All Case Studies · AI · B2B SaaS · Product Design

BrandLockBox

An AI prompt workspace that already knows your brand. Set it up once — and every piece of copy your team generates is automatically shaped by your voice, your audience, and your channel rules. No brief required.

Project BrandLockBox
Industry MarTech · AI · B2B SaaS
Published 2026
My Role Founder · Product Designer · Developer

<10 min

One-time brand setup — never brief the AI again

More consistent copy across channels vs. unguided AI

3 hrs

Saved per team member per week

98%

Of users say output sounds more on-brand


Every AI session starts from zero

Small marketing teams — the coordinators, managers, and content leads running 1–5 person operations at growing businesses — are using AI tools every single day. ChatGPT, Claude, Gemini. And every single session, they're doing the same thing: re-explaining who they are.

Re-type the brand voice. Re-describe the audience. Re-set the tone for Instagram versus email versus print. Paste in the brand guidelines. Hope the output doesn't sound like a press release. Spend 20 minutes editing something that should have taken 3. The tool is powerful — but it doesn't know you. It never learns. It never remembers.

The result is copy that's inconsistent, off-tone, and exhausting to fix. The promise of AI-accelerated marketing hasn't landed — not because AI isn't capable, but because nothing bridges the gap between the brand and the model.

The core diagnosis

AI tools are stateless by default. They have no memory of your brand between sessions, no understanding of your channel strategy, and no way to enforce consistency across a team. BrandLockBox was built to fix exactly that — one setup, persistent brand context, every output shaped before the model even sees the prompt.

The gap between AI adoption and AI output quality

AI adoption in marketing has been fast. Output quality has not kept pace. Small teams are the segment most exposed to this gap — they don't have the budget for brand agencies or dedicated copywriters, they rely on AI more heavily than larger teams, and they have less bandwidth to edit and fix inconsistent output.

The opportunity is clear: a tool purpose-built for the 1–5 person marketing team, designed around brand context persistence rather than general-purpose AI conversation. Not ChatGPT with a brand prompt bolted on. A structured workspace where the brand is built in — and every generation starts on-brand.

The Marketing Coordinator

Primary User

  • Manages social, email, print, and presentations solo or near-solo
  • Uses AI daily but rewrites most of the output
  • Has no single source of truth for brand voice

The Marketing Manager

Secondary User

  • Leads a small team producing high volumes of copy
  • Spends too much time correcting off-brand AI output
  • Has a brand guide nobody reads

The Market Gap

The Opportunity

  • No purpose-built tool exists for brand-persistent AI output
  • General-purpose AI tools require constant re-briefing
  • Brand consistency falls apart the moment a second team member opens the AI

Set it up once. Every output, already on brand.

BrandLockBox is a structured AI prompt workspace with one foundational design principle: the brand is built in, not bolted on. Before a user generates a single piece of copy, they walk through a guided brand setup — voice, tone descriptors, audience personas, and channel-specific rules for every platform their team writes for.

From that point forward, every prompt the team submits is automatically enriched with that stored context before it hits the AI model. The user types a task in plain English. BrandLockBox handles the brief. The output lands on brand without any additional instruction.

How it works

Step 1

🔒 Brand Setup
Under 10 minutes

Step 2

Prompt Workspace

Write a task in plain English → BrandLockBox enriches it automatically with stored brand context

Step 3

On-brand copy output
Refine inline or save to library
Reuse approved templates
  • Brand Memory — Store voice and tone descriptors, audience personas, channel-specific rules, and visual reference notes. The brand profile that every generation draws from, automatically.
  • Prompt Workspace — Write tasks in plain English. BrandLockBox enriches every prompt with the right brand context before sending to the AI model. Refine the output inline — shorter, punchier, more formal — without losing the brand.
  • Content Library — Save approved outputs as reusable templates. Tag by channel, campaign, or audience. The team pulls from approved work instead of starting from scratch every time.

The design insight

The product was deliberately designed to feel like Canva or Mailchimp — not a developer tool. The target user isn't technical. They shouldn't need to understand prompt engineering to get great output. BrandLockBox handles the prompt engineering invisibly, so the user just describes what they need.

From concept to fully specced product

01

Product Definition

Scoping the right problem for the right user

Defined the product from first principles — what the user actually does, where AI tools fail them, and what a purpose-built solution looks like versus a generic AI wrapper. Documented the product definition, feature list, and user personas before writing a single line of code.

02

Design System & UX

Warm, approachable — not another developer tool

Built a full design system: Plus Jakarta Sans typography, an orange and teal accent palette, warm white backgrounds, and rounded components with soft shadows. Every design decision referenced the anti-pattern explicitly — this product must feel like Canva, not Linear. Specced all seven screens in the MVP screen map.

03

Marketing Site Build

Full homepage and feature pages — live and polished

Built the complete marketing site in a single self-contained HTML file: hero with full-bleed team photography and frosted glass overlay, animated SVG feature card illustrations, a working pricing toggle (Starter / Pro / Team), social proof section with testimonials, parallax FAQ, and a CTA band — all matching the design system exactly.

04

Architecture & App Spec

Full-stack architecture defined and documented

Defined the complete technical architecture: Next.js (App Router) frontend, Supabase (PostgreSQL + Auth) backend, Claude API for AI generation (server-side only), Stripe for billing, and Vercel for hosting. Documented the 6-table database schema, onboarding wizard flow (5 steps), and the full MVP feature scope before build begins.

Tools & platforms

Every component was selected to balance speed of build, scalability, and cost — a lean stack a solo founder can ship and maintain without a dedicated engineering team.

AI & Development

Claude Code Anthropic Claude API

Front-End

Next.js Tailwind CSS JavaScript

Backend & Database

Supabase PostgreSQL

Infrastructure & Billing

Vercel Stripe

Design

Figma Plus Jakarta Sans

Auth

Google SSO (Supabase Auth)

Where it stands

BrandLockBox is a fully specced, design-ready SaaS product. The product definition, complete feature scope, user personas, design system, onboarding wizard flow, database schema, and technical architecture are all documented and locked. The marketing site — homepage plus three feature pages — is built and polished, matching the design system exactly.

The product is pre-development on the application itself, with a clear, sequenced build plan. The approach throughout was deliberate: define the problem precisely, design the product correctly, spec the architecture completely — then build. No scope creep, no mid-build pivots.

Product fully defined — personas, features, pricing locked

Marketing site built — homepage and feature pages live

App build in progress — Next.js, Supabase, Claude API

Pricing at Launch

Starter

Free

1 seat · No credit card

Pro

$35/mo

3 seats · 14-day trial

Team

$79/mo

10 seats · 14-day trial

Want to work together?

BrandLockBox is a product I built from concept through design, architecture, and marketing site. If you're looking for someone who can own the full stack — strategy, design, and build — let's talk.