eduba Prepared by Eduba for Alma Operations — Emerge Americas 2026
A working note for the booth

A read on where AI fits in Alma Operations' next 12 months. And where a partner helps.

Teaching non-technical people to use AI is the hardest kind of teaching. The room wants clarity. The room also wants permission to start. Alma Operations has done that work across Tech360 cohorts with America On Tech, the Idea Center at Miami Dade College, and the Miami AI School with Zapier and ChatGPT moving into n8n. Alma Operations was the speaker at Mother of AI and sits on the SMB row at eMerge again this year. The practice has a clear shape.

Written by Matt Creamer, Eduba April 2026 Miami

Sixty, thirty, ten. A rule for sorting AI problems.

My colleague Jake Van Clief has a working rule for this space. About 60 percent of what people call an "AI problem" is actually traditional code and database work. Another 30 percent is rule-based logic. Only about 10 percent is genuinely a job for a large language model. Applied to a typical small-business automation, the Zapier or n8n layer handles most of it. A custom GPT covers the rest. The trick is teaching the distinction early, so clients stop trying to put ChatGPT where a spreadsheet and a trigger would do the job.

How the layers sort out in a typical SMB automation

  • 60% Traditional code and database work. Spreadsheets, records, triggers.
  • 30% Rule-based logic. If this, then that. Zapier and n8n territory.
  • 10% Genuinely a language-model job. Drafting, summarising, classifying.

Teach that split early in a workshop and the rest of the day gets easier. Students stop reaching for ChatGPT by reflex, and start naming which layer each step of their own workflow actually belongs on.

Three things, concretely.

One

Co-delivery on workshops bigger than one calendar can hold

Eduba has run 1,500+ people through enterprise workshops since May 2025. When a corporate invitation lands that is too large to say yes to on your own, a second team on the floor keeps the room, keeps the quality, and keeps the booking.

Two

A shared methodology layer, not a personality

A reproducible frame your students can leave with, so the workshop keeps working when you are not in the room. Jake's work on interpretable context gives a structural way to teach agents and automations that holds up under scrutiny.

Three

A clean handoff path for clients who want the build

When a scoping call ends with "can you build this for us," there is a place to send it. 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.

Evidence worth reading.

Case

KPMG UK, one of the Big Four

Forty-plus executives trained. Regulated-industry audience. Consulting-firm-to-consulting-firm delivery. The shape that matters here: Eduba walks into rooms where the attendees are buyers themselves, and the training is measured on what changes the Monday after.

Reference contact on request. Delivered with a cohort design the Alma Operations team would recognise from their own classroom.

Paper

Interpretable Context Methodology

Jake's paper on folder structure as agent architecture, submitted to ACM TiiS. A structural way to teach context to your students instead of a vibes-based one. Open-source repo under MIT license.

Read the repo: github.com/RinDig/Interpretable-Context-Methodology-ICM-

Distribution instinct, built and measured.

Jake and Matt built an online community to 22,000 members in five weeks, so the curriculum and distribution instinct is practical, not theoretical. Workshop pacing, community trust, and room feel are earned the same way here as they are in a Miami classroom.

Next step: thirty minutes, one workshop you stretched too thin on.

Bring one corporate workshop you turned down or split thin across a calendar that could not hold it. We will talk about what it would have looked like with a second team on the floor, a productised SKU underneath it, and a handoff path for the clients who want the build after.

The call is for scoping. If the fit is partnership, great. If the fit is a warm intro back to Jake or over to NLP Logix, we will say so.