How Redis trained a custom GPT on rebrand documentation to automate brand compliance | Adam Haskew

The Front Lines

Subscribe Content Studio

Subscribe Content Studio

Adam Haskew.

Associate Director, Brand Experience at Redis · Redis

Adam Heskew serves as Associate Director of Brand Experience, where he leads cross-functional initiatives across digital, event, and campaign touchpoints. He partners closely with creative, product marketing, and demand generation teams to develop integrated brand strategies that drive awareness and engagement.

Adam oversees creative direction for flagship brand campaigns, event activations, and executive thought leadership programs, while ensuring consistency in messaging and tone across all public and customer-facing channels. He also analyzes performance metrics to refine messaging and strengthen overall brand impact.

Guest

Adam Haskew

Associate Director, Brand Experience at Redis

Company:

Redis

Location:

Chattanooga, Tennessee, United States

Loading episode...

Listen onApple Podcasts Spotify

In this episode of The Marketing Front Lines, we sit down with Adam Haskew, Associate Director of Brand Experience at Redis. Adam leads the verbal identity of the Redis brand — everything from trade show copy to website language — as one of three people on the Brand Experience team. Faced with a small team, a growing content demand, and an organization already reaching for AI tools, Adam didn't wait for someone else to build the infrastructure. He built it himself: a custom ChatGPT trained on Redis's brand voice, style guide, and buyer personas, now being piloted with 12 people ahead of a full marketing-wide rollout.

Topics Discussed:

Why a three-person brand experience team needs AI infrastructure more than most
What to actually feed a custom GPT to make it produce on-brand content
The three core functions Adam built into the Redis Brand GPT
How to think about AI "tell" auditing before content goes live
Why user testing is the most underrated step in GPT deployment
How to prompt-engineer for both the machine and the human on the other end

Six takeaways from this conversation.

Actionable for Sales & Marketing Tech Builders marketers

Assume Your Team Is Already Using AI — Then Get Ahead of It
Adam's trigger for building the Redis Brand GPT wasn't a top-down mandate. It was the recognition that people were already reaching for ChatGPT in their day-to-day work. Rather than fight it or ignore it, his team asked: how do we channel that behavior toward on-brand outputs? If you're a brand or content leader, this is your call to action. The question isn't whether your team is using AI — it's whether they're using it with any guardrails.
Your Brand Artifacts Are Already GPT Training Material
The raw inputs for the Redis GPT weren't novel — they were documents the brand team had already produced: voice and tone guides, style sheets, buyer personas, and strong content examples that served as quality benchmarks. The insight here is that a recent rebrand or a well-documented brand system is essentially a GPT curriculum waiting to be uploaded. If you've done the brand work, the AI infrastructure work is closer than you think.
Build for Three Distinct Use Cases, Not a Catch-All
Rather than positioning the Redis Brand GPT as a general-purpose writing tool, Adam scoped it around three specific functions: (1) guided content brief creation, (2) on-brand first draft generation, and (3) review and audit of existing content against voice, tone, and accuracy standards. This narrow scoping is what makes the tool reliable. Generic AI outputs fail because the inputs are vague — building with specific jobs-to-be-done in mind forces precision on both ends.
Build an AI "Tell" Audit Into Any Brand GPT One of the more tactical features
Adam built was an automated check for AI giveaway patterns — overused em dashes, generic phrasing, the linguistic tics that signal machine-generated copy to a trained eye. If your brand is producing AI-assisted content at scale, this kind of audit layer isn't optional. It's the difference between content that reads as human and content that quietly erodes brand credibility.
You Have to Think Like the Machine and the Marketer Simultaneously
The hardest part of building the Redis Brand GPT, according to Adam, was learning to translate between how AI processes instructions and how a marketer actually types a request. These are not the same. Prompts that feel logical to a human often produce inconsistent AI outputs — and vice versa. His solution: use ChatGPT itself to help you figure out how to prompt it, then use those learnings to train users on the same. The best GPT builders aren't just prompt engineers — they're translators.
Treat Rollout as a Product Launch, Not a Release
Adam's GPT has been live in testing with 12 users before going marketing-wide. That's not caution for its own sake — it's quality control. AI tools built for brand consistency need stress-testing across different user types, query patterns, and content categories before broad deployment. Skipping this phase doesn't save time; it creates a cleanup problem later. Think of internal AI tool rollouts the way you'd think about a product launch: beta, iterate, then ship.

More episodes

CEO · Rembrand

Marketing Consultant