B2B buyers complete 70% of research before any sales contact

Operators and investors,

Gartner projects that AI agents will intermediate more than $15 trillion in B2B purchases by 2028, handling the research, evaluation, and in a growing share of cases the transaction itself (Gartner). The same research predicts buyer-side agents will outnumber human sellers 10 to 1 by that date, while fewer than 40% of sellers will report any productivity gain from their own AI tooling.

I flagged the same shift last year: 83% of B2B buyers now initiate the first contact themselves, and by the time they do they are already 70% through the buying process, past awareness and consideration.

The seller no longer opens the conversation; the buyer does, after the shortlist is set.

The projection lands on a buying process that had already moved out of reach. Buyer research through the 2020s consistently put the share of the journey completed before a supplier conversation at 70% or more, with Gartner's own studies showing buyers spending 17% or less of their total buying time with any sales rep. Forrester's 2026 State of Business Buying, drawn from nearly 18,000 buyers, now finds 94% using generative AI inside the purchasing process. The human research phase that sales coverage models were built to intercept is being delegated to software.

Evaluation is happening before contact

I see the consequence in mid-market pipeline reviews every quarter. Inbound leads arrive with requirements already set, a shortlist already built, and pricing expectations already anchored, and the sales team's discovery motion runs against a buyer who finished discovering weeks earlier.

When the majority of evaluation happens before contact, a GTM model that spends its budget on outbound touches and its calendar on early-stage discovery is buying coverage of a stage that no longer decides the outcome.

This issue covers 4 things:

• Where the buying journey actually moved

• What agentic procurement changes mechanically

• The rebuild: making the company legible to buyers and their agents

• What this does to coverage models and the sales P&L

Every number on this chart moved in the same direction over the past decade and none of them are moving back.

The way I put it to portfolio CROs: your sales team's calendar is built for the 30% of the journey you can see, and your budget should be built for the 70% you cannot. The companies that accept that arithmetic redeploy spend from coverage to legibility, and the rest keep buying meetings with buyers who already decided.

1. Where the buying journey actually moved

The pre-contact share of the journey has been climbing for a decade, and the instruments changed twice.

• The first shift moved research from analyst calls and trade shows to search, review platforms, and peer communities, which is where the 70% figure took hold.

• The second shift is happening now: the buyer delegates that same research to an AI assistant, which reads more sources faster and returns a synthesized shortlist. Forrester's 94% adoption number describes a research process that is already machine-mediated at nearly every buying organization.

The economics on the buyer side explain the speed. A procurement team that once staffed vendor evaluation with analyst hours now runs the first 3 stages, requirements, market scan, and shortlist, in a fraction of the time.

The evaluation got cheaper, so more of it happens, and it happens earlier.

The mid-market seller experiences this as compressed deal cycles at the end and silence at the beginning: fewer early conversations, later first contact, and a buyer who arrives with 85% of the requirements set, a figure Gartner's buying research has tracked for years in the same direction.

What breaks first inside the seller's organization is attribution and forecasting. When the decisive stages run outside instrumented channels, the CRM records a late-stage inbound as the whole story, marketing takes credit or blame for a journey it never saw, and the forecast inherits a pipeline whose early signals no longer exist.

The entire journey of the sales funnel flags buyers in different stages, with effective SDR teams failing to reach prospects early on, and strong marketing orgs closing with whatever sales power is available since the deal is mostly sealed.

2. What agentic procurement changes mechanically

An AI agent evaluating vendors reads differently than a human does.

It parses structured data, machine-readable pricing, documentation, review corpora, security and compliance pages, and the technical surface of the vendor's own site.

It does not attend webinars, feel brand affinity, or forgive a broken pricing page.

Gartner's projection that a quarter of enterprise software purchases will run agent-to-agent with no human in the loop by 2028 describes a procurement channel where the vendor's first impression is rendered entirely by its data quality.

That changes what counts as sales infrastructure. In our web and RevOps work across mid-market portfolios, the assets that increasingly decide shortlist inclusion are structured data markup, published pricing and packaging, integration documentation, verifiable review volume, and page performance, the same assets our 1,104-company crawl this year found chronically underinvested across PE-backed websites.

A company whose capabilities live in PDFs behind forms, whose pricing requires a call, and whose documentation sits behind a login is invisible at the exact stage where the decision now forms. The cost of that invisibility never appears in a budget line; it appears as pipeline that never arrives.

The seller-side agent story is weaker than the vendor pitch suggests. The sub-40% productivity figure in Gartner's prediction reflects what I see in portfolio sales teams: AI tooling layered onto an unchanged process produces drafted emails faster while the constraint, access to a buyer who no longer takes early meetings, stays untouched.

Melissa Hilbert, VP Analyst in Gartner's sales practice, put the ceiling plainly: "Beyond a certain point, more AI does not mean more productivity. In fact, layering additional prompts and tools onto already complex workflows risks overwhelming sellers and accelerating burnout." Tooling spend follows the vendor narrative while the binding constraint sits in the buyer's process, which no seller-side license can purchase back.

The 69% validation figure is the one that keeps this from being a story about sales headcount going to zero. Buyers delegate the research and still want a human to confirm the conclusion before they sign, which redefines the seller's job as validation and de-risking at the end of the journey. In our RevOps engagements that means fewer sellers, more senior ones, and a public data surface strong enough that the agent's summary of you is accurate, because the rep now inherits whatever the machine concluded.

3. The rebuild: making the company legible to buyers & their agents

The response that works treats the company's public surface as the first sales call.

That starts with the unglamorous technical layer: structured data on every product and service page, pricing published or at least framed, documentation opened, security and compliance answers posted where an agent can read them, and page weight low enough that automated evaluation completes.

This work is measurable, it is cheap relative to a single AE's annual cost, and in our engagements it is owned jointly by RevOps and the web team, with marketing supplying the substance.

The second layer is presence in the corpora agents actually consult: review platforms with volume and recency, comparison content that answers the questions buyers ask their assistants, directories that categorize the market, and category pages that state plainly what the product does, for whom, at what scale.

We operate several such directories ourselves, topanalyticstools.com and aiforbusinesses.com among them, precisely because buyers and their agents consult this layer before any vendor conversation happens. The shift in content economics is that a page now gets read hundreds of times by machines for every human visit, so content built for skim-reading executives underperforms content built to be quoted accurately by a model.

The third layer is instrumentation for a submerged funnel. Intent data, deanonymization, and review-platform signals partially reconstruct the invisible early stages, and the companies that invest there regain some forecasting signal. The honest caveat: reconstruction is partial, and the better use of the budget is usually making the public surface strong enough that the shortlist forms in your favor without needing to watch it happen.

4. What this does to coverage models and the sales P&L

The headcount math changes when the first human conversation moves to the final 30% of the journey. The classic mid-market model, a wide SDR layer generating early conversations and an AE layer running long discovery cycles, is built for a journey that no longer exists at most buyers. What replaces it is smaller and later: fewer, more senior sellers who enter at evaluation stage with commercial and technical depth, supported by the self-service surface that did the earlier work.

The pattern already shows up in fund portfolios.

In our 1,104-company benchmark, the growth-equity portfolios, Insight Partners and Battery Ventures holdings among them, run measurably deeper GTM stacks than the infrastructure and large-platform portfolios, which is consistent with owners who treat the commercial surface as an asset class of its own.

For a PE-backed platform, this is a value creation lever with a measurable P&L shape. Sales cost as a share of revenue falls when coverage matches the real journey, win rates rise when sellers engage at the stage they can influence, and the freed budget funds the public-surface work from section 3, which compounds while headcount spend does not. The transition needs sequencing across 2 to 3 quarters, because cutting the SDR layer before the inbound surface is strong produces a quarter with neither.

This week, run a 60-minute review with your CRO and CMO on one question set:

1. Of the last 20 closed-won deals, at what journey stage did the first human conversation happen, and what did the buyer already know.

2. If an AI agent evaluated us tonight against our top 2 competitors using only public information, where do we lose, and who owns fixing it.

3. What share of the sales budget covers journey stages our buyers now complete without us.

Score each one red, yellow, or green; a red on question 2 is the cheapest one to fix and the most expensive one to leave.

Mario

🛠️ Strong resumes keep failing the first technical screen. DevriX gets 4-5 applications a month from ex-Accenture, ex-Deloitte, and ex-BCG people, and almost none pass, because staffing decks and status meetings are a different skill than shipping production code or fixing an abandoned CRM. Hire operators who have run payroll and dealt with clients - execution and validation are what portfolio companies actually lack.

🎬 Brand power now functions as a distribution asset. Ryan Reynolds built and exited over $14B in brands by pairing entrepreneurial instinct with audience reach. Treat collaborations and media acquisitions as accelerants - for PE and mid-market, distribution amplifies a portfolio company faster than product improvements alone.

🧱 The first 100 days protect the base before they optimize it. Every acquired company has two or three load-bearing processes - the pipeline review, the invoicing cycle, the escalation path for the largest customer - that crater revenue if disrupted. Map the do-not-touch register before Day 1: preserving the cash flow that funds transformation is what separates disciplined value creation from expensive firefighting.

🛡️ Most value-creation plans price the upside and ignore the downside. Model a 15% revenue drop over two quarters and the dependencies surface fast: which streams fund which cost centers, where customer concentration sits, which contracts lock in fixed costs. Stress-test before the stress arrives - building the contraction playbook early costs far less than retrofitting controls in a cash crunch.

Infrastructure remains one of the deepest pools for new PE capital right now. Blackstone Energy Transition Partners agreed to buy Dresser Utility Solutions, and the same market shows AI pushing industrial software vendors to rework pricing and packaging, which moves build-vs-buy and diligence priorities up the agenda for mid-market operators.

Reimbursement policy is a direct input to healthcare-services valuations. CMS proposed its 2027 outpatient payment rule with 340B drug payment cuts and expanded site-neutral payment for off-campus imaging, which shifts revenue, margins, and service-line strategy for providers, imaging, and adjacent portfolio companies.

Founder buybacks of minority stakes are becoming a live liquidity path. Gymshark's founder is in talks to buy back the General Atlantic stake, a move that puts control, secondary liquidity, and valuation timing at the center for SME and mid-market executives weighing similar structures.

Open-source models and frontier labs are increasingly serving different workloads across a product's life cycle. Analysis of why open-source AI has not yet cut into Anthropic helps executives time build-vs-buy decisions, vendor selection, and capability adoption without overreacting to short-term model headlines.

Pet Grooming WooCommerce Brand: 8-year-old WooCommerce brand specializing in premium pet grooming products and accessories for both dogs and cats. Streamlined and efficient workflows ensure a low-maintenance business model, requiring minimal daily oversight. $182,295

Home Decor Amazon FBA: 7-year-old Australian home décor brand specializing in premium faux potted plants. Operated by a small team with streamlined workflows and reliable supplier relationships. $184,695

Trading SaaS: 5-year-old automated trading software business with proprietary MT4 Expert Advisors and broker CPA partnerships. Generates revenue via direct software sales to retail traders and affiliate CPA commissions from regulated trading brokers. $246,845

Men's Wellness Shopify Brand: Growing Shopify brand focused on non-invasive therapy and performance products for men's wellness. Operates independently with 15 full-time team members across creative production, media buying, operations, sourcing, and customer support. $5,000,000

For PE partners and operators seeking alpha

🌐 Scaling $50M - $500M+ mid-market companies with value creation through RevOps, data engineering, and WordPress. DevriX provides full RevOps consulting + delivery with GTM enablement for PE-backed portfolio companies, traditional tech, healthcare, finance, and professional service businesses pacing toward revenue growth initiatives. Our standard retainers between $10K and $60K include revenue lifecycle services for marketing and sales leaders, FP&A for financial teams, pipeline enrichment through websites and dozens of lead sources, automations and delivery integrations, CRO and ongoing testing, product delivery and platform integration solutions, and more through our consulting solutions.

🚀 1:1 Advisory retainers. Supporting operating partners, private equity funds, family offices, and mid-market executives in different capacities, from value creation through due diligence to portfolio digital GTM management in my async advisory programs via Growth Shuttle.

📈 GTM while scaling. European and international businesses can opt in for doola LLC and their “Business in a Box” model. Scaling founders can find smaller digital opportunities on Flippa. And additional opportunities across my investments can be found here.