Amagenta Studio is a Colombian branding agency with international clients. It wanted to stop delivering brand books nobody reads and start delivering a system that enforces the brand on its own. Since July 2026, WebMinds, an AI and software development company based in Atlanta, Georgia, that works with companies in the United States and Latin America, has been building INWARD with Amagenta: an AI brand operating system where each client's marketing team loads its brand DNA, designs templates and lets any employee turn one sentence into a professional, on-brand piece. In fourteen weeks it went from first commit to a live event demo and a V1 MVP ready for its first enterprise pilots.
Key facts
- 14 weeks from first commit to pilot-ready: an empty repository in July 2026 to a closed V1 MVP in October 2026.
- 5 brand checks on every piece: banned words, palette, brand typography, logo presence and legibility.
- 65 brand-adaptive templates across 14 formats, from social posts to roll-ups.
- 18 of 18 tables under row-level security, red-teamed before the first pilot.
- 57 acceptance test cases delivered with the V1 MVP.
Context: a brand book that governed nothing
Every agency knows the pattern. A brand identity is delivered as a beautiful PDF, the client thanks you, and within months the internal flyer for HR or the slide for a sales event is using the wrong blue, a stretched logo and a font nobody approved.
At Amagenta, knowledge of every client brand lived in PDFs, loose folders and generic AI prompts, and every internal piece pulled the design team in. The agency wanted to turn its method into a product it could sell to companies.
The challenges:
- The brand book governed nothing: rules lived in static PDFs, were applied from memory and broke in every internal piece.
- Design was the bottleneck: any communication, however small, from HR, sales or events, had to go through the creative team.
- Generic AI, generic results: each person used their own AI tool, with no brand context and no quality control.
- Multi-company from day one: one product serving many companies at once, with strict isolation and different roles for the agency, marketing teams and employees.
Why generic design tools and AI generators didn't fit
Design platforms with brand kits exist, and AI image generators can produce something that looks finished in seconds. Neither solves Amagenta's problem.
Brand kits store colors and logos, but they don't stop a user from breaking the layout, and they don't check the result against the brand's own rules. Pure AI generators are worse for governance: the model decides the logo, the colors and the typography, and the output drifts from the brand a little more with every prompt. Amagenta also needed something those tools aren't built to be: a product it owns, serving many client companies on one platform, each strictly isolated, with the agency, the client's marketing team and its employees in different roles.
So the design principle became simple: brand rules should not be a document people consult. They should be the architecture every piece runs through.
What we built: an AI brand operating system for many companies
INWARD is a multi-company Brand OS in production, built from scratch and fully owned by Amagenta. It brings together our AI and automation and web and app development work, along with product consulting.
Brand DNA Engine
The brand identity is loaded once, through an AI conversation or a quick form: logo, palette with roles, typography, tone, audiences and banned words. INWARD turns all of it into rules the machine understands.
Prompt-to-piece generation: the AI proposes, the engine composes
One sentence in natural language becomes an on-brand piece with color variations. The AI decides the content; a deterministic engine composes the design. Concretely, Gemini reads intent, channel and format and returns schema-validated JSON with a template choice and its text. The engine then builds the piece, so logo, typography and colors always come from the brand DNA, never from the model.
Automatic brand validation
Five rules are checked on every piece (banned words, palette, brand typography, logo presence and legibility) and shown to the user before export. Templates go through a blocking lint before they can be published, so a broken template never reaches employees.
Template Lab: governance by architecture
Template Lab is an in-house canvas editor where marketing builds and publishes templates with layers, a grid, a safe area and locked or editable fields. Templates declare which layers are editable and which are locked, and by which role. An employee cannot step outside the brand because the editor does not allow it. Every piece keeps a frozen copy of its template, so redesigns never break history.
Variations that are still the brand
Each piece offers Original, Primary and Accent variations that tint the veil and background with brand colors only. Text color is recalculated by WCAG contrast against its effective background, so every variant stays legible.
One render model, four outputs
The editor, the preview, PNG and PDF share the same geometry and type metrics, with a per-family width factor so headlines never get cut. Exports come out as PNG, JPG and PDF. Any brand typeface loads at render time, with a failure cache so a slow font never takes an export down.
Multi-tenant platform and event mode
The agency, client companies and employees each have distinct roles under row-level security. There is also a public, login-free kiosk mode with a database-enforced generation quota, which is what made a live event demo possible.
Under the hood: red-teamed before the first pilot
INWARD looks like a design editor, but roughly half of the codebase is engine, data and security: 18 tables, all under row-level security, 44 schema access policies plus 9 storage policies, 23 versioned migrations, 61 server actions in 27 modules, and 17 SQL functions plus 11 triggers.
Permissions across the studio and client companies are resolved in the database with security-definer functions and a trigger that blocks self-promotion. Before the pilots, QA agents attacked the platform on purpose, and every critical finding was closed with migrations and re-verified.
We build with Hybrid AI Coding: senior WebMinds engineers own the architecture, the product decisions and every conversation with Amagenta, while AI agents (Claude Code) handle implementation under their direction. Every change goes through independent QA agents for function, UX, security and documentation before it reaches production.
If your team is drowning in small design requests or your brand rules only live in a PDF, get a free assessment and we will look at what can be automated safely.
Timeline
- June 2026: WebMinds proposal for a multi-brand Brand OS with AI.
- July 1, 2026: first commit, with the monorepo and a multi-company schema under row-level security.
- July 2026: conversational brand onboarding, Template Lab on canvas and 58 universal starter templates, plus a shift with Amagenta from building more templates to the brand-rules engine and on-brand variations.
- August 2026: product alignment with Amagenta, login-free event mode, light and dark themes, and adversarial QA of the engine.
- August 26–29, 2026: live demo on tablets at Amagenta's GoFest stand, with visitors generating pieces without an account.
- October 2026: V1 MVP closed, with team templates generating pieces, tinted variations, a piece editor, brand typography, a security red-team and pilot onboarding without SQL; 57-case acceptance suite delivered.
- Ongoing: acceptance testing with Amagenta; enterprise pilots next.
Impact so far
INWARD is a product heading into its first enterprise pilots, so its impact today is structural.
A brand that enforces itself
Brand rules stopped being a PDF. The engine applies them to every piece and the platform shows they were met. An employee with no design training produces professional communication from one sentence, which takes routine internal operations requests off the creative team's plate.
From agency to software company
INWARD turns Amagenta's method into a multi-company product it owns and can sell to enterprises, with its first pilots in preparation.
Enterprise-ready from the first pilot
Row-level security across the whole data model, roles by organization type, a red-team before the first client and a 57-case acceptance suite delivered with its limits declared. More screens are in the INWARD case study.
FAQ
What is an AI brand operating system?
It is software that stores a brand's rules (logo, palette, typography, tone and banned words) in a form a machine can apply, and then uses them to generate and check every piece. In INWARD, AI writes the content, a deterministic engine composes the design, and five automatic checks confirm the result is on brand.
How do you stop AI from going off-brand?
By not letting the AI make brand decisions. Gemini only chooses a template and writes text, returned as schema-validated JSON. Logo, colors and typography always come from the brand DNA, and locked template layers can't be edited by employees.
Can employees without design training really produce on-brand pieces?
That is the point of the product. Marketing designs and locks the templates in Template Lab; employees describe what they need in one sentence and get a piece with on-brand color variations, already validated against five brand rules.
Is INWARD built for agencies or for companies?
Both. Amagenta, the agency, owns and operates the platform, and each client company gets its own isolated space with roles for its marketing team and employees.
Can you build a product like this for our agency or company?
Yes. WebMinds works with agencies and companies in the US and Latin America on AI products where governance matters. For a similar "AI proposes, people approve" pattern in logistics, see the Kila document engine; for a full operational platform, see SWAP Translators.
Turn your method into a product
INWARD shows what happens when brand rules become architecture instead of a PDF. Explore the full INWARD case study for the screens and system design.
Get your free assessment or book a call with our team.