For demand generation and RevOps

AI visibility for demand generation and RevOps teams

AI assistants influence a B2B shortlist by answering category questions that name no company, so the shortlist often forms before a buyer visits any website. Demand generation and RevOps teams can see part of that in their analytics and part of it in what buyers say, and should read both as a floor, never as attribution. The answers below cover how the shortlist forms, how to read inbound that came from assistants, and which channels and programmes to change, followed by a plain guide to reading AI-referred inbound.

Talk to us See all services

A pale flamingo standing on bright yellow legs.

How do AI assistants influence which vendors make a B2B buyer's shortlist?

They influence it by answering the early questions, the ones that name no company. A buyer who asks which tools suit a team of their size, in their sector, with their constraints, receives a written answer that may name a few vendors, put one or two forward as the fit and cite the sources it used. For many buyers that answer is the first draft of the shortlist, and it arrives before anyone has visited a vendor's website.

Three outcomes matter and they are different. Being named means the assistant connects you with the topic. Being cited means one of your own pages was used as a source. Being put forward means the answer chose you, or listed you as an alternative, for the buyer's stated need. A shortlist is shaped mostly by the third, and a team that counts only the first will overestimate its position.

The buyer's own words set the criteria. A question that mentions a regulated sector or a particular integration produces a different list from a general one, so the same vendor can be on one shortlist and absent from another.

The evidence behind an answer is public, and much of it is independent: comparison articles, directory listings, reviews and partner pages. Answers also vary from run to run and between assistants. Nobody outside the assistant makers controls the outcome, so the practical lever is the quality of the evidence available to find.

How can we tell how much inbound pipeline comes from buyers who researched us in AI assistants?

You can tell part of it, and you should say which part. There are three sources of evidence and none of them is complete.

Analytics shows visits that arrive with an assistant named as the referrer, where the assistant sends one and the visitor has accepted analytics cookies. Visits from a buyer who copied a link, typed your name into a search box afterwards or came back direct appear under other sources. A self-reported field on your forms, such as an optional "how did you hear about us" with an option for an AI assistant, records what the buyer says, in their words. Notes from first sales calls record what buyers volunteer and what the assistant told them.

Put the three side by side, with the date each began. Read the total as a lower bound. A buyer who read an answer, closed the window and arrived later by another route is in none of them.

What this cannot tell you is that an assistant caused the enquiry, or how many shortlist decisions were shaped without any trace. So report counts over their bases, keep the definitions the same from month to month, note when you changed a field, and do not assign revenue to an assistant on this evidence. The honest statement is that at least this many opportunities mention one, and that the figure is a floor.

Which demand generation channels and programmes should a B2B team change now that buyers research vendors in AI assistants?

Change little at once, and change the things that follow from the evidence rather than from alarm.

Start by adding the buyer's questions to your planning. Write down the questions buyers ask early, ask them of the assistants, and read which companies are put forward and which sources are cited. That gives your programme a map it did not have.

Then look at where a programme makes evidence that assistants can read. Content programmes that publish in volume may do better by publishing less, and publishing pages that answer a named question and state specific facts. Partner programmes, directory profiles, review sites and analyst coverage deserve a fresh look, because independent sources feed answers and a current, accurate listing costs little.

Change the handoff too. Add the self-reported field to forms and ask sales to note what buyers say they asked an assistant.

Leave paid media, events and outbound as they are unless your own data gives a reason. Nothing here shows that those channels have stopped working.

Keep each change small, with a decision date, and measure the same buyer questions each month so you can see whether the answers moved. Report flat results as plainly as the ones that moved, and do not claim that a channel change caused a change in the answers.

Reading AI-referred inbound without over-claiming

What an analytics tool can see: visits that arrive with an assistant named as the referrer, when the assistant sends one and the visitor has accepted analytics cookies. Group them under one label and note the date you began.
What it cannot see: visits where the buyer copied a link, searched for your name afterwards or arrived direct. These land in other sources, so the group above is a floor.
A self-reported field: an optional question on your forms asking how the buyer heard about you, with a free text box and an option for an AI assistant. Keep the buyer's own words, and read the count as what buyers chose to say.
Sales call notes: ask the first call to record whether the buyer used an assistant and what it told them. This is the richest source and the least consistent, so give it a simple template.
What neither can prove: that an assistant caused the enquiry, or how many shortlists were shaped without a trace. A buyer influenced inside an assistant often leaves no referral at all.
How to report it: counts over their bases, definitions held steady from month to month, a note on each date a field or tag changed, and a sentence stating that the figure is a lower bound and not attribution.

Questions demand generation and RevOps teams ask about AI visibility

Who is this page for?

It is for demand generation, growth and revenue operations teams at B2B technology companies: the people who own the pipeline report and the programmes that feed it. If you run budget, finance or content, there is a page for your role as well.

Can we attribute revenue to AI assistants?

Not on the evidence available. Analytics, a self-reported field and sales notes each show part of the picture, and a buyer influenced inside an assistant often leaves no referral. Report what you can count as a lower bound and do not assign revenue to an assistant.

Do assistants send referral traffic?

Some visits arrive with an assistant named as the referrer and many do not, and the behaviour can change without notice. Check what your own analytics shows and treat it as incomplete.

Why are our analytics counts lower than what buyers tell us?

Analytics only counts visitors who accept analytics cookies, and it cannot see a buyer who read an answer and arrived later by another route. What buyers say on forms and calls fills part of the gap.

Should we move paid budget because of this?

Not on this evidence alone. Nothing here shows paid media has stopped working. Make small changes with a decision date, and measure the same buyer questions each month before deciding more.

How often should we review the numbers?

Monthly is enough for the team doing the work, and quarterly for leadership. Compare the same questions on the same assistants, and note any date on which a tag or a form field changed.

Stand out.
Get chosen.

Create your account in under a minute. Start with your first project and the questions that matter.

Get started Talk to our team

White paper aeroplanes with one yellow plane leading.