MARKETING SUMMIT · 2026

AI & THE FUTURE OF MARKETING

TABLE OF CONTENTS
  1. 01 Why Listen to Me — Cover
  2. 02 Why Listen to Me — Credibility
  3. 03 Common Language
  4. 04 Where Modern AI Started
  5. 05 What's Changed?
  6. 06 How to Think About the Tool
  7. 07 What Can't Be Replaced
  8. 08 Building Your AI Workspace
  9. 09 Workspace — Make a Workspace
  10. 10 Workspace — Set Instructions
  11. 11 Workspace — Attach Context
  12. 12 Workspace — Share the Space (Bonus)
  13. 13 Keep Learning
  14. 14 Up Next
AI & THE FUTURE OF MARKETING
LANCE RODGERS
Sr. Analyst, Creative Solutions Technology
ENTHUSIAST & SKEPTIC
FIG.01
CTX//01
↑ Contents01 / 14Next →
Why listen to me?
> RESEARCHING SINCE 2018
> ACTIVE USE SINCE 2021
$4,032/YR
in subscriptions replaced by self-hosting
FOUNDRY · IMMICH · AUTHENTIK · N8N + MORE
TIME, NOT TALENT
FIG.01
CTX//01
← Back02 / 14Next →
Common Language
FIG.02
CTX//02
SOURCE: UNION OF CONCERNED SCIENTISTS
AI
Anything that attempts to emulate human intelligence.
MACHINE LEARNING
Systems that make predictions in new situations using known data.
NEURAL NETWORKS
Systems of machine learning frameworks that loosely resemble human neurons.
DEEP LEARNING
A buzz word that predates OpenAI's GPT 3.5, the first model to hit mainstream in 2022.
LLM
Large Language Model. A specific flavor of deep learning architecture.
← Back03 / 14Next →
Where Modern AI Started
FIG.03
CTX//03
DEEPMIND'S CAT
"A loose and frankly awful analogy is that our numerical parameters [an array of 16,000 processors used to create a neural network with more than one billion connections] correspond to synapses."
He noted that despite the immense computing capacity used, it was still dwarfed by the number of connections found in the brain.
— DR. ANDREW Y. NG, STANFORD UNIVERSITY, 2012
← Back04 / 14Next →
What's Changed?
SCALE
More data. More compute. Bigger models.
ACCESS
A research lab then. A browser now.
INTERFACE
Code then. Plain English now.
EASIER. NOT SMARTER.
← Back05 / 14Next →
How to Think About the Tool
"Good design is not about inventing from scratch, it's about recognizing patterns that already work and arranging them with care."
— CHRISTOPHER ALEXANDER, ARCHITECT, PATTERN THEORY
← Back06 / 14Next →
What Can't Be Replaced
01
CRITICAL THINKING
Data to insight can be accelerated.
Insight to action is human.
02
EXPERIENCE
Recognizes what's been seen before and invents new patterns
03
SOFT SKILLS
Trust is built over time, not generated on demand
04
JUDGMENT
Taste can't be automated.
Know which rules to break, and when.
AUTOMATE THE GRIND. OWN THE CRAFT.
← Back07 / 14Next →
BUILDING YOUR AI WORKSPACE
EASY AS 1, 2, 3
01
MAKE A
WORKSPACE
02
SET
INSTRUCTIONS
03
ATTACH
CONTEXT
← Back08 / 14Next →
STEP 01
MAKE A WORKSPACE
CLAUDE
PROJECTS
Persistent containers for chat history, instructions, and files
CHATGPT → Projects. Same idea, transfers directly.
GEMINI → Closest is Gems, a persona, not a project.
One Project per job, not one catchall for everything
Give it a sensible name
Reuse and refine a Project instead of rebuilding from scratch
← Back09 / 14Next →
STEP 02
SET INSTRUCTIONS
CLAUDE
PROJECT INSTRUCTIONS
Static directions Claude reads every time
CHATGPT → Custom Instructions. Two fields: about you, and how to respond.
GEMINI → Saved Info. Looser, drawn from past chats and connected apps.
Keep it short, readable in less than 90 seconds
Be specific. Vagueness causes frustration later and wastes tokens
If your instructions describe a process, that portion should be a Skill instead
← Back10 / 14Next →
STEP 03
ATTACH CONTEXT
CLAUDE
PROJECT FILES
Explicit, document-driven, read fresh every time
CHATGPT → Project files and Custom GPT knowledge.
GEMINI → Gem knowledge files, Docs, Drive, Sheets.
Use good documentation. These will inform the whole project
Organize by concept, not by date or meeting name
Keep up to date. Remove old files
← Back11 / 14Next →
STEP 04 · BONUS
SHARE THE SPACE
CLAUDE
SHARED PROJECTS
Your chat stays private by default. Teammates see the shared instructions and files, not each other's conversations
CHATGPT → The opposite default. Everyone in a shared project can read everyone's chats.
GEMINI → No project-level model and uses Google Workspace sharing.
← Back12 / 14Next →
Keep Learning
SEE INSIDE THE BOX
bbycroft.net/llm
Watch an LLM actually run, token by token
HOW IT WORKS
UCS: AI 101
An accessible primer on how AI actually works
WHERE AI ART STARTED
Harold Cohen's AARON
"Generative AI," built in 1973
STAY CURRENT
Nate B. Jones
Follow the AI news cycle for high-level changes and strategy
PRACTICAL KNOWLEDGE
Tina Huang
A data scientist gives advice on how to use these tools yourself
GO DEEPER
IBM Technology
Two videos on how models actually learn
THE WORK HASN'T CHANGED.
HOW WE APPROACH IT HAS.
← Back13 / 14Next →
Up Next
GROUP 1
STRATEGY & INSIGHTS
Erik Ortiz
Josiah Gun
GROUP 2
CONTENT & COMMUNICATIONS
Luis Valtierra
Janae Bell
GROUP 3
PRODUCTIVITY & PROJECT MGMT
Ryan Nahe
Sandra Carrasco
← Back14 / 14↑ Contents