TL;DR
Senior coders, marketers, and designers, that is. The industry as a whole has essentially stopped hiring entry-level and junior-level talent for a few years, which could make senior-level talent incredibly valuable in the coming years. For the sake of argument, we’re going to assume that LLMs continue to become more capable over the next 5 years. But even if they don’t, there’s still going to be a major hole in the market where 2023’s entry-level hires should have become 2028’s mid-level talent. Let’s explore how this is playing out across the three core agency disciplines: software, marketing, and design. AI’s Already Eating SoftwareWriting code is one step in a longer chain. Architecture decisions, system integration, testing, and deployment often determine the success or failure of a project. The importance of “non-code writing activities” becomes clear when we look at engineering benchmarks. Tools like SWE-bench evaluate whether systems can resolve real GitHub issues, which typically require coordinated changes across multiple files that must pass a full test suite. Today’s top models are solving three-quarters of these issues. On the other hand, you have security and reliability. A 2025 analysis by Veracode, which tested more than 100 large language models across 80 coding tasks, found that AI-generated code introduced known security vulnerabilities in 45% of cases. Many of those aren’t the things entry-level talent will notice, but senior-level talent will flag them immediately. So you have a situation where AI can solve the simple stuff that early-stage talent would normally work on, but it has trouble with senior-level tasks, and at the same time, the talent pipeline has been severely disrupted. We’ll still get some new seniors from the mid-level talent that’s still in the pipeline, Next, There’s MarketingIf we map what’s happening in the software space to marketing activities, we see a similar, but less pronounced pattern. It’s been most evident in the content space, but it’s seeping into ads, conversion-rate-optimization, SEO/GEO, and more. Google’s own Ads roadmap describes campaign creation flows where the system can summarize a landing page and generate keywords, headlines, descriptions, and images, with humans editing and shipping. And there are already examples from agencies of cost and delivery-time compression in our surveys. Finally, DesignDesign, specifically web design, is in a weird spot, and I still blame Bootstrap for it. Almost every company that purchased design services wasn’t buying art. They were buying attention or conversions. And this sucks when you conflate design with art because the market has forced an optimization of design. This is why everything from websites to phones to cars looks the same now. It’s lame. I don’t like it. But from a “growing an agency”-perspective, there are fascinating advances in churning out tons of highly effective design assets. I’d expect the seniors in this space to follow their coding and marketing counterparts to become something more like a conductor, if they haven’t already. What This Means for AgenciesWe’ve been seeing agencies trend toward strategy (vs. execution) for years now. Five years ago, the argument was better pricing power. Now that argument’s changed to relevance. For those that have made the shift, there’s a lot of opportunity here with AI being more painful for your more execution-focused competitors. As execution becomes commoditized, the strategy and process for determining what to make will be even more valuable. The people who enable that are the senior practitioners. We’re already seeing evidence of shops reconfiguring themselves around a smaller team of experts, supported by contractors and AI. That may be the model going forward, but it’s still too early to tell, especially with how uneven AI’s capabilities are across the various agency disciplines. Hope this helps you better navigate these changes as they’re happening quickly. Until next time!
|
