Why I Started Context Required
I'm tired of AI content that doesn't survive Monday morning.
There’s a gap between AI demos and Monday morning.
I’ve sat through the pitches. Watched the slides. Seen the “art of the possible” presentations where everything works perfectly and the future looks inevitable.
Then Monday comes. Someone has to actually build it. Someone has to make it work inside a real business with real constraints, messy data, skeptical teams, and budgets that don’t include “transformation.”
That someone is usually me.
I’m Director of AI Architecture at a marketing agency. My job is to build operational infrastructure. The unsexy systems that actually move numbers. Utilization dashboards. Research automation. Knowledge capture. Client operations workflows.
None of it is revolutionary. All of it has to work.
The problem with AI content
Most of what I read about AI falls into two categories:
Hype: “AI will change everything. Here are 47 tools that will 10x your productivity.” Written by people who’ve never had to make a workflow survive contact with actual users.
Theory: Thoughtful analysis of where AI is heading, what it means for society, the philosophy of machine intelligence. Interesting at dinner parties. Useless at 9am when something’s broken.
What’s missing is the middle ground. The practical layer. The “here’s what actually worked, here’s what didn’t, here’s what I learned” perspective from people in the trenches.
That’s what Context Required is for.
Why “Context Required”
In programming, “context” is the information surrounding a piece of data that gives it meaning. Without context, data is noise.
The same is true for AI implementation.
Every vendor demo works in a vacuum. Every case study is cherry-picked. Every “best practice” was best for someone else’s practice.
What’s missing is always context: What’s your actual workflow? What are your real constraints? What does your team actually need?
Context is required. Hence the name.
What I’ll write about
This newsletter is where I share what I’m learning as I build AI systems for a living.
Expect:
→ What’s actually working. Tools, workflows, approaches that have survived real implementation.
→ What isn’t working. The failures, the dead ends, the “great in theory” ideas that collapsed in practice.
→ Patterns I’m seeing. Across projects, across clients, across the industry.
→ Honest assessments. Of tools, trends, and the gap between promise and delivery.
→ Systems thinking. Because AI without operational infrastructure is just an expensive toy.
I won’t write about AI changing the world. I’ll write about AI doing useful things on Tuesday.
My background (short version)
Before tech, I ran a restaurant. Scaled it from zero to $1M in annual sales.
Restaurant margins are 3-5%. One inventory mistake costs thousands. One no-show on a Friday night is a crisis. There’s no room for theory in a kitchen. Only systems that work.
That obsession with operational reality followed me into AI. I don’t trust anything until I’ve seen it work under pressure. I don’t recommend anything I haven’t tested. I don’t get excited about demos.
I get excited about things that are still working six months later.
Who this is for
Context Required is for:
→ Operators who need AI to work, not just demo well
→ Skeptics who’ve been burned by hype and empty promises
→ Builders who want to know what’s actually worth their time
→ Leaders who have to make decisions about AI without getting a PhD first
If you’re looking for breathless optimism about the AI revolution, this isn’t it.
If you’re looking for what actually survives Monday morning, you’re in the right place.
What’s next
I’ll publish weekly. Usually Tuesday or Wednesday.
If there’s something specific you’re trying to figure out, reply to this email. I read everything, and it shapes what I write about.
Let’s get to work.
—James