Research — Anusha Ramesh Kannan
Working Paper SSRN · March 2026
Abstract ID 6436240
Original Research

The Dual-Journey Model of Marketing in the Agentic Era

What changes when AI agents become the first audience for B2B brands — and the mechanism by which the human journey programs them.

→ THESIS

Marketing has been built around one journey for sixty years. That journey is splitting in two.

When a B2B buyer delegates research and shortlisting to an AI agent, the agent walks in with a stored preference — something the buyer said in passing weeks earlier, something the agent retained, something the buyer may not even remember saying. The agent is not a neutral evaluator. It carries the buyer's prior brand experience into every comparison it makes on their behalf.

This paper names that mechanism as agent bias and the window in which it forms as the pre-delegation layer. Both are already operating in production systems. Perplexity stores favorite brands as structured preference data. Mem0 retains behavioral patterns across sessions at a benchmarked 26% accuracy gain. Amazon AgentCore extracts semantic preference facts and applies them to future recommendations. The buyer's journey did not end at purchase. It now feeds the agent's memory.

What has not yet been theorized is the mechanism connecting the two — human and algorithmic audiences — and its strategic implication: the human work has to happen first, because it is the upstream that programs the agent. From the paper, Section 5

A February 2026 Harvard Business Review article identified the gap explicitly: marketing must adapt to a customer that may be a machine with its own decision logic, but the mechanism connecting the two audiences had not yet been theorized. This paper, posted the following month, names that mechanism.

→ CONTENTS

What the paper covers

Sections
  1. What the field already knew, and what collapsed in practice
  2. Context matching — how agents evaluate brands at the semantic layer
  3. Agent memory in production — Perplexity, Mem0, AgentCore
  4. Agent bias as a mechanism connecting the two audiences
  5. The dual-audience problem and the AI-to-AI layer
  6. The Gen Yodha case — what real ICP development reveals
  7. The pre-delegation layer — where human marketing now lives
→ KEY TERMS

Definitions, as used in the paper

Agent bias
The embedding of prior human brand experience into agent decision-making. Operates in current production systems via stored user preferences, episodic memory, and semantic memory layers.
Pre-delegation layer
The period during which a human consumer is still the decision-maker, still reachable by emotional and relational marketing, still forming the brand preferences that will shape their agent's behavior. The new upstream of the funnel.
Context matching
How an AI agent evaluates a brand: by reading whether the brand's published content reflects specific, sustained understanding of a customer problem versus generic positioning aimed at a broad market.
Dual-audience problem
A brand now has two distinct audiences with two different decision mechanisms: the human who forms preferences, and the agent that acts on them. Strategies built for one fail the other.
→ CITATION

Cite this paper

For academic, journalistic, or analyst use. Direct citation establishes dated priority.

APA
Kannan, A. R. (2026). The Dual-Journey Model of Marketing in the Agentic Era. SSRN Working Paper. Abstract ID: 6436240. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6436240
SSRN Abstract ID · 6436240 Author ID · 10757736 Posted · March 2026
→ WHERE THE WORK CONTINUES

The paper is the dated record. The implications run further.

Three places where the framework is being applied in public, against current infrastructure shifts (Visa Intelligent Commerce Connect, HubSpot AEO, Shopify agentic storefronts) and against the operational question of what marketing organizations need to redesign.

Newsletter
The Thinking Marketer
Weekly essays applying the framework to current shifts in agentic infrastructure and B2B buying behavior.
Podcast
The 360° Marketing Show
Episodes that walk through the model with live demonstrations across Perplexity, ChatGPT, and Claude.
Forthcoming
Second book — in active research
A book-length treatment of what marketing organizations need to redesign for the agent era. Interviews underway with marketing leaders and founders globally.
→ JOIN THE RESEARCH
Open Call

Are you making AEO budget decisions or thinking about agent-era marketing strategy at your company?

The research behind the second book is open. Senior marketers, CMOs, and founders shaping how their organizations adapt to agent-mediated buying — your perspective is what the book is built on.

Apply to be interviewed
About the Author

Anusha Ramesh Kannan is a marketer and independent researcher based in Chennai. She is the author of The Inner Game (April 2026), founder of Gen Yodha, and host of The 360° Marketing Show. Her current research examines how marketing operating systems are adapting to agent-mediated buying.