AI is changing agency work, but it isn’t closing the door on web shops, marketers, or freelancers. If anything, it’s exposing how much money many of us leave behind because we still price and package work like it’s 2018.

In this Agency Action conversation, Toby Cryns and Kurt von Ahnen make a simple point: agencies don’t lose when AI gets faster, they lose when they keep saying “that’s not what we do.” The bigger risk is thinking too narrowly about what clients will pay for.

AI isn’t the problem, narrow positioning is

We keep hearing the same fear: AI is invading everything, so agencies are about to get squeezed out. That fear makes sense if our agency only sells hours for design, copy, or routine site edits. It makes less sense when we step back and look at the full client relationship.

Kurt von Ahnen framed this through consulting work outside the WordPress bubble. In one training session, small process changes, with no new hires and no new equipment, created a path to more than $1 million in added annual revenue for each of two stores. That kind of result matters because it shifts the question. The issue isn’t whether AI can do a task faster. The issue is whether we can spot revenue that was already sitting in plain view.

A lot of agency owners still talk themselves out of that work. We hear versions of the same lines over and over:

• “I only work with nonprofits.”
• “That’s not the kind of project we do.”
• “What would my clients think if we offered that?”
• “We can’t charge for that.”

Those are business rules we made up. We can change them.

If we run the agency, we decide what we sell, how we package it, and what it costs.

That doesn’t mean we should take every client. Bad-fit clients are still bad-fit clients. Poor communication, mismatched budgets, and constant red flags will drain a team no matter how good the scope looks. Still, that’s different from rejecting a whole category of work because it doesn’t match an old label on our website.

The point is simple: we don’t need to let AI define the edges of our business. We can widen those edges ourselves.

Stop turning away revenue because the work changed shape

One of the strongest ideas in the discussion was the difference between rejecting a client and rejecting a project type. Those are not the same decision.

Many of us have learned the hard way that some projects fall apart because the client isn’t a fit. Maybe they can’t communicate clearly. Maybe their budget is far below the level of work they want. Maybe their process collides with ours at every step. Referring those people elsewhere is healthy.

What hurts is turning away work we could do, or could learn to package, because it sits a little outside our comfort zone.

A great example came from a WordPress build that later moved to Squarespace after the client hired a marketer who preferred that platform. The original site had already been built and paid for, so it wasn’t a total loss. Still, there were at least two missed opportunities. The agency could have sold training and onboarding to help the new marketer work with the WordPress site. Or it could have handled the migration itself and been paid again for the transition.

That lesson matters because it happens in smaller ways every week. A redesign request comes in. A homepage needs fresh thinking. The old process says design and dev will take 10 to 20 hours, so the quote lands far above the client’s comfort zone. The better response is to ask whether AI can shrink the prep work and help us get to a decision faster.

Instead of spending days on polished concepts, we can use AI to produce a few solid homepage directions in an hour or two. Those are not final designs. They are sales tools. They help us pitch faster, scope faster, and move to development with less paid discovery.

The market is already moving this way. Tools such as Promptless for agency WordPress work and Kintsu’s agency dashboard exist because agencies need faster ways to handle briefs, edits, and repeat client work. We don’t need to copy any one tool’s model to see the bigger point. Time is compressing, so our packaging has to change with it.

Productize more of the work, even on WordPress

Agencies don’t need to live inside one platform

WordPress agencies often act as if every client must live in a world of endless customization. Meanwhile, plenty of businesses are happy with tools that have hard limits. Squarespace, Wix, Webflow, Duda, Drupal, and Joomla all came up in the discussion for a reason. People buy them because they can get something working without opening a custom scope every time they want to move a button.

That doesn’t mean WordPress is the wrong answer. It means our business model may be.

When every project turns into “we can build anything,” margin gets messy. Sales cycles get slower. Clients struggle to picture cost because every detail feels custom. SaaS products avoid some of that because the boundaries are clear. Fill in the fields, pick from what exists, publish the site, move on.

We can create the same clarity inside a WordPress agency. We can say yes to WordPress while still defining what a starter package includes, what it excludes, and what costs extra.

The startup package model works because the boundaries are clear

Kurt described a productized offer at Manana No Mas that shows this well. The package uses an AI-assisted builder to create a basic site structure, usually a homepage, services page, about page, and contact page. Then the agency installs and configures a set group of tools, such as forms, CRM connections, and LMS components. After that, the client gets training on how to manage the basics.

There is no pixel-by-pixel adjustment phase built into the starter offer. There is no endless revision loop. Text and images can change, but deep custom work is separate. That boundary protects margin and keeps the entry price low enough for smaller businesses.

This side-by-side view helps explain the shift:

Offer type	What the client gets	What we control	Revenue model
Custom build	Fully tailored design and features	Scope, revisions, custom code	One-time project, larger ticket
Productized starter site	Fixed pages, fixed tools, limited edits	Process, timeline, package rules	Faster sales, repeatable margin
Ongoing growth package	Content, support, ads, reporting, updates	Retainer terms and response times	Recurring monthly or annual income

The key takeaway is that productized work doesn’t replace custom work. It creates a cleaner front door, and then it gives us space to upsell smarter.

That matters even more for learning and membership sites. When we build with an LMS such as LifterLMS, launch day is rarely the finish line. Course sites often add payment gateways, recurring billing, memberships, quizzes, assignments, PDFs, private areas, groups, translations, and WooCommerce sales flows over time. They also need theme checks, plugin conflict review, support workflows, and privacy-minded handling of student data. Those needs are not edge cases. They are normal follow-on services, and they should be priced that way.

Expand beyond the website and keep the relationship

Recurring services are often easier to sell than a rebuild

A lot of agencies still build the site, collect the check, and walk away. That model leaves recurring income on the floor.

The conversation moved through several service lines that many web shops could sell right now: email marketing, SEO, content writing, Facebook ads, TikTok ads, Google Ads, CRM setup, platform management, and support retainers. Some agencies already subcontract this work, which is fine. The missed opportunity is pretending those services sit outside the business when clients are already buying them somewhere else.

One example from the discussion said it all. A client was spending $60,000 a year on Google pay-per-click, while the agency still felt nervous about charging a small fee for upkeep work. That gap tells us something. Clients often have budget. We simply don’t ask for the larger share of it.

If we need proof that recurring revenue is a healthier model, Pressable’s guide to recurring revenue for WordPress agencies makes the case clearly. Support plans, SEO, reporting, hosting, and ongoing improvements create steadier income and stronger client relationships than a long string of one-off builds.

Paid ads fit that model too. The ad spend itself is separate from management fees, and Meta’s ad budget and pricing overview shows how flexible those budgets can be. Whether we run campaigns ourselves or subcontract an expert, there is room for management revenue, strategy calls, and monthly reporting.

Bigger offers still win when they solve a bigger problem

Another useful point from the episode is that clients do not always choose the cheapest path. Agencies often assume they will, and that assumption hurts sales.

In one example, a client asked for a contact form and received two options: a basic form and a much richer version that cost several times more. The client chose the richer version. The same thing happened with a mapping feature. There was a simpler plugin-based option and a much more advanced option inspired by Airbnb-style map behavior. Again, the client leaned toward the higher-cost version.

That pattern shows up in course and training work too. A proposal might begin with a modest setup fee, then jump sharply when we add video production, voiceover, editing, or advanced build work for multiple courses. The number can feel big until we compare it to the value. Training projects, especially inside LMS builds, can expand fast when we include media production, quizzes, reporting, student support, or advanced assessment tools. LifterLMS alone has a large ecosystem of add-ons for payments, quizzes, assignments, forms, social learning, private areas, and ecommerce. A course launch is often an operations project, not a page-build project.

If the scope is real, the price can be real too.

Support and content are where agencies leave easy money behind

Membership and CMS support can become a real business line

Support work sounds small until we price it as a service line instead of a favor.

The example in the conversation was simple. If we build a membership or course site, we can set up a support email address that forwards to our agency. Then we handle the routine issues: password resets, login confusion, missing course access, dashboard problems, and account questions. For the client, that’s a relief. For us, it’s recurring revenue built on a system we already know.

This matters a lot for LMS sites. Anyone who has supported a course platform knows the questions pile up around enrollments, payment status, quizzes, assignments, notifications, and user roles. Add translations, add-ons, theme compatibility, and occasional plugin conflicts, and support becomes ongoing work by default. The knowledge base around LifterLMS shows how broad this gets. Beyond the core plugin, there are documented paths for Stripe, PayPal, WooCommerce, forms, social features, private areas, advanced quizzes, translations, and systematic troubleshooting. That means a learning site needs ongoing help long after launch.

Content work is still valuable even when AI helps create it

Most clients do not want to be webmasters. They want to run their shop, manage their team, fix the cars, coach the staff, or serve the next customer. The website matters, but it isn’t their craft.

That’s why content and SEO remain strong agency services. AI helps us draft faster, research faster, and create first versions faster. It does not remove the need for editorial judgment, business context, publishing workflow, or consistency.

Kurt shared a strong example from a race team website. The team had posted a result on social media, but the website itself had gone quiet because the people running the team were busy doing race team work. AI made it possible to turn one social post and supporting event details into a usable website draft with context about the track and the event. After a quick review, that became a publishable post and a clear proof of service value.

That is where agencies can win. We don’t need to sell “AI content.” We can sell fresh site content, better publishing cadence, smarter outbound links, and regular SEO motion for clients who will not do it themselves.

Be the growth partner, not only the builder

The sharpest insight in the whole discussion was the mindset shift from “we build websites” to “we help clients grow.”

If we stay trapped in the first label, then every request sounds like a design job, a dev task, or a maintenance ticket. When we adopt the second label, the same request can become a support plan, an ad management retainer, a CRM cleanup project, a content package, an LMS upgrade, or a training engagement.

The Ducati story used in the conversation makes the point well. A dealership sold a premium bike to a young buyer, assumed money would be tight, and skipped the gear and accessory sale. The buyer came back the next week wearing thousands of dollars in branded equipment purchased somewhere else. The store missed the revenue, but it also failed the customer. It could have packaged that gear into the financing and made the whole purchase easier.

Agencies do this all the time. We build the site, then the client hires one person for GoHighLevel, another for SEO, another for ads, another for content, and someone else for platform support. The business becomes fragmented. Communication gets messy. Results get harder to track. We also lose the revenue because we assumed the client wouldn’t buy more from us.

That assumption is often wrong.

Final thoughts

AI is forcing agencies to work faster, package better, and stop hiding behind narrow service labels. That pressure is real, but it also shows us where the revenue has been all along.

When we move beyond one-time builds, productize what can be repeated, and offer support, content, ads, training, and platform management, we stop competing only on hours. We start building a business that fits how clients already spend money.

The money isn’t hidden because clients refuse to pay. It’s hidden because too many of us never put the offer on the table.