The old agency equation has changed

For a long time, size carried an obvious promise.

A large agency could put more people on a project. More strategy hours, more design capacity, more production support. For a complex B2B company, that could feel reassuring. If the website was large, the market complicated, or the internal stakeholder group heavy, size looked like a form of safety.

Those advantages still matter. Established organizations have distribution, legal support, data, governance, and experience at scale. Nobody serious should dismiss that.

But AI is changing the equation.

The old advantage was coverage. More people meant more production capacity, more specialists, and more hands around the problem. But with AI disrupting how work is planned, produced, and delivered, the question becomes less about headcount – and more about how efficiently a team can adapt its workflows.

AI does not reward slow operational habits. Technology is changing, new tools appear every day, AI workflows are becoming normal, and ten-year-old best practices are being challenged. More importantly, client needs are shifting too.

At Oimachi, that shift is practical. We are two people, so friction shows up quickly. If a workflow feels wrong, we can change it, test a new tool, or build our own. We do not need a six-month transformation program to find out whether a better way of working exists. The implementation cost is much lower.

The advantage is how little distance there is between noticing friction and changing the system.

‍

AI adoption is no longer the interesting part

Most companies are using AI in some form now. McKinsey's 2025 State of AI survey reported that 88 percent of organizations use AI in at least one business function, up from 78 percent the year before. [cite:McKinsey|https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai|The State of AI, 2025. Used for broad adoption and workflow redesign context.] Stanford's 2026 AI Index points to the same broad pattern: adoption is high, while agent deployment remains early. [cite:Stanford HAI|https://hai.stanford.edu/ai-index/2026-ai-index-report/economy|2026 AI Index economy chapter. Used for organizational adoption context.]

The interesting question is no longer whether a company has tried AI.

The interesting question is whether AI has changed how work happens.

McKinsey's survey makes that distinction useful. Most organizations are still experimenting or piloting. About one-third report scaling AI programs across the organization. The high performers look different because they redesign workflows, scale faster, and use AI for innovation rather than pure efficiency.

That tracks with what we see in our own work. The teams that get the most out of AI are not the ones with the longest tool lists. They are the ones willing to look at the work itself and ask if the shape still makes sense.

Inside a design studio, that gets practical fast.

‍

Rapid editable workflows

One of the biggest changes in our own studio is that workflows have started to feel editable.

It may sound small. In practice, it changes a lot.

Years ago, if a process did not fit, the normal response was to search for a better SaaS tool, accept the compromise, or add a manual workaround. The work bent around the available software. A tool could solve 70 percent of the problem and create a new habit around the remaining 30.

Now, for certain internal problems, the question has changed.

Could this become a lightweight interface, a script, a cron job or a prompt system with enough structure around it that the output becomes reliable?

At Oimachi, we built our own project management tool Flowie, using AI-assisted coding and now use it every day. Not a grand software product. Not a SaaS startup in disguise. A tool shaped around how we want to run projects, document decisions, track finance, prioritize work and reduce the frictions that collect during a week.

It also replaced parts of our subscription stack.

The cost saving is nice, but control matters more. The workflow now fits our studio. When something feels off, we can change the tool instead of changing the way we think.

That shift is quiet, but meaningful.

At enterprise scale, replacing a workflow can mean procurement, security review, rollout planning, training, documentation, and months of alignment before the new thing touches real work. Some of that friction protects the organization from bad decisions.

In a period where the tools change every month, that friction also has a cost.

An independent studio can afford a different loop: notice friction, make a prototype, use it for real, keep it if it works, remove it if it doesn't.

‍

‍

The layoff story is too shallow

AI is often framed as a story about a need for fewer people.

The evidence partly explains why. Companies are trimming teams, reducing costs, and trying to show investors that AI creates efficiency. Gartner reported in May 2026 that about 80 percent of organizations piloting or deploying autonomous business capabilities had reported workforce reductions.

Gartner's conclusion is more interesting than the headline. The research found that workforce reductions did not appear to translate into ROI. Their blunt line is worth sitting with: "Workforce reductions may create budget room, but they do not create return." [cite:Gartner|https://www.gartner.com/en/newsroom/press-releases/2026-05-05-gartner-says-autonomous-business-and-artificial-intelligence-layoffs-may-create-budget-room-but-do-not-deliver-returns|May 2026 research note. Used as tension against layoffs-as-ROI narratives.]

Plenty of AI conversations skip that part.

Fewer people can mean less cost. They can also mean less memory, less craft, weaker accountability, and fewer people who understand why the work took a certain shape in the first place.

The better AI question is not how low the headcount can go. The better question concerns both the quality of the system around the remaining people and adoption rate of new technology.

For a studio, this matters because the goal is not to remove the designer from the design process. The goal is to remove the dull weight around the work: repetitive setup, meeting documentation, tedious text formatting, asset naming, bug-finding, CMS mapping, internal admin. Just to name a few things.

When those parts improve, the human work becomes more important, not less. Taste becomes more visible. Judgment has more room. Direction matters more because production becomes less of a bottleneck.

Without taste and discipline, AI helps a studio create more average work faster. It can chase every new tool, mistake output for progress, and flood its own process with options nobody has properly judged.

The advantage appears when speed meets standards, and new tools can change the work itself.

That is the difference between making better work faster and making average work louder.

‍

The first move matters more now

A 2019 Nature paper looked at more than 65 million papers, patents, and software products from 1954 to 2014. Its finding was simple enough to be memorable: smaller teams tended to disrupt science and technology with new ideas, while larger teams tended to develop existing ones. [cite:Nature|https://www.nature.com/articles/s41586-019-0941-9|2019 paper on small teams and disruptive work.]

That does not make tighter groups morally superior or magically smarter. The paper also makes clear that both modes matter. Scale can develop and spread an idea. A tighter group is often better suited to the first move, the unexpected move, the move that does not yet look safe enough for rolling out to hundreds of employees.

This distinction feels relevant in 2026.

AI has disrupted traditional workflows. Research, design, development, QA. None of these areas are unaffected.

That is where a studio like Oimachi can compete.

Not by having more people.

By keeping fewer layers between an idea and a working prototype.

‍

Some things shouldn't be automated

The most progressive independent studios in the AI era will do more than use AI tools. They will redesign themselves around the new possibilities. However, It might also mean being more honest about what shouldn't be automated.

Client trust. Strategic judgment. Taste. The ability to understand a brand as a whole. The sense that a website feels credible or not. These are not mechanical tasks.

AI changes the production surface around them. It does not remove the need for people who can make those calls.

The advantage means more than "we are faster." Fast can be sloppy. Fast can be shallow. Fast can be wrong at scale.

The better version is this: the studio can change its operating system while keeping senior judgment close to the work.

For B2B companies choosing a partner, that should matter. The coming years will reward teams that can keep rebuilding their own methods to make room for the work they truly excel at.

‍

If this is the moment

If there has ever been a good time for a small studio to compete with big agencies, this is probably it.

Big agencies are not finished. They will still win work where procurement comfort, global rollout, and large program management matter.

For companies that need agility, fast workflow adaptation, AI-assisted production, technical implementation, and a website or brand system that an in-house marketing team can maintain, a small focused studio has become a much more serious option.

We do not want to become a large agency. We want the work to stay close to the people making the decisions. We want to notice opportunities and change the tool or workflow without having to ask for permission.