Artificial Intelligence is Disrupting Digital Agencies

How does your agency add value?

Why do your clients hire you?

Go deeper than “They needed a website.”

They could get a website anywhere. Why do they choose you?

It’s surprising how many owners I speak with who don’t have a good answer. This question is at the very core of why your company exists. With the rapid onset of AI tools, having a good great answer has become critical.

I’ll connect those dots in a moment, but first, some new data on AI adoption by agencies.

Survey results & AI adoption by digital agencies

We recently published an updated 2023 Digital Agency Tools Report. One of the first questions I asked in it was “What AI tools are you using?”

I expected to see a ton of varied responses from a small group of shops. Basically, I thought we’d see the early adopters playing around with it.

I was wrong.

We’re WAY past early adopters on AI.

Over half the shops in the survey reported using some form of AI in their workflows.

I guess it makes sense, ChatGPT grew to over 100M users in ~2 months. Even with that, it does still surprise me how quickly shops integrated AI into production.

AI impacts on digital agencies

It’s looking more and more likely that AI will be able to handle much of the execution work in the near future.

Here’s Sequoia’s forecast:

GPT-4 can already do a great job on text content.

With the release of Midjourney5, images are lifelike.

And the early verdict on GitHub’s copilot has been pretty positive.

So, thinking back to that original question: “How do you add value?” Agencies add value by performing functions that their clients don’t have the capability to do, at least not as well. The lack of capability has historically stemmed from the rapid rate of change of tech and high cost barriers (dollars and time) of upskilling their internal teams. What we’re seeing with the advent of AI is, these historically high cost barriers are crashing down.

This puts a lot of the value that agencies deliver at risk.

It’s not all equal though. Some value adds are more at risk than others. If you fall closer to the Factory side in the Factory-Consultancy spectrum (find out where your agency lands), you’re probably more at risk of disruption from AI.


Recent margin pressure

For three years now, agency profitability has been under immense pressure from record demand and talent scarcity. These forces, along with generally high inflation, have conspired to significantly increase salaries.

As we showed in our “New Agency Economics – Margins Are Shifting” newsletter a year ago, those forces weighed on average net margins and reduced them to an average of ~17%.


AI is reversing the last three years of margin pressure

One of the most significant benefits of adopting generative AI is the potential for cost savings and increased efficiency. By automating repetitive tasks and streamlining workflows, AI tools can help save time and resources in areas like content creation, graphic design, and web development. This frees up teams to focus on higher-value tasks, leading to improved profitability and significantly more options.

If we look back at those early indicators of 15-25% productivity improvements and we assume utilization rates remain the same (aka there’s plenty of work coming in to keep the production teams steady). We can map those productivity improvements almost directly to cost savings. When we apply them to the standard digital agency’s cost structure (mostly salaries), we get an entirely new margin profile.

This is what an income statement looks like with a 20% productivity improvement to production-level (COGS) employees. Since the majority of COGS are employee-related expenses, we modified 95% of the COGS line with a 20% productivity improvement.

That’s a 62% improvement in net income over the industry average baseline margin profile.

When we extend this to the OpEx line, things get even more interesting.

Here we’re keeping half the amount fixed and only adjusting 50% of the OpEx costs by the 20% improvement since you have more non-employee costs here.


This raises net margin by another 12% over the production-level improvement for a total improvement of 81%.

Now, that OpEx improvement might be a bit optimistic since we’re seeing mostly quality improvements there vs. productivity improvements, but there’s potential for both.


How AI is changing agency revenue generation

Everything we just reviewed focused on cost reduction and margin expansion, but we’re also seeing advances in sales and marketing.

While I didn’t get any hard numbers on productivity increases for sales, it’s easy to imagine a sales development rep increasing their output significantly with AI tools. A large portion of their job includes content creation (email drafting and research) that can be almost completely automated.

The improvements to marketing aren’t just in how much content they can produce but also in the quality. Like devs, they’re able to spend more time on tricky parts like hyper-targeting campaigns and conversion rate optimization. This lets them generate more leads that are converting at even higher rates.

Even if this results in a slight increase in lead volume, close rates, or upsells, it’s worth looking into. On the other hand, if it results in a broad-based uplift to the entire marketing and sales process, it instantly enters “must do immediately” territory.

Tactics for using AI tools in your digital agency

So, what exactly are shops doing to achieve these gains?

Content marketing is the obvious one.

Agencies selling any kind of content generation are using AI to increase their output. The owners I spoke with are still instituting significant checks and balances here, and I’d expect this to continue for some time.

We’ve seen some attempts at image generation and branding but nothing too concrete yet. The models either need additional tuning to individual brands/styles or aren’t capable enough yet.

Code creation and debugging

That quote about replacing Stack Overflow with ChatGPT sounded crazy when I first read it, but it seems less far-fetched after a quick conversation. They’re using ChatGPT for everything from initial code creation to reviews to comment generation.

This is changing the role of devs at their shop.

It’s freeing them up to spend time on the more challenging (and value-adding) problems vs. the tedium of the job. It’s also putting some upward pressure on more senior devs.

Sales and marketing

Outreach is about to explode.

Just about every shop that responded was using AI tools in sales. Some are using it on the persona and lead research end, while others are having it draft email outreach campaigns. There have been some pretty garbage attempts at using AI for personalization, but I’d expect these to get better and more natural quickly.

I’ve seen shops using ChatGPT to build buyer personas and even design entire marketing campaigns.

With all that said, no one trusts AI completely. Each agency has an experienced “handler” checking the content generated for errors.

Back-office tedium

Ops professionals are using AI to fine-tune employee onboarding materials, knowledgebases, presentations, and more. Instead of productivity gains, I’m seeing shops opt for higher-quality output in this area.

The key risks from AI as I see them today are:

Competition and commoditization: As AI tools become more prevalent and accessible, it becomes easier for competitors to offer similar services. This can lead to increased competition and commoditization of services that were once considered unique or specialized. Your service mix and “how you add value” is a critical question.

In-housing: Competition from your clients is another area where AI is causing challenges. We’ve already seen a few instances where owners have mentioned their clients are asking them to justify their costs since “AI can do it for free.”

Data privacy and security: We’ve already seen issues with security at ChatGPT There’s definitely some risk associated with feeding client data into these tools.

Legal stuff: Who owns what is always a fun question /s. Getty is already suing Stability AI There are also a ton of questions around using code “generated” by AI that’s essentially swiped from open source libraries.

Loss of human touch: As AI tools become more sophisticated, they may replace certain tasks traditionally performed by human workers, such as copywriting, design, or coding. This may lead to a loss of human touch and creativity in the work produced, which can impact the overall quality and appeal of the agency’s offerings.

What’s an Agency to Do?

Get up to speed: The first thing is to get up to speed on the current AI capabilities. This has been evolving quickly, so it’s a bit tough. Once you have an idea of the current capabilities, it makes planning for future disruptions easier. Here’s a fantastic high-level overview that JP Holecka gave a few weeks ago Then here’s a great dev tools landscape overview and a solid generative AI overview, both from Sequoia. If you’re still looking for more, McKinsey did a decent overview as well.

Unique value proposition: Next, zero in on your unique value proposition. This can include emphasizing your expertise, niche specialization, or tailored solutions that leverage human creativity in combination with AI tools, or even the absence of AI tools. It should also work in concert with your view of where AI is headed.

Human-AI collaboration: If you’re on the “AI is the future” side, encourage a culture that promotes human-AI collaboration. Emphasize the importance of your team’s expertise, creativity, and intuition in complementing AI-generated insights or content. By integrating AI as a support tool rather than a complete replacement, you can maintain a human touch in your offerings.

New-new agency economics

After all that, it looks like margins will be shifting again. Luckily, they’re expanding instead of contracting this time. Even if all AI progress stopped today, the current productivity gains can’t be ignored.

Those that implement AI assistance early stand to gain a significant advantage. Almost doubling their net margins today, and if we look a bit further down the road, there’s even more upside potential.

It’s not without its risks, though. It requires a handler since it can just be flat-out wrong. The potential for the penalization of AI-generated content by search engines is huge. Finally, the long-term value of “everyone can make 10x more content and flood the internet with trash!!!” is questionable at best.

Even with those risks, measured use of this tech seems to provide more benefits than not.

Want help navigating this?

I know how challenging it can be to keep up with the rapid pace of the digital agency world. That’s why I created our Agency Assessment, to help owners uncover and realize their shop’s hidden potential.

Optimizing your agency gives you options.

Those options let you react appropriately to changes in the environment. Changes like AI disruption, economic shocks, etc.

It’s an intense look into every aspect of your agency that results in a fully custom report that pinpoints exactly where you should focus your team’s effort for the maximum results.

Check it out at the link below. If it sounds interesting, reach out and we can chat through the details.

Related Articles

The Factory – Consultancy Continuum

The Factory – Consultancy Continuum

This post originally appeared in our newsletter on September 13th, 2022. Sign up to get our latest research, insights, tools, and resources delivered straight to your inbox twice a month. Factory-Consultancy Evaluation Complexity 2.0: The Factory – Consultancy...

read more

Our latest insights on running digital agencies

Exclusive research opportunities, thoughts, and advice on the digital agency industry every two weeks.

"Best / most valuable updates from anyone in ages. I get hit up soooooo much, keep this kind of newsletter / value up."

Wil Reynolds

CEO, Seer Interactive

Share This