Every (the company) knows what's up

Every started out as a publication back in 2020, and has since focused heavily on AI products, incubating their own products and advising companies on how to leverage new tools.

This week they published a post about how they built a 7-figure AI consulting business in the past year. It's long and worth reading in full, because on the key issues it's in 100% alignment with Subtle. Here are the things they are doing well:

1. The publication is, effectively, lead gen for the consultancy. But it's not half-assed "copy" optimized for search engines and partly generated by AI. It's actually really good writing, proven by the fact that Every is already itself a publication with seven figures of ARR (source for this was Dan on Twitter; I misplaced the tweet). The regular insightful coverage of AI products, how people use them, and announcement of their own original products attracts interest.

2. They earn credibility by building their own products. It's one thing to be an AI commentator, going on podcasts and the like. It's another thing to be getting your hands dirty, actually shipping product and learning, seeing what products people actually use. As they say, "Be a practitioner, not a management consultant."

Practitioners see what frameworks miss. You learn how a tool works, where it breaks, how people use it, and what makes it stick. That firsthand experience is your best diagnostic tool. It helps you move faster, explain more clearly, and spot what matters before the spec sheet tells you.

3. They are smart about, for lack of a better word, exploiting the situation that many companies find themselves in with respect to AI. That is, there's a rising awareness that knowledge work is being disrupted by LLMs and LLM-powered products, but there's no playbook. Every's consulting business has gleaned several nuances of this problem set in their first year:

We assumed the hardest part of adopting AI would be technical. Until now, software has been deterministic: You click X, and Y happens. But AI is non-deterministic. You need to manage it, not just use it. As a result, the real blockers to AI adoption are cultural.

Imagine a marketer who has one hour to write a campaign brief. They can either use that hour to do it the way they always have—manually, with help from templates and a tried-and-true process. Or they can take a risk: Spend that hour experimenting with a tool like ChatGPT, hoping it helps but worrying it might not. Most choose the first option—not because they're uninterested in AI, but because their job is to produce results, not to experiment. When time is tight and expectations are high, reliability wins. Ironically, a prompt written and refined today could lay the groundwork for a faster process tomorrow, but in the moment, it's scary to trade certainty for potential.

AI adoption doesn't happen just because a tool is available. It happens when teams have the space, the support, and the cultural permission to try something new. Play, believe it or not, is the greatest unlock for AI adoption.

People are super busy and barely keeping up with their normal job. They're behind on work and racing to hit deadlines, prep for important meetings, while juggling the rest of their life. They do not have time to figure out how to reinvent their personal working style, much of which has been honed and habituated over years or decades. Hiring a consultancy like Every (or us) creates the space in which to do this with guidance from experts (insofar as there are experts in this nascent field).