ZoomInfo's "Agentic Confusion" Admission: What It Means for Your GTM AI Buying Decisions

A GTM data vendor just blamed its own numbers on AI confusion

I'm Hlib Storchak. I build and run outbound systems for B2B founders and sales teams, and I've booked 2000+ meetings for B2B clients doing it. Most weeks, at least one client asks me some version of the same question: which AI SDR or agentic GTM tool should we actually buy. My honest answer has always been some version of "it depends, and check the vendor's claims yourself before you trust them." What I did not expect was a public GTM data company telling its own shareholders that the market has gotten too confusing to sell into, on the record, on an earnings call.

That is what happened. In May 2026, ZoomInfo lowered its full-year revenue guidance and told analysts the drag was not a bigger competitor stealing share or a product problem. It was confusion. Its own customers could not cleanly tell what to build internally, what to buy from a vendor, and where the real differentiation actually sat. Three months later, on the next earnings call, the same executives said the confusion had not gone away.

What ZoomInfo actually said, on the record, twice

On ZoomInfo's Q1 2026 earnings call (May 11, 2026), CEO Henry Schuck described the issue as customer uncertainty over "what can be built versus bought, what vendor or internal team delivers what, and where the differentiation really lives," adding that "this led to a pause in purchasing decisions, and our software customers were particularly affected" (The Motley Fool, ZoomInfo Q1 2026 earnings call transcript). CFO Graham O'Brien echoed it, pointing to "more customer confusion in the marketplace around what AI can and cannot do." The company cut full-year 2026 revenue guidance to a range implying roughly a 4% decline at the midpoint, splitting the blame between downmarket restructuring, caution from software-vertical customers, and pricing-model transition effects, with the AI confusion sitting squarely inside that software-caution piece.

On the Q2 2026 call three months later (August 5, 2026), Schuck did not walk it back. Asked about the competitive environment, he said "there are a lot of people claiming to be able to do a lot of things in go to market that created some of the confusion that we saw in Q1," then added plainly, "and we're still seeing that sort of confusion in Q2" (ZoomInfo Q2 2026 earnings call transcript, via Yahoo Finance). To be fair to the numbers, Q2 revenue actually rose slightly year over year and beat analyst expectations, and shares moved up after hours. Schuck also called the confusion "temporary." But the specific claim I care about here, that buyers still cannot tell what they are looking at, held steady across two straight quarters. Nobody said it was fixed.

What "agentic confusion" actually means once you unpack it

Strip the earnings-call language down and "agentic confusion" is three separate, more concrete problems stacked on top of each other:

1. Build versus buy is unclear. A modern coding agent can now stand up a passable internal enrichment or outreach tool in an afternoon, which makes a buyer legitimately unsure whether a vendor's product is worth paying for at all.
2. Vendor versus internal team is unclear. When a data vendor, a sequencing platform, and an in-house GTM engineer all claim to own "the agent layer," a buyer cannot tell who is actually responsible for the outcome if it breaks.
3. Real differentiation is unclear. Every vendor in this category currently uses nearly identical language, agentic, autonomous, AI-powered, which means the words on the page have stopped doing the job of telling a buyer what is actually different.

None of that is a ZoomInfo-specific problem. It is a category-wide symptom, and ZoomInfo is simply the vendor that said so out loud, with numbers attached.

Why this is bigger than one vendor's earnings call

Public companies do not blame quarterly misses on vague vibes. Guidance cuts get scrutinized by analysts who ask pointed follow-up questions, and a CEO who says "confusion paused purchasing decisions" twice in a row, in two separate quarters, is describing something his own sales team is watching happen deal by deal, not a hunch. That is a useful, rare data point, because most of what gets published about agentic AI adoption comes from the sell side: vendor surveys, vendor-commissioned research, vendor blog posts about how fast their category is growing. An admission of friction, from the demand side, said against the company's own short-term interest, deserves more weight than another adoption-curve chart.

It also matches what I actually see running outbound and vetting tools for clients this year. Sales cycles for AI SDR and agentic GTM tools have stretched, not because buyers are less interested, but because they are doing more diligence before signing, and finding that diligence harder than it should be.

The buyer's side: a slow purchase decision is the rational response

If every vendor pitch uses the same four adjectives, autonomous, agentic, AI-powered, end-to-end, and none of them will tell you plainly what happens when the agent gets something wrong, slowing down is not indecision. It is the correct response to a genuinely confusing market. The mistake I see clients make is treating that hesitation as a personal failing, something they should push through faster, instead of treating it as a signal that the vendor has not yet earned a fast yes.

A faster, better path is to stop evaluating vendors on their positioning language entirely and evaluate them on a short list of specific, checkable claims instead. That is the point of the framework in the next section.

The seller's side: vague positioning made this worse

The other half of this is that vendors did this to themselves. When an incumbent renames its whole product line around "agentic AI" without changing what the product concretely does, and three competitors do the same thing in the same quarter, buyers lose the ability to tell a real capability shift from a marketing refresh. I have written before about Outreach's own agentic rebrand and what to check before trusting the label on any vendor doing the same thing. ZoomInfo's earnings call is the first time I have seen a vendor admit, with a stock price attached, that this category-wide vagueness is now actively costing revenue rather than generating it. That should change how sellers write their own positioning: narrower, provable claims are starting to close faster than broad ones, not the other way around.

Six questions that cut through agentic confusion

I use a version of this list with clients before they sign anything with an AI SDR or agentic GTM vendor. None of it requires technical expertise, and a vendor who cannot answer these plainly is telling you something important on its own.

What I tell clients before they sign anything

The mistake I see most often when I sit in on a client's vendor evaluation call is letting the vendor set the vocabulary. If the vendor says "agentic," ask them to replace that word with a sentence describing exactly what happens, in order, from trigger to output. If they cannot do it in one sentence without another adjective, that is real information, not a communication gap on your end. I have watched this single substitution, adjective for verb, cut a two-month evaluation down to two weeks more than once, simply because it forces the vendor to either answer plainly or reveal that there is not much concrete behind the language.

What changes if you're the one pitching AI-built GTM tooling internally

If you are the one trying to get budget approved for an internally-built agent or a vendor purchase, ZoomInfo's numbers cut both ways for you too. Broad claims ("this will transform our whole outbound motion") now read as a yellow flag to a finance leader who has read the same confused vendor pitches you have. A narrow, specific claim ("this replaces four hours a week of manual list-building for two reps, here is the before and after") is easier to approve precisely because it is checkable. Pitch the smallest true claim you can defend, not the biggest one you can imagine.

What doesn't change: the fundamentals still decide the deal

None of this changes what actually wins outbound. A confused AI market does not fix a bad ICP, a weak offer, or a list nobody wants to hear from, and it will not book you a single meeting on its own. Every AI SDR or agentic tool still has to plug into a working outbound system to be worth anything. Vet the tool with the six questions above, then judge it the same way you would judge any hire or vendor: on whether it moves your actual pipeline numbers, not on the language it uses to describe itself.

FAQ