The Canary Has Thoughts
We often say digital is infinite, and all we mean is it can feel opaque and unbounded. It originally depended on constraints to set its limits, and scope was usually the primary one. And this was good thing! Scope meant you could only build projects where your small team and ours could see all the edges. They were understandable, easy to maintain and easy to fix.
There's a budget. Budget buys developer hours. Hours determine which features, pages and integrations make it to launch. Everything else went on a list for v2, which, tbh, often never came.
There's just no doubt anymore that equation is changing. Northern Ground has built digital products for a long time. Meticulous marcom sites for beloved local brands to massive platforms for tentpole global enterprise. The shift happening now with AI is the largest change in production we've ever seen.
The good news: code is getting cheaper
Agentic development has cut the time it takes to draft, test and revise code. Work that once took a team weeks can now take days. Prototypes that used to be a separate budget item are now part of the first conversation. We can build three versions of an interface and test them, instead of arguing over giant post-its.
For clients, this means budget goes further. Features that seemed too ambitious or indulgent before can now stay in. Internal tools, content systems and integrations that were out of reach for mid-sized projects are now practical. Scope just no longer needs to be the main limit on what a project can be.
The bad news: cheaper code ain't better
- 01
Mercurial workflows
Models, harnesses, IDEs, design tools and platforms change every month. What was best practice in Spring is outdated by Fall. And agentic tools are becoming more and more black box but we still need to understand and test everything since we're responsible for every piece and outcome. Knowing which new concepts are reliable, which are beautiful mirages and how to remain masters of our tools of production takes daily attention, education and experimentation.
- 02
Brand strategy
When almost anything is possible (and at the risk of quoting Jurassic Park) the question becomes not if something can be built, but whether it should be. Ten agents can generate a hundred options. But they work only from the brief they are given. They cannot sit in the room and notice that the real challenge is elsewhere. They have no stake in the outcome. And of course they agree with whoever is asking.
A good strategy team can tell you when an idea is wrong. The relationship matters more than it sounds. People are candid with partners they trust and enjoy working with, and good strategy depends on that relationship. A faster project timeline means more decisions, more quickly, more often. That makes the people behind those decisions more important, not less.
- 03
Architecture
AI tools can and will produce software that's a lot like my love life: it looks fine on the surface but is a total mess underneath. Insecure, brittle, unable to scale. Code may be cheaper, but a weak foundation costs more than ever because teams now build on top of it faster. Decisions about data, security and system design need engineers who have both seen what breaks and why and who know where AI-generated code tends to fail: invented dependencies, outdated security patterns, logic that passes tests but breaks under real use. This is where decades of experience matter.
Matt RetzerExperienced developers can now develop quicker, which results in more robust applications then previously possible within the project scope, while still maintaining good architecture, security, maintainability and scalability. You can finally have all three legs of the stool: cost, quality and speed.
What this means for you
These three shifts are not specific to software. They are the pattern. Wherever AI makes production cheap, the same things become scarce: fluency with fast-changing tools, judgment about what is worth making, and the structure that keeps the output secure and durable.
A law firm will feel it in drafting. A marketing team will feel it in content production. A finance group will feel it in analysis.
If your industry has not felt this yet, it will. The organizations that do well will move their people toward judgment and structure before they are forced to.
The question to ask a digital studio has changed
It used to be, how much can you build for this budget? Now it's: can parts of running my business be made faster and cheaper? How do you decide which of these new tools can help me? If we build something together, how do we make sure it lasts?
And our best shot at an answer: Tools are still tools. We use the fastest and best available, and we put our people where they matter most, in strategy, design judgment and architecture. The result is more ambitious work, delivered sooner, for less cost, on a foundation built to grow.
If you have a project that was cut down to fit a budget in the past, it may be time to look at it again.
Take it from a canary. It's about to get interesting.