The Scoreboard Nobody Expected
HubSpot and Clay didn't lose because they're bad products. HubSpot has spent years building a genuinely useful operating system for marketing and sales. Clay built one of the sharpest workflow layers for enrichment and personalization that B2B has seen in years.
They lost a popularity contest to a general-purpose chatbot with no CRM integration, no native sequencing, no proprietary data model, and no GTM-specific training.
That's the part worth sitting with. In a market obsessed with specialized, agentic, purpose-built AI, the tool GTM leaders say moves the needle most is the one with no GTM specialization at all.
The pattern shows up everywhere else in the data too. According to Responsive's Buyer Intelligence 2025 study, generative AI has overtaken traditional search for roughly a quarter of B2B buyers, and nearly two-thirds now use GenAI as much as or more than search when researching vendors. G2's Answer Economy report goes further: 51% of B2B software buyers now start their purchasing process in an AI chatbot rather than a search engine, and 69% ended up choosing a different vendor than they originally planned, based on what the chatbot told them.
Buyers are already living inside plain-vanilla AI tools to make decisions. Sellers are still shopping for specialized ones to influence those decisions. That mismatch is the real story.
This Was Never a ChatGPT Story
Here's the mistake I see founders make when they read a stat like this: they treat it as an endorsement. "ChatGPT wins, so let's build our GTM around ChatGPT."
Wrong read. ChatGPT winning isn't proof that ChatGPT is the best tool. It's proof that tool sophistication was never the bottleneck.
Think about what ChatGPT actually does well in a GTM context. It drafts emails, rewrites messaging, summarizes calls, generates talking points, helps reps prep for meetings. It doesn't know your ICP. It doesn't know your competitive wedge. It doesn't know why a customer should pick you over the three other vendors in the shortlist.
It only knows what you tell it. And that's exactly the point.
When a GTM leader says ChatGPT is their most impactful tool, what they're actually saying is: "The thing that helped me most this year was having a fast, flexible way to articulate what we do." That's not a tooling win. That's a clarity exercise wearing a product interface.
I've watched this play out inside startups directly. A team with a sharp, differentiated one-sentence pitch uses ChatGPT to generate ten variations of outbound copy, all of them on-message, all of them converting at a reasonable clip. A team with a muddy pitch uses the exact same tool and generates ten variations of generic, interchangeable copy, faster than they could before. Same tool. Opposite outcome.
AI doesn't fix a confused go-to-market. It just executes the confusion faster, and now it does it in your competitor's voice too, because they're prompting the same model with the same vague positioning you are.
The Gartner Trap: Horsepower Without a Steering Wheel
This is where the 2026 budget conversation gets dangerous.
According to Gartner's CMO Spend Survey, 70% of CMOs say becoming an AI leader is a core goal for 2026. Only 30% believe their organization actually has the infrastructure to pull it off. AI now absorbs 15.3% of marketing budgets on average, rising to 21.3% at better-equipped organizations, while total marketing budgets stay flat at 7.8% of revenue.
Read those two numbers together. Marketing budgets aren't growing. AI's share of that flat budget is growing fast. Something else is getting cut to fund it, and in most startups I work with, that something is the strategic work: positioning, messaging discipline, ICP refinement, the unglamorous foundation that determines whether any of the AI spend produces anything.
Marketing Week's State of B2B Marketing 2025 research adds a second layer to this. Most B2B marketers report a greater strategic role in the business than a year ago, but investment doesn't reliably follow, and over half of B2B brands still fail to understand marketing's potential beyond lead generation.
Put it together and you get the Gartner trap: a majority of CMOs chasing AI leadership as a mandate, a minority with the foundation to execute it, and a budget environment that keeps buying horsepower while starving the steering wheel. You end up with a very fast car pointed in a direction nobody chose on purpose.
What This Means in Tel Aviv and Herzliya Boardrooms
Israeli B2B startups are walking into 2026 planning with the same instinct as everyone else: the AI gap feels like a tooling gap, so the fix feels like a tooling purchase.
I've sat in budget reviews where the proposed line item is a new "AI-powered GTM suite" promising agentic outbound, auto-generated account plans, and predictive lead scoring, priced at a premium specifically because it's AI-native. Meanwhile the actual positioning deck the sales team uses still changes every quarter, and the one-sentence pitch gets answered three different ways by three different AMs in the same week.
That's not a tooling problem. Buying a more sophisticated AI layer on top of unclear positioning doesn't close the gap. It widens the blast radius. You're now generating confused messaging at volume, with AI confidently producing variations of something that was never clear to begin with.
The founders who get real lift from AI this year aren't the ones with the most expensive stack. They're the ones who did the unglamorous work first: a tight ICP, a specific point of view on the problem, a one-sentence answer to "what do you do" that doesn't need a follow-up question. Then they feed that clarity into whatever tool is on hand, often something as plain as ChatGPT, and the output compounds instead of scatters.
Before adding AI tooling spend, ask: can every rep, every piece of content, and every landing page answer "what do you do and why does it matter" the same way?
If the answer changes by team or by quarter, the AI layer will amplify that inconsistency at scale, not fix it.
Buyers are already researching you inside AI chatbots. If your positioning isn't sharp enough for a human to repeat, it's not sharp enough for a model to represent you accurately either.
Fix the Sentence Before You Buy the Software
There's a simple test I use with every startup I advise before any AI tooling conversation happens: can the CEO, the VP Sales, and a brand-new AE all describe what the company does, who it's for, and why it wins, in one sentence each, and do those three sentences roughly agree?
If they don't agree, no tool fixes that. Not ChatGPT, not a $150,000 agentic platform, not a six-month implementation of the latest intent-data suite. You'll just get three AI-assisted versions of a disagreement, produced faster and distributed wider.
If they do agree, something interesting happens. Plain tools start performing like sophisticated ones, because the thing doing the heavy lifting was never the software. It was the clarity you fed it. That's the actual lesson behind ChatGPT's win over HubSpot and Clay: in a market flooded with specialized AI, the differentiator isn't which tool you use. It's whether you've done the strategic work the tool is supposed to amplify.
Positioning before performance marketing. Clarity before velocity. The order hasn't changed just because the tools got smarter.
Is your 2026 AI budget about to amplify clarity or confusion?
If you're weighing AI tooling spend for next year, let's first confirm the positioning that tooling is supposed to scale. A short conversation can save you a six-figure mistake.
