The Terminal Was Never the Problem. Access Was.

I’ve been running KORTX for more than a decade. In that time, we’ve managed north of $300 million in media spend and watched the advertising technology landscape transform from “complicated” to “absurdly complicated” to whatever we’re calling it now.

And through all of it, one pattern keeps repeating: the most powerful tools get locked behind walls that most advertising professionals can’t (or won’t) climb.

This week, Anthropic released Claude Cowork—and it might be the first real crack in that wall.

First, a Confession

I’m a Claude Code devotee. Borderline religious about it, actually.

As someone who’s managed software teams at holding companies and still writes code regularly, I live in the terminal. Claude Code has become my go-to for everything from data manipulation to building internal tools that would’ve taken weeks the old way.

But here’s the thing I’ve had to acknowledge: every time I show Claude Code to a media buyer or account manager at KORTX, I watch their eyes glaze over the moment they see that black screen with blinking cursor.

It’s not that they couldn’t learn it. They absolutely could. It’s that they shouldn’t have to.

The Real Barrier in Advertising Has Never Been Intelligence

Let me be direct: I’ve worked with media planners who can optimize a seven-figure campaign across 47 audience segments while simultaneously managing client expectations and hitting ROAS targets. The cognitive load of advertising is genuinely difficult.

What these same people can’t do is navigate a command line to automate the soul-crushing data reconciliation tasks that eat 15 hours of their week.

This isn’t a skills gap. It’s an interface mismatch.

Claude Cowork essentially takes the autonomous capabilities of Claude Code—the ability to read files, create documents, process data, execute multi-step workflows—and wraps it in the chat interface people already know. Same engine, different steering wheel.

For advertising specifically, this shift matters more than it might in other industries.

Why Advertising Is Uniquely Positioned for This

Our industry runs on repetitive data tasks that are just complex enough to resist simple automation:

• Pulling platform data into unified reporting formats
• Cross-referencing campaign performance against media plans
• Generating insertion orders from spreadsheet data
• Reconciling invoices against delivery reports
• Building audience segment lists from CRM exports

These aren’t hard problems intellectually. They’re hard problems practically—death by a thousand clicks, each one slightly different from the last.

Traditional automation requires either expensive development resources or tools so rigid they break the moment a platform changes its export format (which is always).

Here’s what it looks like when it works. Our account managers and ad operations team used to check KPIs, performance and pacing by hand: campaign by campaign, platform by platform. So our product team built automated monitoring that watches every campaign against the criteria that matter, across every platform. Anything off track goes straight into an issue-tracking system, where it’s assigned to a person with an SLA clock running.

The result: far fewer errors and make-goods. Problems get flagged early and fixed before they ever touch the client or the business.

But notice what it took: a product team. Most agencies don’t have one. That’s the gap.

Agentic AI—the kind that plans and executes rather than just suggests—handles this naturally. But until now, accessing that capability meant either hiring developers or becoming comfortable with tools that felt designed to exclude you.

What I’m Actually Excited About

Cowork itself isn’t revolutionary technology. Anthropic is transparent about that—the team built Cowork in about a week and a half, with Claude Code writing the code. It’s the same underlying capability, just more accessible.

What is exciting is what comes next.

I can already see the trajectory for adtech-centric MCP integrations that would let non-technical advertising professionals accomplish things that currently require entire teams:

• Connect directly to platform APIs without writing code
• Build custom reporting workflows that adapt to your specific needs
• Automate campaign setup across multiple DSPs from a single brief
• Create data pipelines between your CRM, ad platforms, and BI tools

The Model Context Protocol (MCP) is the real unlock here. It’s what lets Claude connect to external systems and actually do things rather than just talk about doing things.

When these integrations mature—and they will—the advertising professional who adopts them will have capabilities that previously required a dedicated ops team.

The Mission Hasn’t Changed

At KORTX, our mission has always been to democratize technology for advertising professionals. We’ve spent more than a decade trying to make sophisticated capabilities accessible to people who shouldn’t need a computer science degree to do their jobs well.

This is the same goal, different mechanism.

The best tool in the world is worthless if the people who need it most can’t use it. That’s been true for every generation of advertising technology, and it’s still true for AI.

Cowork isn’t going to change my daily workflow—I’m staying in the terminal, thank you very much. But I’m genuinely curious what happens when every account manager and media planner has access to the same autonomous capabilities that technical people have been hoarding.

The interface was always the obstacle. Now it’s not.

Let’s see what happens next.

---

What would you automate first if the terminal weren’t in the way? Reply to the email or tell me on LinkedIn. I read every answer.