Agentic AI will save local media: Generative AI was useful; agentic AI is game-changing

Guy Tasaka

An AI agent ... can be given a task or goal and then work through the steps required to complete it. That task could take a minute, hours or days. It might involve dozens or hundreds of steps. The agent can plan, use tools, check its own work, get stuck, recover and report back.”

Guy Tasaka, managing partner at Tasaka Digital and Editor & Publisher’s technology columnist

I don’t want to bury the lede, so I’ll start by repeating the mantra I’ve been saying for the past three years: AI will save local media.

Generative AI was useful. It was interesting. It helped people write, summarize and create. But the real paradigm shift was always going to be agentic AI.

I’ve spoken at several dozen conferences in front of hundreds of media executives discussing the benefits of AI, but for a long time, agentic AI was still mostly conceptual. In the past three to six months, that has changed. Agentic AI has moved from difficult and fragile to mainstream and game changing.

Our industry suffers from a lack of resources and a lack of budget due to the implosion and erosion of revenue models over the past 20 years. We can no longer do more with less. We are often doing less with less, and the business models we employ and the communities we serve are paying the price.

So, ask yourself: What would you do with staff you never had, with skills you could never hire in your market?

Now add this: that staff works 24/7 for almost free. No HR issues. No training issues. It runs the way you want and need it to run.

Sounds too good to be true? It’s not. I’ve witnessed it firsthand and helped executives jump-start their journey. And once you start building the staff you always wanted, you start wanting to add more of these agentic workers.

What are agents?

Let’s start with where most people are today: generative AI and chatbots.

If you’re using Claude or ChatGPT, you’re probably using it like a chatbot. You ask a question, get a response, ask a few follow-ups and then start again with a new chat. You may access tools or a knowledge base to accomplish that task, but for the most part you’re still using it for a singular task.

In the past, I’ve spoken about agents through tools like n8n, which are really multi-step automations that can use AI in the process. But most of those workflows still need a trigger to start the process.

An AI agent is different. An agent can be given a task or goal and then work through the steps required to complete it. That task could take a minute, hours or days. It might involve dozens or hundreds of steps. The agent can plan, use tools, check its own work, get stuck, recover and report back.

That is the shift.

There are two categories of agentic tools that have exploded in the past six months. The first is coding agents. Tools like Claude Code, OpenAI’s Codex and Cursor have dominated headlines because they can be given a goal and resources to build working software. They plan, build, test, deliver and iterate.

If you are interested in the development side of AI tools, consider this: Anthropic reportedly used Claude Code to help build Claude Cowork, its tool for accomplishing multi-step business tasks, in 10 days. You heard that right. Software-development tools are now being used to build other AI-powered business tools. AI developing AI.

How does this benefit you as a local media publisher?

Imagine having your own tools that run parts of your business. Imagine tools that handle mundane tasks, or tasks that have been sitting on your “too hard” pile because they required staff you don’t have or software you can’t afford.

That is where agents become interesting.

I need an agent!

These are the exact words I hear from publishers and CEOs when I explain how I use my agents. There are two agent technologies worth watching closely: OpenClaw and Hermes Agent. Very rarely are the best tools free. These are.

Both are open-source projects with active development communities. OpenClaw’s creator, Peter Steinberger, was hired by OpenAI earlier this year, but the project remains open source. Both tools require a little bit of technical knowledge, or a friend who can help. Setup can take as little as 15–30 minutes, and you interact with your agent through a messaging app like Telegram, Discord, Slack or several dozen others.

That last part is important. You don’t have to sit inside a technical dashboard. You can talk to your agent where you already communicate.

I set up my first agent six months ago. Most people think of agents as assistants that check and respond to emails, summarize meeting notes or manage leads. That is useful, but I wanted something different.

I set my agent up to be my thinking partner and collaborator. I instructed it to challenge my thinking and learn how I think and solve problems.

My agent brainstorms ideas with me and gives me hard critiques. It helps me build PRDs — product requirement documents, the foundation for building software. I’ve given my agent access to tools like Claude Code and Codex so it can operate those tools on its own. I upload meeting transcripts, then debrief with the agent and discuss next steps and independently take action on my approval.

As I describe it to anyone who wants to understand, imagine having a $300,000-a-year system architect, software developer and business strategist at your beck and call for roughly $100 a month.

What couldn’t you do? What couldn’t you build?

That is my use case. My agent gave me capabilities I had wanted for my entire 40-year career. It gave me the staff and development resources I could never afford.

And once you experience that, the question changes. It is no longer, “Can AI write this?” It becomes, “What work should I assign first?” Dare to dream. Dream big.

What would you do with these capabilities? It’s not about what you can or can’t do anymore. It’s about prioritizing what you should do.

Guy Tasaka is a seasoned media professional with a 35-year track record of leading change in the industry. He has collaborated with renowned organizations, such as Macworld Magazine, Ziff-Davis and The New York Times, where he honed his expertise in research, strategy, marketing and product management. As the former chief digital officer at Calkins Media, Guy was acknowledged as the Local Media Association's Innovator of the Year for his work in advancing OTT and digital video platforms for local news organizations. He is also the founder and managing partner of Tasaka Digital, specializing in helping media and technology companies navigate business transformations using his extensive experience and forward-thinking approach. He can be reached at guy@tasakadigital.com.