Forrester reports that nearly 90% of B2B buyers now use generative AI in their purchase research. Buyers ask ChatGPT and Claude to compare vendors, define requirements, build shortlists, and narrow a decision, usually before visiting a vendor’s website at all. Gartner separately predicts that by 2028, 60% of lead-scoring decisions will be made by AI, according to LeadSpot’s 2025 AI-Driven Demand Generation Benchmark Report.
Those two numbers describe the same change from opposite ends of the funnel: buyer research has already moved to AI tools and lead scoring is moving the same direction. Most demand generation campaigns are built for neither as of today.
B2B Buyer Research Now Runs Through AI
The Forrester findings describe current behavior: a buyer researching a category today is likely to run that research through an AI tool before a sales conversation happens, and usually before a vendor’s website enters the conversation at all. That research creates the vendor shortlist and by the time a prospect fills out a form or takes a call, a real preference is frequently already formed.
Insights written for a human skimming it and a web page written to be extracted, summarized, and cited by an AI answer engine read MUCH differently, even when they cover the same topic. A page can rank well in old school search while still failing to get pulled into an AI answer, since the two systems weigh structure, specificity, directness, and clarity differently.
Lead Scoring Is Moving to AI
Gartner’s 60%-by-2028 prediction covers the other half of the funnel. Standard lead scoring today mostly runs on the same signals used to define a Marketing Qualified Lead: a form fill, a content download, a job title, a page visit tracked by a pixel. Gartner expects a majority of that scoring work to move to AI within a few years, evaluating leads on a broader set of signals than a marketing team typically tracks by hand.
That raises the bar for what counts as a valuable signal in the first place. A scoring process that runs entirely on manually tracked behavioral data has less to work with than one that can also read and weigh the actual content a lead engaged with, which puts real pressure on making that content specific and well-structured.
AI-Cited Content and AI-Scored Leads Share the Same Traits
Content that earns an AI citation during buyer research tends to share traits with content that would give an AI scoring system something real to evaluate: specific claims, clear structure, genuine depth, and a direct answer near the top rather than generic filler. A whitepaper written to rank on a keyword and a whitepaper written to answer a direct question well are different documents, and the second kind performs better on both ends of this change, the buyer-research side and the eventual scoring side.
This is the same logic behind LeadSpot’s approach to content syndication: distribute genuinely educational content, structured for direct answers rather than SEO padding, through channels where AI-empowered buyers are already researching. LeadSpot’s FAQ page covers the same principle directly: content that still earns a download is content an AI tool cannot fully replicate on its own.
Two Changes for Content and Lead Data
Two changes are worth making before either trend fully plays out. First, audit whether existing gated content answers a specific question directly, with a clear claim near the top, rather than building slowly to a conclusion the way traditional long-form content often does. Second, treat lead qualification data, decision-making authority, active initiative, current vendor situation, and timeline, as content in its own right, the same structured data LeadSpot’s HQL model already confirms before a lead is delivered, since a future AI scoring system will have more to evaluate on a lead that includes that data than one that includes only a name and a job title.
Frequently Asked Questions
How many B2B buyers use AI for vendor research?
Nearly 90% of B2B buyers now use generative AI tools during purchase research, according to Forrester, cited in LeadSpot’s 2025 AI-Driven Demand Generation Benchmark Report. Buyers use tools like ChatGPT and Perplexity to compare vendors and build shortlists, often before visiting a vendor’s website directly.
Will AI replace lead scoring?
Gartner predicts that by 2028, 60% of lead-scoring decisions will be made by AI. That figure describes a change in how scoring gets done, weighing a broader set of signals than most manually built scoring models track today, rather than scoring disappearing as a function.
How should B2B marketers prepare for AI-driven lead scoring?
Building structured qualification data into the lead capture process now, decision-making authority, active initiative, current vendor situation, and timeline, gives a future AI scoring system real signal to work with, rather than leaving scoring dependent on behavioral proxies like form fills and page visits alone.
Want help structuring content and lead qualification for what’s already changing? Book a call with LeadSpot, or read the full 2025 AI-Driven Demand Generation Benchmark Report.