60%
Structure work.
Intake forms. SOW templates. Revision logs. Client asset libraries. Invoice flow. None of it needs an LLM. It needs a structure.
A working read. Not a pitch deck.
A read on where AI fits in Bold Creative's stack. And where it does not.
Pulled up your site before walking over. boldcreative.world has an SSL cert mismatch, so I landed on your bio.site instead. Nice hook on Make BOLD Moves. Here is the rest of the read.
Four people. Miami. Branding, marketing strategy, product development.
The Bold Creative ops lead came in last year because projects were starting to slip. The new domain is live but the cert is broken, so prospects who search you land on a portfolio page marked Coming Soon. eMerge is carrying a lot of weight this week.
This is a read on where the next 12 months of AI noise actually helps you. And where it does not.
02
The 60 / 30 / 10 frame
30%
Rule work.
Brand guideline enforcement. Voice consistency checks. Approval gates. Rules and lightweight logic do the job cheaper than a model.
10%
Real AI work.
Brand-aware copy drafting. Moodboard generation. Pattern spotting across a client's existing assets. That is where a model earns its keep.
The trap for a creative shop is putting a model where a template would do the job. Cheaper, faster, more predictable. The opposite trap is smaller but real: insisting on templates where a model could actually draft the first pass of a brand voice from an intake conversation.
03
The KPMG read.
The output was not a tool. It was a methodology those executives carried into their own client work.
KPMG UK (Big Four) brought 40+ executives through a consulting-firm-to-consulting-firm engagement. The structure holds for a brand studio.
You do not need to become an engineering shop. You need a way to make AI-shaped work quotable, deliverable, and defensible when a client asks for it.
- 40+ executives trained
- 1,500+ people trained across enterprise engagements since May 2025
- 95% still using the tools 30 days after the workshop
04
The paper that fits your shop.
Interpretable Context Methodology (ICM).
Folder structure as agent architecture. Published, reproducible, applicable to a shop that juggles ten client brands at once.
L0 identity through L4 working artifacts. A layered way to organize every client brand the same way so the next project does not start from a blank folder.
05
The partner under the layer.
Eduba partners with NLP Logix for work that sits below the orchestration layer. NLP Logix has been in machine learning since 2011 and runs over 150 data scientists. When a client asks you for a production data pipeline or a custom-trained model, that is NLP territory. Not yours. Not ours.
06
Who is bringing this.
Jake Van Clief / Eduba
- Marine Corps veteran. Eight years. Cryptographic systems, F-35 / F-18 avionics.
- MSc Future Governance, University of Edinburgh.
- Published in ACM TiiS (ICM) and arXiv.
- 1,500+ people trained across enterprise engagements since May 2025.
- Built an online community to 22,000 members in five weeks.
The next move
30 minutes with Matt.
Bring one client project where scope slipped. We walk the stack together and mark which pieces belong where: structure, rule, or real AI. No deck. No sell. You leave with a marked-up stack either way.
Book 30 minutes with Matt CreamerMatt Creamer, CRO / Eduba. calendly.com/thecro-eduba/30min