You Are Less Behind Than Your Feed Says — Altvina Altvina Insights

Published July 10, 2026 · Altvina Insights · 5 min read

You Are Less Behind Than Your Feed Says

"Are we behind on AI?" is really three questions in one coat. Two of them can wait. The third was never about AI.

It's 9 p.m. and you're scrolling. Another founder just posted that their firm has "gone all in on AI." Someone in the comments agrees. You close the app and the question follows you to the kitchen: are we behind?

Here's the useful move: pull that question apart. It's really three questions, each with its own clock, and only one of them deserves your week.

Question 1: Are we behind the rules?

Short answer: no, and the rulemakers just proved it.

The strictest AI law in the world is the EU's AI Act. In May 2026, the EU agreed on a package that pushes back most of its own big deadlines:

  • Rules for most stand-alone high-risk AI systems, due August 2026, moved to December 2027.
  • Rules for AI built into regulated products, due August 2027, moved to August 2028.

One rule kept its date. Starting August 2, 2026, AI that talks to people has to say it's AI, and AI-made media has to be labeled. But read who that rule covers: firms that ship AI products or publish AI-made content to the public. A US firm using AI internally to draft documents that a human reviews and sends is not on that list. If your firm does ship AI products or publish synthetic media, that's a conversation for your counsel.

So the toughest regulator on earth just gave most of its own rules an extra year and a half. If they aren't rushing, your feed's daily countdown is not the real clock. Slower is not stopped, though. The conditions that end the waiting are below.

Question 2: Are we behind other firms?

A feed is a stage. Every "we went all in on AI" post is also an ad for the person posting it. Some of that confidence is earned. Some of it is performance. From the outside, you cannot tell which is which.

That makes a feed a terrible measuring stick. The only comparison that holds up is your firm this quarter against your firm last quarter. Everything else is your inside against someone else's stage.

Question 3: Are we behind on our own work?

This is the only urgent question, and notice: it isn't about AI at all. It's about where your work actually piles up. Here's a fifteen-minute way to find out, all of it right here.

Step 1: Write the wishlist. List everything you wish AI would take off your plate. Real items from your firm. For many firms it looks like: proposals, status reports, meeting notes, first-pass research, invoices, the follow-up emails that never get sent.

Step 2: Ask why a human still does each one. Push past the first answer. Almost every honest answer lands in one of three buckets:

  • No written process. The work lives in someone's head. No document exists that a new hire, or a tool, could follow.
  • Judgment only you hold. The work is mostly a decision, and you're the only one trusted to make it.
  • Volume outgrew the method. A good process exists, but it was built for ten of something and you now have sixty.

Step 3: Act on each bucket.

  • No written process? Not an automation candidate yet. Pointing a tool at an undocumented task automates the confusion. Write the process down first; a page is enough. Sometimes the writing alone returns the time, and the item leaves the list with no technology at all.
  • Judgment only you hold? That's a structure problem, not a software problem. The real question is who else could hold that judgment, and what guardrails they would need. The queue of work waiting on your decisions is the real cost.
  • Volume outgrew the method? These are your genuine automation candidates, for AI or anything else. Start with the smallest one, prove it works, then take the next.

What this usually leaves is a much shorter list: one or two real automation candidates, a few processes that need writing down, and one uncomfortable question about who holds the decisions. The wishlist was never really about AI. It was a bottleneck inventory the whole time. That shorter list is the deliverable. Nothing beyond this page is needed to act on it.

Three signs waiting has stopped being safe

Calibration cuts both ways. Waiting is fine, until one of these is true:

  1. People in your firm are already pasting client material into AI tools, with no shared rule about what client data goes where. That's not waiting. That's adoption without rules, and it happened without you.
  2. A client asks directly how AI touches their work, and nobody in the firm can answer in a sentence you would put in writing.
  3. The same volume-outgrew-the-method item has shown up three quarters in a row, and the queue behind it is longer each time.

If none of those are true today, waiting is a defensible position, not a lag.

Permission

If you ran the sort and found mostly process problems and one structural question, you are not behind. You are accurately measured, which is rarer and more useful.

So do less than your feed suggests, and do it more deliberately:

  • Write down one process this week.
  • Name one decision that should not need you.
  • Watch the three signs above.

You do not have to become an expert in any of this unless you want to. That is why Altvina exists: helping owners work through questions like these is the job, whether the answer turns out to involve AI or a plain process fix. If the one urgent question, where your own work actually queues, is sitting on your plate right now, a fit call is a short conversation to see whether it is something we should work through together: altvina.com/fit-call. And "not yet" is a fine answer.

Continue the series

This is part 5 of a 5-part series on Both Columns, Honestly. The full arc:

Content and Accuracy Disclaimer

This article was drafted with AI assistance and reviewed by the Altvina team. We rigorously fact-check all content to ensure reliability.

Should you notice any inaccuracies or outdated information, please contact us so we can correct it. Your feedback helps us maintain high standards of accuracy and transparency.