We did not set out to build a product. We set out to fix a problem we had been solving manually for 25 years.
Alphabet® is a brand agency. We have built brands, brand guidelines, and brand systems for tourism boards, healthcare networks, government agencies, and enterprise organizations since the late 1990s. Our work has been good. Some of it has won awards. Almost all of it has encountered the same failure mode.
The brand guidelines we delivered worked for about six months. Then the gap between the document and the reality started widening. Teams changed. Tools changed. The work moved faster than the guide could follow. And the knowledge that made the brand work, the nuance that no document could fully capture, lived in the heads of the people who had been there longest.
Every client. Every time. For 25 years.
The moment the thinking shifted
In late 2023, we watched AI tools go from novelty to daily workflow in every client organization we served. Teams that previously produced 30 pieces of content a month were producing 150. The quality ceiling went up. The brand consistency floor fell out.
We were fielding more brand review requests than ever. Not because the work was worse. Because there was so much more of it, and none of the AI tools generating it knew anything about the brand.
That is when the thinking shifted. We had been treating brand consistency as a human problem: hire better people, train them better, review more carefully. But the problem was structural. Brand rules lived in a format (documents) that the dominant content creation mechanism (AI) could not read.
Brand rules are logic. Tone, voice, visual standards, messaging frameworks, audience personas, and content strategy can all be modeled as structured data. We had been delivering that logic as prose for 25 years. The medium was the failure.
What we tried first
We started where most agencies would: prompt engineering. We created elaborate brand prompts for each client. System instructions packed with voice guidelines, vocabulary lists, audience descriptions, and example outputs. We trained the team to paste these into every AI session.
It helped. For about three weeks. Then the prompts diverged. Different team members used different versions. New hires did not know the prompts existed. The context window filled up and the AI started ignoring the guidelines at the bottom. And every new AI tool (Canva's AI, Figma's AI, the next thing that launched on Tuesday) required its own integration.
Prompt engineering was a workaround, not a solution. It distributed the enforcement burden to every individual in every session. The same structural failure as brand guidelines, in a different format.
Building the system we needed
What we needed was a structural layer between the brand and every AI output. Not a prompt. A system. Something that encoded brand rules as structured data and made them available to any AI generation, any tool, any team member, without anyone having to remember to paste anything.
We started building. The first version was ugly. An internal tool that stored brand rules as structured fields and injected them into AI context automatically. We tested it on our own client work.
The results were immediate and unmistakable. Content generated with structured brand context was categorically different from content generated with prompts. Not slightly better. Different in kind. The AI was not approximating the brand's voice. It was operating inside the brand's complete system of rules.
Then we added Brand Fit. A validation layer that scored every output against the client's own rules before a human reviewed it. The score came with specific improvement suggestions. The review cycle shifted from "this doesn't feel on-brand" to "the score flagged three specific issues, all of which are already fixed."
Within a month, our internal review time per client dropped by roughly half. New team members were producing on-brand content in their first week. And the AI output quality improved with every piece we published, because the system learned from every human edit.
What surprised us
Three things we did not expect:
Surprises from the build
- The encoding process improved the brand. When you force brand rules out of prose and into structured fields, you discover ambiguities and gaps the original guide papered over. "Warm and professional" becomes a conversation about what warm actually means in measurable terms. The encoding process itself is a brand strategy exercise.
- Creative quality went up, not down. We expected resistance from the creative team. What happened was the opposite. When the system handled compliance, creative review focused entirely on ideas, angles, and storytelling. The creative directors said it was the first time in years they were doing purely creative review.
- Clients wanted the system, not just the output. When we showed clients their Brand Fit scores and the structured encoding, they did not just want better content from us. They wanted the system for their own teams. Brand intelligence became a service, not just a workflow.
Where we are now
The internal tool became Maloo®. We rebuilt it as a platform that any organization can use: brand encoding, AI content generation with embedded brand context, Brand Fit validation, multi-brand workspaces, content scheduling, analytics, and direct publishing to social and CMS platforms.
We are currently in controlled early access, onboarding organizations with real brand complexity. Tourism and destination brands. Healthcare systems. Agencies managing multiple clients. Enterprise teams with distributed content operations.
The product is real. The problem it solves is real. And the 25 years of agency work that preceded it are the reason we believe it works: we did not build Maloo® from a thesis about what brands need. We built it from two decades of evidence about what breaks.
If you manage a brand and the gap between your guidelines and your published content keeps growing, that gap is structural. We know. We lived it. And we built the fix.
Maloo® is built by a brand agency, for anyone who uses brand to grow their business.
25 years of brand work. One growth system.