Client Brain — AI-Powered Knowledge Base Initiative

Overview

A recurring problem across client accounts is that institutional knowledge about client business models, terminology preferences, and content rules lives in people's heads rather than in a shared, queryable system. When this knowledge isn't captured, content errors slip through — wrong business model descriptions, forbidden phrases, or factual misrepresentations — that erode client trust and create rework.

This initiative proposes building a structured Client Brain knowledge base for every account, using AI to actively govern content against client-specific rules before it goes out the door.

The concept surfaced during a pre-call prep session for the ^[1] account, where a social post incorrectly described reverse ATMs as leased (they are owned by Bluepoint; business owners benefit from foot traffic, not revenue). Neither the account manager nor the strategist knew the correct model. A similar pattern exists with ^[2], where terms like "peel oil," "crushing," and "chocolate" (vs. "chocolate flavored") have caused repeated friction.

The Problem

Client-specific knowledge accumulates informally over the course of an engagement but is rarely written down in a structured way:

• Business model nuances — who owns what, who pays whom, how value flows to each party
• Forbidden terminology — words, phrases, or framings the client has explicitly rejected
• Competitive sensitivities — topics or comparisons to avoid
• Audience insights — what resonates on sales calls, what objections come up

Without a system to capture and enforce these rules, the same mistakes recur. Content writers, designers, and strategists onboarding to an account have no single source of truth.

"We're building institutional knowledge about these clients, but we're not doing a particularly good job of writing it down anywhere."
— Mark Hope

Proposed Solution

1. Client Brain Knowledge Base Section

Add a dedicated Institutional Knowledge section to each client's existing Client Brain workspace. This section would contain:

• Business model summary (who owns what, revenue flows, customer value props)
• Terminology rules (approved language, forbidden phrases, preferred framings)
• Content guardrails (topics to avoid, claims that require verification)
• Sales intelligence (common objections, what resonates with prospects)
• Running learnings log (updated after each client call)

2. AI Content Governance Layer

Feed the knowledge base into an AI content-checking workflow. Before any content (blog post, social copy, ad creative, web page) is sent to the client, it is run against the client's rules:

• Flag forbidden terms or phrases
• Catch factual claims that contradict the documented business model
• Surface missing context that the client has previously flagged

This replaces the current manual checklist approach (which exists for Bluepoint but is not systematized across accounts) with an AI-assisted pass that scales across all clients.

Example prompt pattern:

"Review this content against the following client rules: [rules]. Flag any violations or potential issues."

3. 30-Minute Discovery Call Per Client

To seed the knowledge base, schedule a focused 30-minute discovery call with each client using a standard question set:

• What is your business model? Where does revenue come from?
• Who are your primary customers and what do they care about?
• What terminology do you prefer or want to avoid?
• What do prospects push back on? What resonates?
• What have we gotten wrong in the past that we should never repeat?

These calls are framed as a quality initiative, not a remediation exercise. The output feeds directly into the Client Brain knowledge base.

Suggested question areas:

Implementation

Immediate Next Steps

• [ ] Mark Hope — Schedule a call with Evoke and Gavin to discuss AI content governance workflow and how to implement the Client Brain knowledge section (action item from 2026-03-12 prep call)
• [ ] Develop standard discovery call question template
• [ ] Pilot with Bluepoint (given active friction) and Citrus America (given recurring terminology issues)
• [ ] Roll out 30-minute discovery calls to remaining accounts within 30 days

Ongoing Process

• At the end of each client call, capture learnings: "What did we hear or see this week that we should add to the knowledge base?"
• Assign ownership of each client's knowledge base to the account manager
• Review and update rules whenever a client flags a content error

Known Client Examples

Related

• ^[5]
• ^[6]
• ^[7]