The AI Influence Economy | 2026 Spotlight × Profound Report

November 13, 2026
Share

The AI Influence Economy

How Earned Authority in B2B Buying Influences AI Visibility & Sentiment

Introduction: What Changed in Twelve Months

Your buyers are asking AI systems which vendors to shortlist. The answer they get is assembled from whatever sources the engine trusts. So the first question shouldn't be how to rank higher. It should be who the engine is listening to, and what those sources say about you.

To find out, Spotlight and Profound read more than a million AI answers across B2B software and services markets and traced the sources behind them: what the engines cite, and who they name. The year-over-year figures come from four of those markets, measured twelve months apart on the same six engines.

Three numbers set the scene:

  • The demand: 33 million. That's how many B2B buying questions go to ChatGPT alone every day, up from 20 million a year ago. Across every major AI assistant, a conservative estimate is 65 to 100 million a day. The shortlist now forms before a buyer ever reaches your site.
  • The supply: 10 became 8. The engines cite a quarter fewer sources per answer than they did twelve months ago, down 23% to 24% in all four markets. More questions are being answered from less evidence.
  • The evidence: +72%. Analyst firms and customer reviews gained share of every answer, in all four markets (+43% to +187%). Over the same period the trade press fell in every market and Wikipedia dropped out of the top fifty everywhere.

More questions, answered from fewer sources. The sources holding their place are the ones with a human voice of authority behind them.

What you'll walk away with

  • You'll see who is named, who is cited, and who dropped out. The most-cited source in HR technology is now YouTube, which was absent from the top five a year ago. In IT services, McKinsey left the top five vendors and EY took its place.
  • You'll see that no two markets have the same trust layer. Clutch is the most-cited source of any kind in IT services. In HR technology and data security it doesn't reach the top five at all.
  • You'll see why showing up doesn't mean showing up well. The vendor with the biggest visibility gain in its category is also the one the engines describe worst on price and deployment.
  • You'll see why counting citations alone can mislead you. Most answers that name an analyst firm never link to it. Measure this channel by citations and you see a fraction of what it does.

Part 1: Who Speaks for You

AI visibility is an orchestration problem, not an optimization problem

Most of this market has decided that AI visibility is a technical discipline. It already has acronyms (AEO, GEO) and a familiar shape: work out what the engines reward, then produce more of it. Schema, structure, freshness, more pages, better markup.

That work matters. Think of it as your résumé. You have to have one, it has to be accurate, and it has to be a fair reflection of what you offer. Nobody gets hired without one. In a job search, the résumé gets you the interview and the references get you the offer. References get checked late, once someone has already decided they're interested in you.

The machines run it backwards. Everything said about you by someone else (analysts, review platforms, communities, creators, journalists, customers) is your reference list, and the engines read it first. They decide from that reading whether your résumé is worth opening at all.

Those references are your market's trust layer: the signals the engines weigh when a buyer in your category asks who to buy. Every market has one. No two are alike, and yours likely moved this year.

When your references agree, the story is clean and the buyer moves. When they disagree (your website says one thing, your reviews another, the analyst covering you a third), the engine hedges, the answer gets vaguer, and the buyer keeps looking.

The work is getting every channel and every voice telling a consistent story about you, so the buyer meets one version of you instead of five.

The prize is bigger than a mention. According to 6sense's 2025 Buyer Experience Report, 77% of the time the business goes to the vendor the buyer contacts first. That's not because the first vendor is better. It's because they frame the problem, set the criteria, and become the standard everyone after them is measured against. AI answers are increasingly where that first contact gets decided.

Counting citations is the wrong scoreboard. A citation tells you an engine linked to a page. It doesn't tell you whether the engine recommended you, or how it described you when it did. So the question this report works on isn't "how do I get cited more?" It's which voices the machines are listening to, and why it matters that they all tell the same story about you.

Ten sources became eight

We asked the same questions twelve months apart: identical prompts, verified word for word, on the same six engines, with roughly the same number of answers. The engines cited a quarter fewer sources.

Sources cited per answer

Category

July 2025

July 2026

Change

HR Technology

10.3

7.9

−23%

Data Security & Privacy

10.9

8.3

−24%

Digital Commerce

10.6

8.1

−23%

IT Services & Consulting

10.0

7.6

−24%

Four markets and four independent sets of questions showed the same contraction. The answer layer isn't expanding to make room for everyone who wants in. It's getting smaller, and every answer a buyer reads now rests on a thinner evidence base than it did a year ago.

And the shortlist got shorter

The contraction isn't only happening to sources. It's happening to vendors.

Share of answers naming the category leader

Category

Leader

July 2025

July 2026

Change

Digital Commerce

Shopify

36.9%

52.3%

+15.4pp

HR Technology

Workday

29.6%

42.6%

+12.9pp

Data Security & Privacy

Microsoft

50.1%

60.9%

+10.8pp

IT Services & Consulting

Accenture

30.2%

35.9%

+5.7pp

In all four markets the leading vendor gained presence, and in three of them the gain was ten points or more. Shopify is now named in more than half of all digital commerce answers, and Microsoft in three of every five data security answers.

Those gains came from somewhere. Adobe lost two and a half points and WooCommerce two. BambooHR led HR technology a year ago and now sits third. IBM was the third most-named vendor in data security and is no longer in the top five.

The two contractions compound. Fewer sources back each answer, and the answers name a narrower set of vendors. When a buyer's shortlist starts with an AI answer, you don't lose the deal later. You were never in it.

The sources changed, not just the number

The count fell by a quarter. The mix changed far more than that.

Share of AI answers drawing on each source type, July 2025 to July 2026

Source type

HR Tech

Data Security

Digital Commerce

IT Services

Trade & business press

69.4% → 42.8%

51.8% → 25.2%

20.0% → 15.2%

22.1% → 12.3%

High-volume comparison sites

0.0% → 20.1%

0.0% → 22.7%

0.3% → 13.0%

0.0% → 8.5%

YouTube

4.4% → 20.5%

2.4% → 9.6%

3.2% → 11.2%

1.0% → 10.3%

Wikipedia & reference

7.7% → 0.6%

11.9% → 1.4%

13.8% → 3.3%

6.9% → 1.4%

  • The trade press was the backbone, and it's being dismantled. This is the largest single displacement in the data. If your program was built on earned media, this is the number to sit with.
  • Video arrived from nothing. A year ago YouTube wasn't in the top twenty sources in any category we studied. Today it's the most-cited domain in two of them.
  • Wikipedia was deprecated everywhere. It ranked seventh in HR technology and now ranks 210th.
  • Comparison content arrived from zero. Four sites account for almost all of it (worldmetrics.org, zipdo.co, gitnux.org and wifitalents.com), each publishing close to 72,000 "best tools" lists across ninety-plus categories. All four describe an AI-assisted editorial process in their own words. They were first cited in February. By July, one of them was a top-five source in every market we studied.

The engines aren't being fooled here. They're being efficiently served: a buyer asks for the best ecommerce platform, and these sites have a page titled exactly that, ranked, current and structured. If your answer to AI visibility is to publish more, this is what you're publishing against.

The engines are already diverging on comparison content

That share is an average, and the average hides a split. In data security, between April and July 2026:

  • Google AI Mode cut its use of AI-generated comparison content by 91%, from 0.89% of its citations (419) to 0.08% (34).
  • Microsoft Copilot grew its use nineteenfold, from 2.76% of its citations (115) to 51.60% (6,338).

The same direction held in all four markets. Every Google property moved away from this content at scale, while Copilot, Perplexity and ChatGPT all moved toward it. A source that rises on being easy to retrieve can fall just as fast, and no amount of formatting saves it once an engine decides to weigh it differently. What your buyer is told now depends on which engine they asked.

One engine appears to have made that call by judgment. The industry is now building the means to make it automatically. In August 2026 Anthropic began watermarking Claude's text output with an invisible statistical mark, and it's one of roughly 190 signatories to the EU transparency code. A watermark doesn't prove who wrote a page. It shows that a model was probably involved. An engine doesn't need proof to act: for a site publishing 72,000 comparison lists, a likely mark is enough to stop citing all of them.

A prediction: an answer layer where machine-made comparison sites already reach one answer in five won't stay that way once the engines can identify them cheaply. What survives is what a machine can't produce.

One layer gained while others dropped

The contraction didn't fall evenly. The press fell in all four categories and Wikipedia lost 75% to 90% of its share. Meanwhile analyst firms and customer review platforms gained ground in every single one. The third-party layer didn't grow. It reshuffled away from editorial and toward the structured and the verifiable.

Analyst firms and review platforms, share of all citations

Category

2025

2026

Change

IT Services & Consulting

1.2%

3.5%

+187%

HR Technology

2.4%

4.2%

+78%

Digital Commerce

1.1%

1.9%

+77%

Data Security & Privacy

3.8%

5.4%

+43%

The engines cite fewer sources than they did a year ago, and the ones holding their place are the ones you don't write. So how far can your own content take you?

Part 2: What the Machines Reward

Nobody wins on their own website

We plotted 915 vendors across ten B2B categories against a single variable: how much of what the engines say about them is backed by their own content. (Six AI assistants, vendor-shopping questions only.)

Below roughly half, owned content is no obstacle. The strongest vendor in each of those bands reaches 54% to 70% visibility. Plenty of vendors lean hard on their own material and still win, which is why "owned content is a liability" is the wrong conclusion.

Past half, it stops. The best performer in the 50–60% band reaches 9% visibility, and no band above it recovers past 24%. By 80–90%, the strongest vendor sits at 1.3%. Not one vendor in 915 is both heavily reliant on its own content and highly visible. That's not a gentle correlation. It's a cliff, and it sits at about fifty percent.

If more than half of what backs your AI mentions is your own material, you're not approaching a visibility ceiling. You're past one.

The clearest illustration is a vendor whose content the engines reward heavily. In data security, sentinelone.com is the single most-cited domain in the category: ahead of every publication, every analyst firm and every review platform. 94.3% of those citations point at one library, /cybersecurity-101/, a set of explainers on what ransomware is, what a SIEM does and how zero trust works.

In twelve months, SentinelOne's visibility went nowhere. The explainers get cited when a buyer asks what something is. They're almost never there when a buyer asks who to buy, and not one page in that library names a competitor.

Owned content can succeed as content and fail as influence. Apply that test to your own best-performing pages. Are they turning up when a buyer asks what something is, or when a buyer asks who to buy? Only the second one moves a shortlist.

The most visible vendors have analyst coverage and customer evidence

Split every vendor into five bands by AI visibility, then compare what backs the top band against the bottom.

Share of the vendor's AI answers backed by…

Least visible fifth

Most visible fifth

Difference

Analyst coverage

5.6%

11.6%

2.1× more at the top

Customer evidence

16.3%

21.5%

1.3× more at the top

Two things are true here, and both matter.

  • Customer evidence is the bigger channel at both ends. More vendors have it, and it shows up in far more answers than analyst coverage at every level of visibility. If you're starting from nothing, this is the larger surface.
  • Analyst coverage is the sharper differentiator. It roughly doubles between the bottom band and the top, while customer evidence rises by a third. Analyst backing is what the most visible vendors have and the least visible ones don't.

Vendors in the top fifth have both. Programs that treat analyst relations and customer evidence as competing line items are arguing about a question the data has already answered.

Analyst influence is not a fixed ranking

To see how the analyst channel behaves, we read 68,299 AI answers in one category, digital workplace, across six platforms and ten topics. We counted how often each research and advisory firm was named in the text or cited as a source, April to August 2026. Thirty-one firms turned up. Nine carry almost all of it, and they're not moving in the same direction.

Change in answers naming or citing each firm, April to August 2026

Firm

Trend

Change

Gartner

Rising

+57%

Forrester

Falling

−14%

IDC

Flat

−4%

Futurum

Rising

+281%

IoT Analytics

Falling

−42%

Synergy Research

Rising

+659%

Constellation

Falling

−23%

Omdia

Rising

+152%

ISG

Flat

+2%

  • One firm is almost as big as the other eight combined. Gartner finishes at 143.6 mentions per thousand answers, up 57% in five months and three times Forrester. Three firms hold 74% of all analyst presence. The other quarter is where the movement is: the three fastest-growing firms all sit outside that group.
  • Below the top two, the order is being rewritten inside a single quarter. Gartner and Forrester hold first and second all five months. Every other firm changes rank at least once. Synergy Research Group, Futurum and Omdia all started from under 6 mentions per thousand. A tier list built on last year's rankings doesn't have them on it.
  • A firm's standing is not the same in any two markets. Forrester fell 52% in data security while rising 173% in IT services over the same twelve months.

There's no such thing as a firm's overall AI influence. There's only its influence in your category, this quarter. You can now watch that month by month, which means the return on an analyst relationship can be evidenced while there's still time to act on it.

Why citations undercount analysts

Most answers that name an analyst firm never link to it: two-thirds to four-fifths of them, depending on the market. The engine says "Gartner positions them as a leader," and the citation list points somewhere else entirely. Measured by citations alone, this channel looks like a rounding error. Measured by what the engines say, it's substantial and rising.

How analyst wins reach the engines

Analyst research sits behind a paywall, so the engines can't cite it. The conclusions travel anyway: written up in the trade press, posted on LinkedIn, quoted in analyst blogs and announced by every vendor named in the report. The engine knows who the analysts rated as leaders. What it doesn't have is the firm's own page to point at.

So amplification is a mechanism, not a victory lap, and there's a number for it. In digital workplace, Google is one of the vendors. Of every citation pointing at a Google URL, 17.5% goes to a page about what an analyst said. Analyst-validation pages are the best-performing content on Google's own domain. One press release about a Magic Quadrant placement pulled 641 citations on its own.

The announcement, the blog post, the LinkedIn write-up and the customer-facing summary are how a paywalled finding reaches a machine at all. An analyst win you don't activate is a win the models may never see.

That points at work you've already paid for. Every placement you've earned is an asset you can still activate: the mention in a report from eighteen months ago, the rating nobody wrote up, the quote that went into one deck and nowhere else. A placement announced once gives the engines one surface to reach it through. The same placement written up, summarized and quoted across owned and third-party channels gives them many, and those surfaces keep working long after the report goes back behind the paywall.

A prediction: the analyst firms that open more of their research will convert reputation into citation, and they'll do it while citations are getting scarcer.

Not all reviews do the same job

Review platforms need no argument made for them. In July 2026 they appeared in 27% of all data security answers and 21% of HR technology answers. The useful question is which platform, doing which job, in your market.

How AI uses each review platform

Platform

How AI uses it

What to optimize for

G2

Volume validation. Precise star ratings and ease-of-use pros and cons, quoted at high review counts ("4.8/5 across 14,000 reviews"). Cited most often, and most precisely.

Review volume and recency; ease-of-use and satisfaction Grid categories. The review count itself is leverage, because the engines cite the scale.

Gartner Peer Insights

Credentialing and shortlisting. Willingness-to-recommend percentages, "Customers' Choice," and structured sub-scores for integration, deployment and support.

Clearing the review-count threshold, since thin counts leave niche vendors invisible. Customers' Choice and the enterprise sub-scores.

TrustRadius

The justification engine. Balanced pros-and-cons synthesis plus verbatim reviewer quotes the engines lift directly into their reasoning.

Detailed, quotable, use-case-rich reviews. This is where an answer gets the language to explain why you fit.

PeerSpot

Numbers and mindshare. Supplies mindshare percentages and willingness-to-recommend data no other source has. Strong in enterprise, security and IT.

Presence and mindshare for enterprise and technical products; richer reviews, since it's aggregate-heavy and depth helps.

Capterra

SMB corroborator. Almost always co-cited to reinforce G2 or Gartner, and rarely drives the narrative alone. Skews small business and ease of use.

Presence and rating for corroboration, especially with SMB buyers. Not a standalone driver.

The platforms aren't interchangeable, and the difference isn't quality. It's function.

  • The horizontals get you onto the list. G2 and Gartner Peer Insights carry the breadth and volume the engines reach for when assembling a shortlist. If you're starting from nothing, start here.
  • The category specialist can outrank them in its own market. In IT services, Clutch is the single most-cited domain in the category, ahead of every publication, analyst firm and vendor site, at twelve times the volume of the next review platform.
  • The depth platforms shape how the answer reads. TrustRadius and PeerSpot sit near the bottom on raw volume (0.04% to 0.21% of citations), but their content is long-form and specific about deployment and support. When an answer explains why a vendor suits a particular buyer, the language often comes from here.
  • Capterra corroborates. It rarely leads and frequently confirms.

The leading review platform is different in every market

G2 leads in HR technology and digital commerce. Gartner Peer Insights leads in data security. Clutch leads IT services. The leading platform changed in two of these four markets in a single year.

That turns review generation from a volume exercise into a targeting one. Between the role table and the market rankings, you can say which platform does which job and which one leads in your category. That's the basis for a precision advocacy program: the right customer, on the platform that does the job you need done. Breadth reaches the shortlist and depth justifies the choice, so you stop asking every willing reference for a review everywhere.

Part 3: The Third Voice

The independent experts your buyers already read

Analysts and review platforms are two of the three voices this report keeps naming. The third is a person rather than an institution: the independent expert, the practitioner, the creator with outsized standing in a market. Your buyers were already reading these people, and now the machines are too.

They don't sit in one place. The same individual can be the byline in a trade publication that turns out to be a one-person shop, the practitioner answering a question in a subreddit, and the outside voice in a video a vendor co-produced.

In ITSM (100 buyer prompts across eight AI platforms), each channel does a different job. Some shape what a buyer thinks the category is, others decide who makes the list, and others supply the reason one vendor beats another.

Channel presence by buying stage, ITSM (1.00 = each channel's own average)

Buying stage

Trade press

Reddit

YouTube

LinkedIn

Education: what is this category

1.12

1.23

0.68

0.56

Shortlisting: which tool should we buy

1.20

1.24

0.97

0.86

Comparison: this one or that one

1.27

1.38

1.28

0.36

Cost: what will it run us

1.31

0.96

0.68

1.38

Lived experience: what was it actually like

n/a

1.28

0.96

1.03

Implementation: how do we run it

1.05

0.62

0.49

0.55

No channel carries a buyer the whole way, and two of them are near mirror images. Reddit is strongest on comparison and weakest on implementation. LinkedIn is the reverse: almost absent when a buyer is comparing, and the strongest voice in the set when they ask what it costs.

So activate independent voices on purpose, channel by channel, against the narrative you're trying to land or counter. A roster that treats every expert and every channel as interchangeable is spending against an average that describes none of these stages.

Third-party presence beats publishing

Part 2 showed the ceiling on your own content. This is the other side of it. Third-party content doesn't write itself. The people described above produce it, and being present in their work is measurable at the vendor level.

Own publishing against third-party presence, two ITSM vendors


InvGate

ManageEngine

Content on its own site

24,560

18,259

Presence in third-party content

15

1,239

Named in AI answers

7.7%

23.2%

InvGate publishes more on its own site than ManageEngine does, and is named in a third as many AI answers. The difference isn't output. ManageEngine appears in other people's content roughly eighty times more often. Across all 23 vendors we measured, presence in third-party content predicts AI visibility better than a vendor's own publishing does.

An influencer program is how a vendor gets into that content. These are the people writing it, and how you work with them decides where you get named.

Working with independent voices: what to do, and why

Recommendation

Why

Don't select influencers on follower count alone

On one vendor's own influencer list, the person with the ninth-largest following is picked up by AI nearly eight times more than the other eleven combined.

Point influencers toward comparisons, not think-pieces

Thought leadership gets read, but lands in answers that name no vendors. "Which tool" and "how to choose" pieces put a vendor in front of a buyer.

Get influencers bylined in trade press

Trade press is the single largest thing AI reads in this market, bigger than Reddit, YouTube and LinkedIn put together.

Ask for LinkedIn Pulse articles, not just feed posts

Pulse articles are picked up roughly eight times more often than timeline posts.

Publish podcast transcripts, or put the episode on video

Podcast audio is almost never picked up. The same conversation on video, where transcripts are generated automatically, appears hundreds of times more often.

Support the publications influencers own

One independent expert's own articles are picked up by AI six times more often than Forrester, IDC and Everest Group's research combined.

Place your own experts in independent publications, including categories you don't yet lead

A competitor's employees published on IT operations integration, a category they aren't known for. On that exact buyer question, their mention rate more than doubled, from 7% to 16%.

None of that requires a bigger program, only a more precise one: the same budget, aimed at specific people, specific formats and the surfaces where buyers are asking.

Getting experts and customers to write about you somewhere an engine can retrieve it is half the problem. What they say when they do is the other half.

Part 4: What They Say About You

Winning the mention, losing the description

Everything so far has been about being named. This part is about being described, and those aren't the same achievement.

Workday posted the largest visibility gain in HR technology: 29.6% to 42.6% of all answers in twelve months. Same questions, same six engines, July to July. By the measure most of this market uses, that's the year's best result in the category. It didn't come from the question everyone competes on.

Workday visibility by topic

Topic

2025

2026

Change

Employee engagement

16.3%

44.4%

+28.1pp

Workforce analytics

50.4%

77.2%

+26.8pp

Remote workforce management

3.7%

15.9%

+12.2pp

Time and attendance

1.2%

11.7%

+10.4pp

HR software solutions

45.2%

47.2%

+1.9pp

The gains are all in specific use cases. On the generic category question, "HR software solutions," Workday barely moved. It's winning the buyer who already knows what they need to do, not the one still asking what to buy.

The company it keeps changed too. In answers that name Workday, employee engagement specialist Culture Amp now appears alongside it 13.0 points more often than a year ago. BambooHR, its long-standing core HR rival, appears 10.8 points less.

Now look at what those same answers say about Workday.

Share of sentiment that is positive, by theme (point-in-time, May 15, 2026)

  • Compliance: 99.0%
  • Payroll: 98.8%
  • Analytics: 95.8%
  • Scalability: 82.3%
  • Pricing: 11.8%
  • Deployment: 4.1%

Workday has the biggest visibility gain in its category while it loses on two narratives. A buyer asking which HR platform to shortlist now meets Workday more often than anyone. A buyer asking what it's like to deploy one meets a description that's four percent positive. Same engine, same market, same week.

That's why visibility on its own isn't influence. You can win the mention and still lose the description, and more visibility doesn't fix it. It puts a worse description in front of more people.

The fix is narrower and more useful than a bigger program. Two attributes are named, and that's the brief. Find the sources producing those two descriptions and work on them, rather than on the number that already looks like a win. Which sources those are depends on the market you sell into.

Part 5: No Two Markets Alike

We profiled four B2B markets (HR technology, data security, digital commerce, and IT services and consulting) to see how much the trust layer varies from one to the next, and what that does to the shortlists the engines build. For each: the vendors the engines name, the domains they draw on, the review platforms they cite and the analyst firms they mention. Every list is a top five, both years, with what moved.

The names will be different in your market. The pattern won't.

  • No two trust layers are shaped alike. Clutch is the most-cited source in IT services and doesn't appear in the top five of HR technology or data security. The analyst firms in third, fourth and fifth place barely overlap between any two markets. A program designed for one of these markets would be misallocated in the other three.
  • A year moves more than most planning assumes. The leading review platform changed in two of the four markets. Video entered the top five in every one. Wikipedia left the top five in three of the four and fell out of the top fifty in all of them.
  • The vendors move with it. In three of the four markets, the leading vendor added ten points or more of presence while the sources behind those answers were reshuffled underneath them. Vendors that had held top-five positions for years dropped out entirely.

A note on units: they differ by column and don't compare across columns. Vendors are a share of answers naming the vendor. Domains and review platforms are a share of all citations. Analyst firms are a share of answers naming the firm in the response text.

HR Technology

20,910 answers and 214,823 citations (2025) → 21,018 answers and 165,549 citations (2026)

Rank

Vendors

Top cited domains

Review platforms

Analyst firms named

1

Workday 42.56% (was #3, 29.63%)

youtube.com 4.23% (new)

G2 1.85% (was #2, 0.66%)

Gartner 1.35% (was #1, 0.36%)

2

Rippling 40.60% (was #2, 32.71%)

peoplemanagingpeople.com 3.03% (was #1, 7.42%)

Gartner Peer Insights 0.92% (was #1, 0.75%)

Forrester 0.09% (was #2, 0.13%)

3

BambooHR 35.08% (was #1, 35.53%)

worldmetrics.org 2.11% (new)

Capterra 0.49% (was #3, 0.31%)

ISG 0.06% (new)

4

Gusto 26.02% (was #4, 25.21%)

g2.com 1.85% (new)

GetApp 0.48% (was #5, 0.16%)

Info-Tech 0.05% (was #3, 0.06%)

5

ADP 24.66% (was #5, 22.57%)

reddit.com 1.79% (was #3, 1.86%)

Software Advice 0.25% (was #4, 0.28%)

Brandon Hall 0.05% (was #5, 0.01%)

Dropped out: forbes.com (#2) and selectsoftwarereviews.com (#4) from top domains; Nucleus Research (#4) from analyst firms.

Data Security & Privacy

18,837 answers and 205,306 citations (2025) → 18,596 answers and 153,748 citations (2026)

Rank

Vendors

Top cited domains

Review platforms

Analyst firms named

1

Microsoft 60.86% (was #1, 50.11%)

sentinelone.com 2.75% (was #1, 3.61%)

Gartner Peer Insights 2.35% (was #1, 1.98%)

Gartner 5.58% (was #1, 4.59%)

2

CrowdStrike 34.54% (was #2, 28.84%)

worldmetrics.org 2.53% (new)

G2 1.76% (was #2, 1.01%)

Forrester 1.00% (was #2, 2.08%)

3

Palo Alto Networks 31.04% (was #4, 24.05%)

gartner.com 2.51% (was #4, 2.03%)

GetApp 0.36% (new)

IDC 0.13% (was #3, 0.20%)

4

SentinelOne 19.96% (was #5, 20.63%)

youtube.com 1.81% (new)

TrustRadius 0.21% (was #4, 0.10%)

Info-Tech 0.04% (was #4, 0.04%)

5

Okta 17.68% (new)

g2.com 1.76% (new)

PeerSpot 0.20% (was #3, 0.25%)

Frost & Sullivan 0.02% (was #5, 0.04%)

Dropped out: IBM (#3) from vendors; microsoft.com (#2), expertinsights.com (#3) and wikipedia.org (#5) from top domains; Capterra (#5) from review platforms.

Digital Commerce

18,852 answers and 199,132 citations (2025) → 18,600 answers and 150,988 citations (2026)

Rank

Vendors

Top cited domains

Review platforms

Analyst firms named

1

Shopify 52.34% (was #1, 36.90%)

youtube.com 2.11% (new)

G2 0.65% (was #2, 0.32%)

Gartner 1.50% (was #1, 0.52%)

2

BigCommerce 26.63% (was #2, 27.57%)

shopify.com 1.71% (was #1, 4.13%)

Clutch 0.55% (was #3, 0.20%)

Forrester 0.32% (was #3, 0.14%)

3

Amazon 23.29% (was #3, 21.55%)

bigcommerce.com 1.70% (was #4, 1.05%)

Gartner Peer Insights 0.42% (was #1, 0.33%)

IDC 0.04% (was #4, 0.03%)

4

Adobe 13.67% (was #4, 16.18%)

worldmetrics.org 1.31% (new)

TrustRadius 0.04% (was #5, 0.04%)

IoT Analytics 0.04% (was #2, 0.22%)

5

WooCommerce 12.80% (was #5, 14.84%)

linkedin.com 1.24% (new)

GetApp 0.04% (was #4, 0.05%)

Info-Tech 0.02% (new)

Dropped out: wikipedia.org (#2), reddit.com (#3) and forbes.com (#5) from top domains; Frost & Sullivan (#5) from analyst firms.

IT Services & Consulting

13,197 answers and 132,367 citations (2025) → 13,019 answers and 98,888 citations (2026)

Rank

Vendors

Top cited domains

Review platforms

Analyst firms named

1

Accenture 35.88% (was #1, 30.20%)

clutch.co 2.65% (new)

Clutch 2.65% (was #1, 0.73%)

Gartner 2.91% (was #1, 1.01%)

2

Deloitte 31.63% (was #3, 26.67%)

linkedin.com 2.17% (was #1, 2.01%)

G2 0.21% (was #2, 0.23%)

Forrester 1.08% (was #2, 0.40%)

3

IBM Consulting 27.51% (was #2, 28.45%)

youtube.com 2.01% (new)

Gartner Peer Insights 0.19% (was #3, 0.14%)

IDC 0.63% (was #3, 0.28%)

4

Capgemini 18.54% (was #4, 17.29%)

corsicatech.com 0.91% (new)

TrustRadius 0.07% (was #4, 0.03%)

Everest Group 0.22% (was #5, 0.02%)

5

EY 16.54% (new)

worldmetrics.org 0.79% (new)

Capterra 0.01% (new)

ISG 0.08% (new)

Dropped out: McKinsey (#5) from vendors; managementconsulted.com (#2), wikipedia.org (#3) and reddit.com (#4) from top domains; PeerSpot (#5) from review platforms; IoT Analytics (#4) from analyst firms.

Part 6: What To Do

Start with the shape of your own market

Every category has a trust layer, and no two are shaped alike. Some run on peer review, some on analyst authority, and some on trade press and community with almost no analyst signal at all. Find out which one you sell into before you spend another quarter against it.

The mandate, by function

Function

What to do

What it gets you

CMO

Stop funding output. Fund the evidence you can't manufacture, and read each engine separately, because a composite hides the variation.

A defensible answer to "why does the AI recommend them and not us?", and a reason to move budget that survives the CFO.

Analyst Relations

Measure what the engines say, not only what they link. Then activate every analyst win on a surface the models can reach.

The budget argument you've never had evidence for, and a use for placements you already own.

Customer Marketing

Find which platform leads your category before funding review volume. Breadth to reach the shortlist, depth to justify the choice.

A specific platform to fund and a specific reason, and a move from reference desk to shaping the answer.

Influencer Relations

Build the roster from what the engines cite, not from follower counts. Reach describes an audience; citations describe influence.

The first quantified evidence that your channels move AI answers, and a roster you can defend.

Product Marketing

Audit the narrative, not the mention. Find the two attributes where you're described worst and treat those as the brief.

You stop guessing which objections are killing deals, and learn where each one came from.

Content & SEO

Stop competing on output. Publish to be cited as a source, and know that being the source isn't being the choice.

An exit from a volume race you can't win, and budget redirected to work that compounds.

Communications / PR

Change what goes into trade press, not how much. Put independent experts and analyst findings into the placements you already earn.

Trade press is still the largest single source the engines read, and the largest decline in this report. The channel isn't the problem; what fills it is.

The window is open now

Ten sources became eight. There are fewer slots and far more questions, and the sources that kept their place were the ones with a person behind them: analysts, customers and the practitioners who already work in your category.

The slots that opened went to content produced at a scale no marketing team can match, and there's no point racing it. But the layer that gained anyway is the one a better-funded competitor can't simply buy more of. Most of what it takes, you already have.

The analyst briefings, the reference customers and the experts who know your product are already sitting in your organization. That work is spread across five or six budgets and measured five or six different ways: briefing counts, review volume, follower reach. Every channel keeps its own books, and none of the numbers convert.

The answer layer is the first place all of them show up together, in the same answers, in front of the same buyer, where you can see what each one is worth. That's the measure that's been missing. It's not how many analysts you briefed or how many reviews you collected. It's which voices your market runs on and what they say about you when the shortlist forms. The honest answer to "what did it get us?" has always been a story rather than a number. Now it can be a number.

So read every layer, not just the one you own, and put one person in charge of whether all of it tells the same story. There are fewer slots than there were a year ago, and no extra credit for arriving late. Find out what the engines already say about you, then decide what it's worth to change it.

No single voice moves a market. Perception is assembled, and decisions are made, from every voice your buyer already trusts. It moves when your story holds up in all of them.

Methodology

Prompt volume. The 33 million figure is from Profound, an AI visibility platform that measures prompt activity and answer content across the major LLM assistants. It represents daily B2B buying-related prompts on ChatGPT: questions about vendor selection, product comparison, category evaluation and purchase decisions. The prior-year figure of 20 million comes from the same source on the same basis, so the year-over-year change reflects growth in query volume rather than a change in method. Prompt volume at this level isn't published by the platforms and exists only through third-party instrumentation, so read it as a measure of scale and direction. It's a separate dataset from the citation panels and shares no sample with them.

The cross-platform estimate. The 65 to 100 million range extends the measured ChatGPT volume across the rest of the market using ChatGPT's share of assistant usage. Third-party estimates of that share run from 54% (Similarweb, May 2026) to 77% (StatCounter, June 2026). We adjust downward from the low end because web-traffic measurement structurally undercounts enterprise-embedded assistants. Microsoft Copilot's roughly 420 million monthly users sit largely inside Microsoft 365, and Gemini's inside Workspace, which are the surfaces where B2B buying research is most likely to happen. The range assumes ChatGPT accounts for between one third and one half of all B2B buying prompts. A higher multiple is defensible; we report the floor.

The year-over-year panel. July 2025 against July 2026, across four B2B markets: 143,029 AI answers and the 1.32 million citations behind them. Restricted to the six engines present in both windows: ChatGPT, Google AI Mode, Google AI Overviews, Google Gemini, Microsoft Copilot and Perplexity. The same questions were asked in both windows, so what moved is the market, not the measurement.

Other datasets. The analyst-firm trajectories cover digital workplace, April to August 2026: 68,299 answers across six platforms and ten topics, with Gartner Peer Insights excluded throughout. The buying-stage and independent-expert analysis covers ITSM, November 2025 to August 2026: 197,944 answers across eight AI platforms, drawing on 199,423 distinct cited URLs. Each is a self-contained window, and figures aren't comparable across them.

Units differ by measure. Vendor figures are a share of answers naming the vendor. Domain and review-platform figures are a share of all citations. Analyst-firm figures are a share of answers naming the firm in response text, so they're text-derived, while everything else is citation-derived.

About Spotlight and Profound

About Spotlight. Spotlight orchestrates influence across every credible market voice (analysts, customers, independent experts and AI systems) to put clients on the shortlist and move buyers to buy. We work with more than 190 B2B organizations across software, tech and services. spotlight-io.com

About Profound. Profound is the full-stack marketing platform for the marketer of the future: understand, analyze, build and measure how your brand appears in AI search. Coverage spans ChatGPT, Google AI Overviews, Gemini, Copilot, Perplexity, Claude, Grok and DeepSeek. tryprofound.com