Who this guide is for: Brand and marketing teams launching a first pilot with Spherical Brand AI who want to use existing brand material as their starting point — no technical background required.
You don't have to start from zero. Most companies already have brand material — guidelines, campaigns, approval experience — that just needs to be structured and translated into machine-readable form. This guide shows how a pilot team can turn existing brand knowledge into a first working Brand Core Prompt in two weeks.
Before You Start: What You Need
You don't need perfect brand documentation. What's enough: current brand guidelines (even if incomplete or outdated), 5–10 examples of recently approved and rejected assets or copy, and one or two people from the brand team who can explain approval decisions, even if those decisions were never written down.
Week 1: Review and Diagnose
Days 1–2: Bring Brand Material Together
Gather all existing brand documents in one place — guidelines, tone-of-voice examples, audience descriptions, current campaign assets. Completeness matters less at this stage than honesty: note where knowledge only exists verbally or in individual people's heads.
Days 3–4: Run the Spherical Brand Due Diligence
Based on the material gathered, the brand is reviewed across the nine dimensions of the Spherical Brands Framework®. The result isn't a score — it's a concrete list: what's clear enough to translate into rules right away, and where does a strategic decision still need to be made first.
Day 5: Prioritize the Gaps
Not every gap needs to be closed before the pilot. Prioritize the three to five points that matter most for the first test phase — usually tone of voice, no-go topics, and the most important audience distinctions.
Week 2: Translate and Test
Days 6–8: Draft the First Brand Core Prompt
The prioritized insights become the first version of the Brand Core Prompt: brand personality and tone of voice, guardrails for language and claims, and concrete examples of correct and incorrect on-brand behavior — each phrased so an AI system can apply it unambiguously.
Days 9–10: Test With Real Use Cases
Run the Brand Core Prompt against three to five realistic examples — a social media post, a product description, a customer inquiry. Review the results together with the brand team: where does the system make the right call, and where doesn't it?
Days 11–12: Refine
The failure cases from testing are the most valuable information of the week. Every incorrectly handled request points to a place where the decision logic is still too vague or contradictory — and gets worked directly back into the Brand Core Prompt.
Days 13–14: Approve and Plan Next Steps
The pilot team signs off on the first stable version of the Brand Core Prompt. From here, it's versioned going forward and grows with every additional use case — governance processes for ongoing approvals are added in the next step.
What's Realistic After Two Weeks
After two weeks, you'll have a working but still young Brand Core Prompt — enough for a limited pilot area, not for a full rollout across every channel. That's intentional: a Brand Core Prompt gets more precise through use, not through one-time upfront perfection.