Skip to content

B2B channel mix after AI search: where buyers find you now, and why trust still runs through people

Ten years ago, the B2B growth playbook was easy to draw. Publish useful pages, rank on Google, turn visitors into enquiries, hand the enquiries to sales. The website was the front door, and search was the street that led to it.

That street is getting quieter. Buyers now ask an AI assistant to explain the problem, list the options and compare suppliers, and many never click through to anyone's website. The questions that used to bring visitors are being answered on the results page or inside a chat window.

None of this means marketing stopped working. It means the jobs inside your channel mix have moved. This guide explains the traditional mix in plain terms, shows what AI search changed and uses five peer-reviewed studies, plus recent buyer research, to suggest where a B2B SME should put its next hour and its next dollar. We say where the evidence is strong, where it is thin and where we are drawing our own conclusions.

What is a channel mix, in plain terms?

Your channel mix is the set of ways people come to know, trust and buy from you, and how you split time and money between them. Marketers usually sort channels into four groups by who controls them.

TypeWhat it meansExamplesWhat you control
OwnedPlaces you run yourselfWebsite, newsletter, customer list, your own eventsAlmost everything, but people still have to find it
PaidAttention you buySearch ads, LinkedIn ads, sponsorships, trade show standsWho sees it and when, for as long as you pay
EarnedWhat others say about youReferrals, reviews, press, mentions in AI answersVery little directly; you earn it with good work and evidence
SharedConversations on platforms you do not ownLinkedIn posts, podcasts, communities, messaging groupsYour own voice, not the platform's reach

Two words carry most of the weight. Reach is how many of the right people see you. Trust is how willing they are to believe you. Every channel does some of each, in different proportions.

How did the traditional B2B mix work?

For most of the last fifteen years, the default mix for a growing B2B company ran in a line:

  1. Get found through search. Write pages and articles around the questions buyers type into Google.
  2. Turn visits into leads. The website explains, persuades and captures an enquiry or a download.
  3. Use paid to speed things up. Search ads and LinkedIn ads push more of the right people to the same website.
  4. Close through people. Sales calls, demos, visits and trade shows turn interest into a contract.

The website was the hub and organic search was the engine that filled it. Paid was an accelerator you switched on when you needed more. People did the closing.

What has AI search changed?

The engine has slowed. Several independent measurements point the same way.

  • Pew Research Center tracked the browsing of 900 US adults in March 2025. When Google showed an AI summary, users clicked a traditional result in 8% of visits, against 15% when there was no summary, and clicked a link inside the summary in 1% of visits. AI summaries appeared far more often for longer, question-style searches: 60% of searches starting with "who", "what", "when" or "why" produced one.
  • Ahrefs compared 300,000 keywords using desktop Search Console data from December 2023 and December 2025. The presence of an AI Overview correlated with a 58% lower click-through rate for the page ranked first.
  • Forrester's 2025 Buyers' Journey Survey found that 94% of business buyers use AI somewhere in their buying process. Twice as many buyers named generative AI or conversational search as a more important source of information than any other source, ahead of vendor websites, product experts and salespeople. Forrester concludes that driving traffic to your site to retarget and nurture prospects "will be much less effective".

Treat the exact numbers with care. Pew measures US consumers, Ahrefs uses one company's dataset and reports a correlation, and buyer surveys differ: TrustRadius, below, found 63% of software buyers used AI in their purchase. The direction is consistent: more questions are answered before anyone clicks.

Here is what that does to the jobs in your mix. Your website does not become less important. Its job changes. It moves from being where buyers discover you to being where they verify you: the place they check after an AI tool, a colleague or an ad has put your name in front of them. We cover what buyers look for in that check in how B2B buyers research suppliers before making contact.

Why does being known early matter more now?

Because research has become confirmation. In TrustRadius's 2026 B2B Buying Disconnect report, a January 2026 survey of 1,862 technology buyers, 79% had already heard of the product before they started researching. 83% shortlisted three or fewer products, and 67% bought their first choice.

LinkedIn's B2B Institute calls this the 95:5 rule, after Professor John Dawes of the Ehrenberg-Bass Institute: at any moment only about 5% of B2B buyers are in the market. It is a rule of thumb, not a published study, but the implication holds. Marketing rarely pushes a buyer into the market. Its job is to make sure your name comes to mind when the buyer gets there.

So the question for an SME is not only "how do we get found?" It is "how do the right people already know us when the need arrives?"

Why is paid now the most controllable way to stay known?

Because it is the one discovery channel where you still decide who sees you, how often and when. Organic search used to offer something close to that for the price of good content. As AI answers absorb more clicks, that bargain weakens. AI answers are not ad-free either: in 2025 Google announced it was expanding ads inside AI Overviews, in English in selected countries.

Three cautions keep this honest.

Paid is rented, not owned. The reach stops the day the spending stops. What you keep is what the spending leaves behind: a name buyers recognise and contacts on your own list.

Advertising works, but modestly, and partly over time. The best-known meta-analysis, by Sethuraman, Tellis and Briesch (2011), pooled 751 short-term and 402 long-term estimates from 56 studies. The average short-term elasticity was 0.12: roughly, 10% more advertising went with about 1.2% more sales in the same period. The median was lower still, at 0.05, and only 57% of the estimates that could be tested were clearly above zero. The long-term average was 0.24, about double. Effects were higher for products early in their life cycle and had declined over the decades, a fall the authors link partly to advertising clutter. The authors suggest more precise targeting and distinctive messages as ways to counter that.

The B2B evidence is thin. That meta-analysis deliberately excluded business-to-business advertising, noting that B2B studies are few. Its data also ends in 2004, before most online advertising. So nobody can tell you from peer-reviewed research what a LinkedIn budget will return for a B2B SME. What the research supports is a direction: advertising pays back slowly, and precise targeting matters.

Put together with the 95:5 idea, the job of paid is to keep your name familiar to a defined group of accounts, steadily, so it is already on the list when they go into the market. If you judge it only on this month's leads, you will cut the part that does the long-term work. It is also getting dearer: Gartner's 2025 CMO Spend Survey, of 402 mostly large companies, found paid media takes 30.6% of marketing budgets and that media inflation means less for every dollar.

Why does trust run through people, not pages?

Because buyers now check almost everything, and what survives the check is what vendors do not control.

In the TrustRadius survey, 47% of buyers trusted online resources less than a year earlier, up from 39%. Of buyers who used AI, 94% fact-checked its answers at least some of the time, and 72% did so always or very often. Vendor marketing collateral ranked last among the resources buyers consult. Of the 53% who spoke to a peer during the purchase, every one found it at least somewhat helpful, and buyers ranked demos, free trials, prior experience and user reviews among the most influential resources.

The peer-reviewed research explains why people carry trust better than pages.

Relationships stick to individuals, and expertise builds them. Palmatier, Dant, Grewal and Evans (2006) combined 637 correlations from 94 studies, with a combined sample of 38,077. Strong relationships mattered more in business markets than consumer ones: their link to seller performance was 0.36 against 0.25. They also mattered more when the customer's relationship was with a person rather than the company: the link to loyalty was 0.56 against 0.46. The most effective ways to build a relationship were the seller's expertise (0.62) and communication (0.54), followed by investment in the relationship and similarity to the customer. How often the two sides met (0.16) and how long the relationship had lasted (0.13) mattered much less.

That last finding is the useful one for events. Showing up more often is not what builds trust. Showing what you know, and having a real conversation about the customer's problem, is.

Asking in person beats asking by email, by a wide margin. In two experiments, Roghanizad and Bohns (2017) had students ask strangers for a favour, either face to face or by email. In the first, 45 requesters each asked 10 strangers to fill in a questionnaire. Both groups expected about five in ten to agree. In person, an average of 7.2 did. By email, an average of 0.2 did: face-to-face requests were 34 times more likely to succeed. In the second study, with a different task, in-person requests still succeeded more than twice as often. The mechanism was trust: recipients trusted and empathised with the person in front of them far more, and the requesters did not see that coming. The participants were students on one campus, not buyers, so take the size of the gap as a signal, not a forecast.

People distrust content they assume a machine wrote alone. In two preregistered experiments with 4,976 US and UK adults, Altay and Gilardi (2024) found that labelling news headlines "AI-generated" made people rate them as less accurate, even when the headlines were true and written by humans. The penalty was small, about a third of the effect of labelling a headline "false". The reason was the assumption that AI had written the whole thing without human oversight. When people were told the AI had only improved the clarity of a human's text, or written a first draft from sources a journalist chose, the penalty disappeared. The study used news headlines rather than marketing, but the lesson transfers: what costs trust is content that looks unsupervised, not the use of AI as such.

Put together: an event, a site visit, a call or a recommendation from someone the buyer knows carries trust that a page cannot, provided the person brings real expertise. For an SME that is good news. Expertise and presence are things you can offer as well as any large competitor.

Do parasocial relationships matter in B2B?

A parasocial relationship is the one-sided sense of knowing someone you have only seen or heard: a podcast host, a writer, a founder whose posts you read every week. You have never met, but you feel you know how they think.

Schramm, Liebers, Biniak and Dettmar (2024) reviewed 281 studies of these relationships published between 2016 and 2020. Across media, the same factors kept strengthening them: perceived similarity, competence and trustworthiness, credibility, authenticity and, for audio, a familiar voice. Real interaction made a difference too: people who had received replies from a personality on social media formed stronger bonds. Podcast listeners sometimes played episodes on topics they did not care about, because of the host. Several studies linked strong bonds with radio hosts and influencers to more positive brand attitudes and purchase intentions, especially when the person and the product fitted. People with more knowledge of persuasion tactics were less swayed.

Three limits matter. The research is about media audiences and consumer advertising, not B2B buying. Most samples were young and mostly female. Most studies were surveys that show association rather than cause. Applying it to B2B is our inference.

It fits what B2B buyers report. The 2025 Edelman and LinkedIn B2B Thought Leadership Impact Report looked at "hidden buyers": people in finance, legal, procurement or operations who influence a purchase without leading it. 95% said strong thought leadership made them more receptive to sales and marketing outreach. 53% said strong content can outweigh brand recognition alone, and understanding of their business challenges (85%) and domain expertise (74%) mattered more to them than "the safest choice" (41%). Edelman's page does not give the survey's sample size, and the report is vendor research, so treat the figures as indicative.

In practice, a buyer who has read your founder's posts for six months, heard your head of operations on a podcast or watched your engineer explain a problem on video arrives at the first meeting already half-introduced. Business buyers know a sales pitch when they see one, which is why the content that builds this familiarity is expertise, not promotion.

What does the new B2B channel mix look like?

The channels are the same. Their jobs have moved. Use this table to check how your current mix is set up.

JobOld defaultNew defaultWhat to measure
Be foundOrganic search to your websitePaid placements aimed at chosen accounts, plus being mentioned where AI tools lookShare of target accounts reached; how AI tools describe you
Be rememberedOccasional campaignsSteady paid presence and named experts publishing regularlyBranded search, direct visits, answers to "how did you hear about us?"
Be trustedWebsite copy and case studiesPeople showing expertise: talks, hosted sessions, visits, calls, peer referralsMeetings held, referrals received, conversations that move to a next step
Be verifiedWebsite as brochureWebsite as evidence: proof, people, prices, consistency, reviewsEnquiry quality; time from first visit to first call
CloseSales teamSales team, often with a founder or expert in the roomWin rate; sales cycle length

If most of your effort still sits in the "old default" column, you are not doing anything wrong. You are building for a route that fewer buyers take.

Worked example: rebalancing an SME's mix

This is an illustrative example, not a client. A B2B engineering services firm with about US$2 million in revenue spends most of its small marketing budget on blog articles written for search, plus occasional Google Ads. Website traffic has fallen for a year, while the owner notices that most new contracts still come through people they have met.

The rebalanced mix keeps the same budget and changes the jobs:

  1. Paid narrows. Broad search ads are trimmed. A modest, always-on LinkedIn budget puts the owner's and the lead engineer's content in front of a list of named target accounts every month, judged quarterly.
  2. Content gets a face. Fewer anonymous articles; more short pieces under named experts, with real project detail an AI tool cannot invent. Where AI helps with drafting, the team says so.
  3. Events are built around expertise. Two big trade-show stands a year become a small hosted breakfast for target accounts each quarter, built around a short technical talk and time for questions, plus one talk at an industry association.
  4. The website becomes the verification layer. Case studies, the named team, certifications, reviews and a clear first step, so anyone who meets the firm through an ad, an AI answer or a breakfast can check it in two minutes.
  5. Measurement changes. Alongside leads, the team tracks target accounts reached, meetings held and the answer to "how did you hear about us?"

Nothing in this plan is exotic. It moves effort to where trust is now formed.

What should B2B SME owners do first?

  • Ask the AI tools. Ask two or three assistants to recommend suppliers in your category and region, and to describe your company. Note what they get wrong and which sources they seem to rely on.
  • Map your current mix. List every channel, what it costs in money and time, and which of the five jobs in the table it does.
  • Put an expert in front. Choose one or two people whose expertise buyers would value and give them a small, regular publishing habit on the platform your buyers use. Teach, do not pitch.
  • Make events about expertise, not attendance. Pick fewer events, prepare a target list, bring something worth learning and follow up within a week. For referrals and introductions, ask in person.
  • Keep paid steady and judge it slowly. Aim a modest budget precisely at the accounts that matter and review it over quarters, not weeks.
  • Rebuild the website for checking. Proof, named people, reviews and a clear first step. The 30-second positioning test and our guide to collecting testimonials help here.
  • Use AI for the work, and say how. Let it research, draft and edit. Keep the opinions, stories and relationship messages human, and be clear about where AI helped. For Singapore owners, our guide to the Singapore Digital Economy Report 2026 shows why AI saves time but does not win customers on its own.

What should founders read?

If you want the evidence behind this guide, start here, in order of usefulness for an owner:

  1. Palmatier, Dant, Grewal and Evans (2006), relationship marketing meta-analysis. What actually builds business relationships: expertise and communication, and people over companies.
  2. Roghanizad and Bohns (2017), "Ask in person". A short paper that will change how you make requests.
  3. Altay and Gilardi (2024), AI labels and trust. Why people distrust content they assume is fully automated, and what fixes it. Open access.
  4. Sethuraman, Tellis and Briesch (2011), advertising elasticity meta-analysis. A realistic sense of what advertising does and how long it takes.
  5. Schramm, Liebers, Biniak and Dettmar (2024), review of parasocial research. What makes a familiar voice persuasive. Open access.

For current buyer data, read the TrustRadius B2B Buying Disconnect report, Forrester's buyer journey research and the Edelman and LinkedIn thought leadership report. Re-check them every year: the numbers move fast.

Machines answer, people decide

AI search has taken over the first question a buyer asks. It has not taken over the last one: can I trust these people with my business?

The winning mix for a B2B SME is not more content or a bigger ad budget. It is a clear division of labour. Pay to stay familiar to the right people. Let your experts become voices buyers recognise. Meet buyers in person whenever you can, and bring something worth learning. Make your website the place where everything they have heard is easy to confirm.

Questions people also ask

What is a B2B channel mix?

The combination of channels a business uses to reach, persuade and win customers, and how it splits time and money between them. Channels are usually grouped as owned (your website, list and events), paid (ads and sponsorships), earned (referrals, reviews, press) and shared (social platforms and communities).

Is SEO still worth it for B2B companies?

Yes, but its job is changing. Pew and Ahrefs data show AI summaries reduce clicks to websites, so ranking alone brings fewer visitors. Clear, specific, well-evidenced pages still help buyers verify you and help AI tools describe you accurately.

Is paid advertising owned media?

No. Paid media is rented: you control who sees it while you pay, and the reach stops when the spending stops. What you keep is the familiarity it builds with buyers and the contacts it adds to your own list.

Does advertising work for B2B companies?

The research on advertising's average effect is mostly from consumer brands: a meta-analysis of 56 studies found a modest short-term effect that roughly doubles over the longer term. Peer-reviewed B2B evidence is scarce, so judge B2B paid media over quarters and aim it precisely at the accounts that matter.

Why are events and face-to-face meetings important in B2B?

Because trust attaches to people. A meta-analysis of 94 studies found relationships matter more in business markets and when they are with an individual rather than a company, and that expertise and communication build them far more than frequent contact. An experiment found in-person requests 34 times more likely to succeed than emailed ones.

Should B2B companies use AI to write marketing content?

For research, drafting and editing, yes. Experiments show people distrust content they believe AI wrote without human oversight, but the penalty disappears when they know AI only drafted or edited a human's work. Keep the voice and the relationships human, and be open about how you use AI.

Sources

  1. Factors Influencing the Effectiveness of Relationship Marketing: A Meta-AnalysisPalmatier, Dant, Grewal and Evans, Journal of Marketing 70(4), 136-153, 2006
  2. How Well Does Advertising Work? Generalizations from Meta-Analysis of Brand Advertising ElasticitiesSethuraman, Tellis and Briesch, Journal of Marketing Research 48(3), 457-471, 2011
  3. Ask in person: You're less persuasive than you think over emailRoghanizad and Bohns, Journal of Experimental Social Psychology 69, 223-226, 2017
  4. People are skeptical of headlines labeled as AI-generated, even if true or human-made, because they assume full AI automationAltay and Gilardi, PNAS Nexus 3(10), pgae403, 2024
  5. Research trends on parasocial interactions and relationships with media characters: a review of 281 studies from 2016 to 2020Schramm, Liebers, Biniak and Dettmar, Frontiers in Psychology 15, 1418564, 2024
  6. Google users are less likely to click on links when an AI summary appears in the resultsPew Research Center, 2025
  7. Update: AI Overviews Reduce Clicks by 58%Ahrefs, 2026
  8. B2B Buyers Make Zero-Click Buying Number OneForrester, 2025
  9. TrustRadius 2026 B2B Buying Disconnect report: press releaseHG Insights, 2026
  10. Trust more, verify everything: the 2026 B2B Buying Disconnect reportHG Insights, 2026
  11. The rise of the hidden buyer: 2025 B2B Thought Leadership Impact ReportEdelman and LinkedIn, 2025
  12. The 95:5 rule is the new 60:40 ruleLinkedIn B2B Institute and Ehrenberg-Bass Institute, 2021
  13. Gartner 2025 CMO Spend SurveyGartner, 2025
  14. Google Marketing Live 2025: your roundup of announcementsGoogle Ads Help, 2025

Let’s make it unfair.

Start with a free briefing call. We look at the business like owners, find where the growth engine is leaking momentum, and show you what to fix first, whether or not we end up working together. Keep the findings either way.

Book a briefing call