Building the case to turn a corporate website into an AI-first revenue channel

Large professional services firms invest heavily in their websites, but those sites are built for a world where buyers navigate directly to them. In 2026, a growing share of B2B buyers start with AI. They query. They get answers. And the brands that aren't structured to appear in those answers don't exist in that moment.

This initiative is built around that recognition. The goal: redesign how a global firm's primary web channel works, including its architecture, its content, its intelligence layer, and its AI-facing infrastructure, and build the business case and operational machinery to make that transformation real.

The initiative isn't a single project. It's a coordinated program with three workstreams on a shared strategic foundation, each designed to reinforce the others. The foundation informs the CMO pitch. The pilot is designed to validate what the pitch promises. Design makes the vision concrete.

These aren't descriptions of what was built. They're the reasoning behind why it was built this way.

1. 01

   Lead with opportunity, not dysfunction

   An incoming CMO has no reason to defend a site they never owned. Opening with what's broken would have created defensiveness before any trust was built. The pitch is framed around what's being missed in the AI discovery layer, not what's wrong with what exists. That shift changes the entire emotional register of the business case, from "fix this" to "capture this."

2. 02

   Reframe the problem as a split funnel, not a traffic problem

   The instinct was to lead with site traffic or engagement metrics. That framing makes the problem feel like an SEO or UX issue. The more accurate, and more urgent, framing is that large language models (LLMs) have created a new upstream discovery layer that most brands aren't present in at all. "The funnel didn't shrink. It split." That one sentence changed how the entire narrative was organized.

3. 03

   Keep the vendor out of the CMO deck entirely

   There was a temptation to include the GEO vendor as proof of seriousness. The decision was to exclude it deliberately. Mixing a mandate conversation with a vendor selection conversation would have muddied both. The sequencing: secure the mandate first, then bring in the vendor. It also means the business case stands on its own logic rather than being contingent on one vendor's claims.

4. 04

   Reject vendor-self-reported statistics

   When building the business case, several compelling data points came from vendors and their marketing materials. All of them were cut. The credibility of the pitch depends on the rigor of its sourcing, so Forrester, Gartner, 6sense, and Adobe Digital Insights became the standard. A new executive audience will scrutinize the data, and weak sourcing would undermine the whole argument.

5. 05

   Order the five strategic bets by buyer impact, not internal logic

   The five bets (Experience, Content, Channel, Intelligence, Technology) could have been ordered any number of ways: by cost, by complexity, by team ownership. The deliberate choice was buyer-first ordering: start with what the buyer experiences, end with the infrastructure that enables it. That sequencing makes the chapter read as a coherent experience design argument rather than a list of capability investments.

6. 06

   Scope the pilot to the smallest version that proves value

   The GEO pilot could have started with the full corporate domain. The deliberate choice was to scope it to a single business unit as the test environment, with clear criteria for what it needs to prove before expanding. That decision keeps the pilot achievable, the governance manageable, and the success criteria clean. It also makes the procurement conversation simpler: a bounded pilot is easier to approve than an open-ended platform commitment.

7. 07

   Separate deliverables for separate audiences

   Three distinct audiences needed three distinct artifacts: the CMO needs a narrative pitch deck, procurement and legal need a business case document, and the executive sponsor needs a use case deck for a working conversation. Combining any two of these would have made all of them worse. The discipline was to build each artifact for the specific decision it needs to enable, then link them rather than merge them.

8. 08

   Design the variable matrix around buyer state, not page type

   The personalization framework could have been organized around content categories or site sections. Instead, it was designed around the buyer's state: their mode, their moment, their segment. That choice means the system can flex across any page rather than being locked to a specific template. It also makes the design team's job clearer: they aren't designing pages, they're designing responses to buyer states.

Designing and writing a full strategic investment narrative for an incoming CMO, with no prior context, a revenue-first orientation, and a 5-chapter structure built around buyer behavior and commercial return.

Status: built and ready; pitch not yet presented

The presentation is structured across five chapters plus The Ask. Chapter 1 establishes the market shift: LLMs splitting the buyer funnel into an upstream AI discovery layer and a downstream human validation layer. Chapter 2 builds the revenue case around what's being missed. Chapter 3 lays out five strategic bets ordered by buyer impact. Chapter 4 makes it tangible through a variable matrix system, an interactive wireframe, and a real-world mid-market use case. Chapter 5 addresses execution: technology requirements, vendor landscape, and a crawl/walk/run roadmap.

The competitive landscape and agent experience detail were moved to an appendix. The logic: the human experience is more tangible for a new CMO. Lead with what they can see and react to. Let the AI infrastructure follow.

1. Chapter 1

   The Rules Changed

   LLMs split the funnel. Upstream AI territory vs. downstream human validation.

2. Chapter 2

   What We Stand to Gain

   Business case anchored in AI-referred buyer conversion quality and dark funnel visibility.

3. Chapter 3

   Five Strategic Bets

   Experience, Content, Channel, Intelligence, Technology: ordered buyer-first.

4. Chapter 4

   What It Looks Like

   Variable matrix system, interactive wireframe, mid-market use case, Figma prototype.

5. Chapter 5

   How We Get It Done

   Tech requirements, vendor assessment, crawl/walk/run roadmap.

Scoping, designing, and mobilizing a generative engine optimization pilot with a net-new technology vendor, from initial vendor evaluation through procurement initiation and working session facilitation.

Status: scoped; in procurement and legal review

The pilot is designed around one guiding principle: the smallest scope that proves value with minimal governance and technical complexity. That means scoping to a single business unit as the initial test environment, with clear success criteria before any expansion. Keeping the pilot bounded isn't a compromise. It's the argument for why the procurement and legal review is worth initiating now.

Two detailed mid-market buyer journey use cases were developed to make the pilot concrete for the executive sponsor: a cybersecurity scenario (Director of IT Security at a manufacturing company, triggered by a peer's ransomware attack) and a finance and ERP scenario (Finance Operations Manager, triggered by a CFO mandate for Series C readiness). Each includes a day-by-day AI query sequence, emotional state annotation, channel behavior mapping, and content gap analysis, making the invisible AI window visible in the channel layer.

Scoping and framing a personalization and design system that makes the strategic vision tangible for both executives and the design team, without getting ahead of what the program can actually build.

The personalization framework is built around buyer state rather than page type. Macro signals (mode, flavor, segment) define what the buyer is in the market for. Micro moments (spark, land, explore, act, follow-up) define where they are in the journey. Five GenAI experience archetypes (Adaptive Advisor, Thought Partner, Pre-Sales Assistant, Signal Amplifier, Lead Incubator) define what the site does for them at each moment.

The variable matrix system translates this framework into something a design team can work from: a defined set of inputs that determine which version of the experience a buyer sees. The interactive wireframe prototype makes this tangible for the executive sponsor without waiting for the Figma build to be complete.