AI GTM Agents: Knowing Buyers Before They Message

In today's rapidly evolving B2B sales landscape, understanding the buyer before the first interaction is paramount. Dr. Sajjan Kanukolanu, VP of Global Operations and Strategy at Position Squared, shared insights on building "AI GTM Agents That Know the Buyer" at the AI Engineer World's Fair 2026 in San Francisco. The core thesis is that traditional Go-To-Market (GTM) strategies often fall short because they fail to adapt to the buyer's advanced stage of research and decision-making by the time they reach out.

The Evolving B2B Buyer

Kanukolanu highlighted that by the time a potential buyer contacts a company, their decision-making process is largely complete. Statistics reveal that 94% of buyers use generative AI for primary research, making these platforms a "black box" for sellers. Furthermore, 67% of B2B buyers prefer a rep-free experience, and a significant 80% of deals go to buyers who are already on a pre-contact favorite list. With only 17% of total buying time spent talking to potential vendors, it's clear that sellers need to leverage AI to gain deeper buyer insights.

Three Critical Problems in AI GTM

Kanukolanu identified three key problems that GTM teams must address to succeed in the AI era:

• AI Limitations: Simply bolting AI onto existing systems doesn't empower it to understand the buyer's role, history, or intent independently.
• Integration Challenges: Existing GTM stacks may not be equipped to capture and integrate the intent signals and CRM context that AI gathers, leading to broken systems.
• Architectural Evolution: The underlying architecture needs to be AI-centric. A "bolted-on" approach hinders scalability, and failure to address all three problems simultaneously leads to failure.

The Three-Layered Architecture for Intelligent GTM

To build a system that truly knows the buyer, Kanukolanu outlined a three-layered architecture:

• Signals: This layer encompasses data from CRM systems (deals, contacts, owners) and enrichment systems, including vital signals from social media platforms like LinkedIn (engagement, job changes, intent). These signals help identify who to target and when.
• Buyer Intelligence: This involves a knowledge base with product context, buyer personas, and ICP criteria. It also includes visitor identity management, ICP scoring, and a context builder to assemble a comprehensive buyer context graph. Routing logic is also defined here to determine the right message for the right person at the right time.
• Action: Based on the intelligence gathered, this layer dictates personalized chat greetings, rep alerts with context and recommended next steps, CRM updates, and sequence triggers for outreach.

The Context Graph

A crucial element of this architecture is the "Context Graph," which unifies every signal, action, and outcome into a single, connected record per buyer. This graph links a person's signals to their account, accounts to multiple people, and incorporates deal-level information and sales/marketing activities. Without this, organizations struggle to prioritize high-intent accounts and contacts effectively.

Key Takeaways for Building AI GTM Agents

Kanukolanu concluded with four key takeaways for GTM leaders:

1. Start with Identity: The ability to identify website visitors is fundamental. A robust system with feedback mechanisms is essential.
2. Separate Score Fit and Intent: Conflating these two can lead to misdirected messaging.
3. Build an Auditable Policy Engine: This allows for easy adjustments and fixes when AI systems encounter issues, rather than relying solely on developers.
4. Let the Flywheel Compound: Every interaction and closed deal should feed back into the system, making the AI models smarter over time.

By focusing on these principles and adopting an AI-centric architecture, GTM teams can move beyond generic outreach and build intelligent agents that truly understand and cater to the needs of their buyers.