profound, that’s what AI Marketer (Aim) is for. You can work with it even when the workspace belongs to your SEO or AEO team.

Aim is the orchestrator inside Profound, a marketing teammate that reads everything your team tracks there, including prompt volumes, Agent Analytics, and benchmarks against competitors. Every week, it surfaces a ranked list of opportunities, including demand gen insights, to act on, like rising prompt volume in a category you don't own yet or a falling benchmark against a competitor.

When your team picks an opportunity, Aim turns it into a project with a goal, the data behind it, and concrete tasks, then builds and runs the agents that execute them. A project built around a rising category might pair a landing page brief and a nurture sequence with new content from the SEO team. Aim tracks the work as it ships, and the project shows the impact each task had on its metric, which covers much of the campaign readout you'd otherwise assemble. You can also ask Aim about the data directly, like why AI-referred signups dipped last month, and get to the cause behind the number. Nothing goes live without approval.

If your SEO or AEO team already uses Profound, ask for access. And if it's choosing an AEO platform, ask to be part of the decision, since much of the data it brings in is demand data.

See where AI search already sends buyers your way, and where it doesn't. Book a demo.

AI agents for growth and demand gen marketers FAQs

How is AI search demand different from search volume?

It counts questions instead of keywords, and it reaches buyers earlier. People describe a problem to an AI engine in full sentences, often before they know which category of product solves it, so the topics and phrasing look different from the keywords you'd find in a keyword tool. Profound's Prompt Volumes is built from licensed panels of real answer engine users, which lets you see those questions, how often they come up, and how they're trending.

The two numbers work best side by side. Where search volume is high and AI demand is low, your existing search programs still carry the load. If AI demand is rising while search volume stays flat, buyers may be moving their research into AI, and that's where to shift effort first.

Can we attribute pipeline to AI search?

Partly. Agent Analytics shows when AI crawlers visit your pages and how many visitors who arrive from AI search convert, and with GA4 connected, it can import your conversion and revenue metrics. Joining that with your CRM lets you follow AI-referred visitors into pipeline, which is how CRS Credit API tracks AI search through to closed deals.

Where does demand gen fit in an AEO program?

Close to the center, even when the SEO or AEO team runs it. Chapter six of our AEO Guide measures an AEO program in three layers: visibility, traffic attribution, and revenue impact. SEO usually owns the first. The other two are demand gen's, because they're the numbers you already report on.

In practice, the AEO team tracks the prompts and coordinates the work, while demand gen sizes the opportunity, builds campaigns around what buyers ask, and reports on what AI search brought in. Working in the same platform keeps those numbers connected instead of split across two teams' spreadsheets.