By John Hurley, Co-Founder of Walleta and Former Co-Owner of Webexcellence

I've spent a good portion of my career in the agency business.

In my early 20s, I co-owned Webexcellence, which grew into one of the largest web development firms in Indiana. At our peak, we had more than 30 people working on some of the largest web development projects in Indianapolis.

Back then, being a great web development agency wasn't just about building a website.

The website was the beginning of the relationship.

The real business opportunity came from everything that happened afterward.

We provided hosting. Email. Website updates. SEO. Maintenance. Ongoing development.

These services created recurring revenue, but they also gave us something that was arguably even more valuable:

A reason to stay connected to our customers.

That experience has shaped how I think about the next generation of agencies.

And I believe we are watching the beginning of that generation right now.

AI Agencies Are the New Web Agencies

Twenty years ago, businesses needed help getting online.

Today, businesses need help figuring out how to integrate AI into their operations.

That means building AI-powered workflows, connecting systems, automating repetitive processes, deploying AI agents, integrating models, and finding practical ways to use AI to improve the business.

This is creating an entirely new category of agency.

The AI Agency.

An AI agency might build an automated customer service workflow for a company. It might create an AI-powered sales assistant, connect multiple AI models, automate internal processes, build an AI agent, or create a sophisticated workflow using tools like OpenAI, Claude, Gemini, Make, Zapier, and dozens of other platforms.

But there is a problem I recognize very well.

Building the solution is only part of the opportunity.

The bigger opportunity is figuring out what happens after the implementation.

Agencies Have Always Needed a Reason to Keep the Customer

When we were running Webexcellence, we understood something fundamental about the agency model:

You don't want to build a website and then disappear.

You want to become part of the customer's ongoing technology ecosystem.

Hosting gave us a recurring relationship.

Email gave us a recurring relationship.

SEO gave us a recurring relationship.

Website updates gave us a recurring relationship.

Maintenance gave us a recurring relationship.

Every one of those services created value for the customer while simultaneously creating predictable revenue for the agency.

That model worked extremely well for web development agencies.

I believe AI agencies have an even bigger opportunity.

AI Doesn't Stay Static

A website might need an update every few months.

An AI implementation can change every day.

Models improve. Pricing changes. New providers emerge. Workflows evolve. Customers want new capabilities. AI usage increases. New agents get added. Businesses discover additional processes they want to automate.

And perhaps most importantly, AI has an ongoing variable cost associated with usage.

That's fundamentally different from the traditional agency model.

An AI agency might build an AI-powered customer service system for a company for $25,000.

But what happens six months later?

The customer is using the system thousands or millions of times. The agency is managing the workflows. Different AI models are being used. API costs are accumulating. The agency is optimizing the system.

The agency is delivering ongoing value.

But how does the agency turn that ongoing value into predictable recurring revenue?

Don't Make the Agency Pay the AI Bill

This is where I think the AI agency model needs to be different from traditional managed services.

The agency shouldn't necessarily own the customer's AI consumption.

If a customer is using OpenAI, Anthropic, Gemini, or another AI provider, the customer can maintain its own account with that provider.

The customer pays the AI provider directly for its actual usage.

The agency, meanwhile, charges the customer a fixed monthly fee for the service it provides.

For example:

AI Customer Service Management — $2,500/month

That monthly fee could include:

• AI workflow management

• AI agent management

• Prompt optimization

• Workflow optimization

• Monitoring

• New automation development

• Performance reporting

• Ongoing support

• AI strategy and consulting

The customer maintains its own OpenAI account and pays OpenAI for its usage.

The agency doesn't have to guess how much AI consumption will cost next month.

It doesn't have to mark up API costs.

It doesn't have to absorb an unexpected spike in usage.

Instead, it is selling what agencies have always been good at selling:

Expertise, management, technology, and ongoing value.

That's a much cleaner business model.

This Is Where Walleta Comes In

This is a big part of the thinking behind Walleta.

Walleta was built around a simple idea:

AI businesses need visibility into the economics of usage-based technology.

For an AI agency, that means being able to understand what is happening underneath the fixed monthly services it sells.

Let's say an agency has 50 customers.

Each customer pays the agency $2,500 per month for AI management.

That's $125,000 in monthly recurring revenue.

But those 50 customers may be using different AI providers, different models, and different workflows.

Without visibility into usage, the agency is flying somewhat blind.

Which customers are consuming the most AI?

Which workflows are generating the most usage?

Which customers are becoming more expensive to support?

Which implementations are highly efficient?

Where are opportunities to optimize?

Where is the agency spending engineering resources?

And perhaps most importantly:

Is the monthly fee the agency is charging still appropriate for the amount of value and support being delivered?

That's where Walleta can provide an important layer of visibility.

Fixed Revenue. Variable AI Usage.

I think this distinction is one of the most important concepts for AI agencies to understand.

The agency's revenue can be fixed while the customer's AI consumption remains variable.

For example:

This creates a much more predictable agency business.

The agency knows what it's going to collect.

The customer knows what they're paying the agency.

And the customer retains control of its own AI infrastructure and provider relationships.

Meanwhile, Walleta gives the agency visibility into the usage and economics happening underneath the service.

The AI Agency Becomes a Managed Service Provider

I think the most successful AI agencies will eventually look less like traditional development shops and more like managed service providers.

They won't simply say:

"We built your AI system."

They'll say:

"We manage your AI system."

That's a much stronger relationship.

The agency can continuously optimize workflows, introduce new AI capabilities, monitor usage, help control costs, add new functionality, and help the customer find additional opportunities to automate its business.

And that creates something every agency owner understands:

Recurring revenue.

The Agency Flywheel

This creates an interesting flywheel.

An agency builds an AI solution.

↓

The agency puts the customer on a fixed monthly service plan.

↓

The customer uses its own AI provider accounts.

↓

Walleta tracks usage and underlying AI costs.

↓

The agency sees how the customer is using AI.

↓

The agency identifies additional opportunities.

↓

The agency builds additional workflows.

↓

The customer's reliance on the agency increases.

↓

The agency generates more recurring revenue.

That's the model I would have wanted when I was running Webexcellence.

We were constantly looking for the next service we could provide that made our customers more successful and gave us a legitimate reason to continue the relationship.

AI makes that opportunity even more powerful because the technology is embedded directly into the customer's business processes.

From Websites to Workflows

When I look back at Webexcellence, the technology has obviously changed dramatically.

But the agency fundamentals haven't.

The best agencies understand their customers.

They solve real problems.

They create ongoing value.

And they build relationships that last.

The difference is that the center of gravity is moving.

We went from websites to web applications.

From web applications to SaaS.

And now we're moving from SaaS toward AI-powered workflows, agents, and autonomous systems.

I believe AI agencies are going to become one of the most important technology service businesses of the next decade.

And the agencies that figure out the recurring-revenue component early will have a significant advantage.

The Best Agencies Will Sell Outcomes, Not AI

There's another important lesson for agency owners.

Don't sell AI.

Sell the outcome.

A customer doesn't necessarily care whether you're using GPT, Claude, Gemini, or some other model.

They care that their customer service team is responding faster.

They care that sales leads are being followed up with.

They care that employees aren't spending 20 hours a week doing something that could be automated.

They care that their business is more efficient.

That's the agency's job.

Figure out where AI creates economic value.

Build it.

Manage it.

Improve it.

And create a business model that allows you to participate in that value over time.

The Questions I Would Be Asking If I Were Running an AI Agency Today

Having spent years building a traditional technology agency, these are the questions I'd be asking today:

• Can I turn my AI implementations into recurring monthly services?

• Can I create a fixed monthly price that reflects the value I'm delivering?

• Can my customers maintain ownership of their own AI provider accounts?

• Can I see what my customers are actually using?

• Can I understand the underlying AI costs associated with their workflows?

• Can I identify customers whose usage or complexity is growing?

• Can I identify additional AI opportunities within existing accounts?

• Can I make myself an ongoing part of the customer's AI infrastructure?

• Can I increase customer lifetime value without constantly selling new projects?

Those are the questions I would have been asking 20 years ago if the technology available today had existed then.

And they're exactly the kinds of questions that led us to build Walleta.

The Future of the Agency

The future of the agency isn't necessarily about charging customers for tokens.

It's about creating recurring value around AI.

The customer can own the AI accounts.

The customer can pay OpenAI, Anthropic, Gemini, or other providers directly.

The agency can charge a predictable monthly fee for the expertise, management, workflows, support, optimization, and ongoing innovation that sit on top of that infrastructure.

And Walleta can provide the visibility needed to understand what's happening underneath it all.

That's a powerful combination.

Because the agency isn't simply building an AI system and walking away.

It's building an AI-powered service relationship.

And that is something I understand very well.

When we built Webexcellence, we learned that the most valuable customer wasn't necessarily the one who paid us the most for a website.

It was the customer who continued to trust us with their technology year after year.

I believe AI agencies have the opportunity to build that same kind of relationship—but on a much larger scale.

The technology has changed.

The agency model hasn't.

Build it. Manage it. Improve it. Stay connected to the customer.

That's the next generation of the agency.

And that's the opportunity I see for Walleta.

— John Hurley
Co-Founder, Walleta
Former Co-Owner, Webexcellence