{"id":62881,"date":"2025-11-11T15:14:25","date_gmt":"2025-11-11T15:14:25","guid":{"rendered":"https:\/\/www.elixirr.com\/en-us\/?page_id=62881"},"modified":"2026-07-29T16:13:16","modified_gmt":"2026-07-29T16:13:16","slug":"latest-ai-insights-you-cant-miss","status":"publish","type":"page","link":"https:\/\/www.elixirr.com\/en-us\/latest-ai-insights-you-cant-miss\/","title":{"rendered":"Latest AI Insights you can’t miss"},"content":{"rendered":"\n
\n \n
\n \n
\n \n Home<\/a><\/span><\/span> <\/nav>\n <\/div>\n<\/section>\n
\n
\n
\n

\n July 2026 AI Insights<\/h1> <\/div>\n \n <\/div>\n\n \n \n
<\/div>\n \n \n <\/div>\n\n \n \n \n <\/div>\n<\/section>\n\n\n\n
\n
\n

There’s\u00a0a conversation I keep having, in different rooms, with different companies,\u00a0roughly once\u00a0a month. Should we do this one first, or that one? Do we go here, or over there? We just had a reorg. We need to hit these numbers. Good people, real budget, real intent. Thirty days later, same questions.<\/span>\u00a0<\/span><\/p>\n

In those same thirty days, Sam Altman said\u202f<\/span>we are in the singularity, this is the moment<\/span><\/a>. His point is that\u00a0we’re\u00a0too early inside it to see it, but the\u00a0takeoff\u00a0has already started.<\/span>\u00a0<\/span><\/p>\n

I mostly agree with him. Which makes that monthly conversation harder to sit in, not easier.<\/span>\u00a0<\/span><\/p>\n

July was more of the same. New models, more capability, a few new features. The drumbeat continued.\u00a0That’s\u00a0why\u00a0it’s\u00a0worth writing about. The compounding is the story now, and this month it pointed somewhere specific.<\/span>\u00a0<\/span><\/p>\n

There is a bigger version of this argument in our latest AI campaign<\/a>, The Seat AI Cannot Take<\/em>.<\/strong><\/p>\n

The shift isn’t that humans are protected from AI. It’s that judgment becomes more important as intelligence becomes cheaper. The work changes, but the need for leadership does not.<\/p>\n

The demo stopped being “write this email”<\/h2>\n

Two features shipped days apart, from the two labs that rarely agree on anything.<\/span>\u00a0<\/span><\/p>\n

Anthropic added\u202f<\/span>Record a Skill<\/span><\/a>\u202fto Claude. You hit a button, screen-record yourself doing a piece of work, and talk through your thinking while you do it. Claude turns that into a saved skill it can run again. No prompt engineering. No script.<\/span>\u00a0<\/span><\/p>\n

OpenAI came at it from the other side.\u202f<\/span>ChatGPT Work<\/span><\/a>\u202ftakes a goal and runs multi-hour jobs across Drive, Slack and Salesforce. Last week it learned to\u202f<\/span>log into password-protected sites<\/span><\/a>\u202fand keep going.<\/span>\u00a0<\/span><\/p>\n

Both labs are betting the unit of automation is\u00a0a whole process, not a task.<\/span>\u00a0<\/span><\/p>\n

Most companies are still scoping AI one task at a time. Map a step, count the minutes, build the case. That habit is two years old, and\u00a0it’s\u00a0how you end up with a pile of things too small to move a number anyone cares about.<\/span>\u00a0<\/span><\/p>\n

What\u00a0actually gets\u00a0captured is the judgment<\/h2>\n

The part almost nobody noticed is that the recording captures you explaining <\/span>why<\/span><\/i>. Not the clicks. The reasoning.<\/span>\u00a0<\/span><\/p>\n

Automation has always been hard in big companies for one reason. The person who knows the work\u00a0can’t\u00a0write the instructions, and the person who can write instructions\u00a0doesn’t\u00a0know the work. Demonstration closes that gap. What lived in a senior person’s head becomes something you can hold.<\/span>\u00a0<\/span><\/p>\n

Think about why coding raced ahead. GitHub. Millions of worked examples, public, well documented, with the reasoning attached. The models had somewhere to learn.\u00a0Knowledge work has no GitHub. Your firm’s judgment was never written down, which is exactly why the models\u00a0can’t\u00a0do it yet, and exactly why\u00a0it’s\u00a0the scarce thing.<\/span>\u00a0<\/span><\/p>\n

Most leaders assume their edge is how the work gets done today. It usually\u00a0isn’t, and most of it is inefficient anyway. The edge is the decisions, the\u00a0taste\u00a0and the direction your best people add on top. Write that down in a form a machine can carry, and the business changes shape.<\/span>\u00a0<\/span><\/p>\n

I heard versions of this all month. One manufacturer wants its own engineering agent rather than its software vendor’s, with decades of design judgment inside it. A bank described its two-year goal as moving people off building spreadsheets and slides and onto judgment. Inside our own firm, we took one person’s process for pricing, turned it into a tool, and got three days of a good manager’s work in about five minutes.<\/span>\u00a0<\/span><\/p>\n

So\u00a0here’s\u00a0the payoff. When people stop moving data between systems and\u00a0writing\u00a0reports, and start interpreting,\u00a0directing\u00a0and deciding at the speed the machine moves, you have a different business. Not a cheaper one. A different one.<\/span>\u00a0<\/span><\/p>\n

The data panic, and the answer\u00a0nobody’s\u00a0talking about<\/h2>\n

Every time I make that point, someone pushes back. Careful,\u00a0you’re\u00a0handing your context to OpenAI and Anthropic, and\u00a0they’ll\u00a0build the next best version of you. Satya Nadella\u202f<\/span>gave the argument a name<\/span><\/a>\u202fthis month.<\/span>\u00a0<\/span><\/p>\n

I think\u00a0it’s\u00a0overblown. Not baseless, but overblown.<\/span>\u00a0<\/span><\/p>\n

I\u00a0don’t\u00a0believe the labs are stealing enterprise data. I do think\u00a0they’re\u00a0mining what they see to work out what to build next. And\u00a0they’re\u00a0pulling companies in a single industry together, healthcare loudest of all, to pool data and build models for that sector. Those models then get sold to everyone, your competitors included. The terms and conditions say what they say, and proving a breach would cost you more than\u00a0you’d\u00a0get back.<\/span>\u00a0<\/span><\/p>\n

The reason\u00a0it’s\u00a0overblown is that the other half of the answer landed this month. Open models caught the frontier.\u202f<\/span>Kimi K3 shipped as the first open-weight frontier model<\/span><\/a>, and the gap to the closed labs is now about ten days.\u202f<\/span>Xi made\u00a0open source\u00a0China’s official AI doctrine<\/span><\/a>. Open stopped being the cheap\u00a0option.<\/span>\u00a0<\/span><\/p>\n

So\u00a0the answer is a hybrid, and\u00a0I’d\u00a0start building it this quarter.<\/span>\u00a0<\/span><\/p>\n

Use frontier models for frontier work. Hard strategic thinking, genuinely complex reasoning\u00a0and orchestrating everything else.\u00a0That’s\u00a0what the premium buys. Then run open models you control for the rest, tuned on your own information, at a fraction of the cost. Cursor found a\u202f<\/span>15x swing in cost from routing alone<\/span><\/a>, and Microsoft is\u202f<\/span>swapping frontier models for in-house ones at up to 89% less<\/span><\/a>.<\/span>\u00a0<\/span><\/p>\n

Two things follow that most people\u00a0haven’t\u00a0connected.<\/span>\u00a0<\/span><\/p>\n

You need a team you\u00a0don’t\u00a0have today. Someone\u00a0has to\u00a0choose, tune, host, route and\u00a0monitor\u00a0those models.\u00a0That’s\u00a0real headcount, and a useful reminder that AI is currently creating more jobs than\u00a0it’s\u00a0removing.\u202f<\/span>Tech CEOs have dropped the replacement story<\/span><\/a>\u202fthey were telling last year. One study found the heaviest AI investors grew headcount 10.2% in the two years after adopting it, with entry-level roles up 12%.<\/span>\u00a0<\/span><\/p>\n

And if\u00a0you’re\u00a0regulated, this month was a gift. I work with banks and insurers stuck for two years because the data\u00a0can’t\u00a0leave the building and the regulator wants explainability. Open models at frontier quality mean they can keep that control and finally move.\u00a0They’ll\u00a0pay for it in\u00a0expertise\u00a0instead of licenses. For most of them,\u00a0that’s\u00a0a good trade.<\/span>\u00a0<\/span><\/p>\n

What it looks like when the CEO decides<\/h2>\n

Against all that, one story from our own work this month.<\/span>\u00a0<\/span><\/p>\n

It started as an executive briefing. Ninety minutes, no commitment. The CEO left the room having decided something, which is rarer than it should be.<\/span>\u00a0<\/span><\/p>\n

That turned into a six-week sprint to reimagine two parts of the business from a blank sheet. Not automate them. Reimagine them. Those two are now heading for live AI-native platforms in under ten weeks. On the back of it, the team has a remit to do the same across the rest of the company and find the next set of bets worth making, the ones with 10x to 100x in them rather than 10%.<\/span>\u00a0<\/span><\/p>\n

A few months from a briefing to a working platform. Nothing about that\u00a0required\u00a0new technology.\u00a0It required a CEO who understood what was possible and was willing to make a call, and a team that was ready to\u00a0operate\u00a0differently.<\/span>\u00a0<\/span><\/p>\n

I love stories like these!<\/span>\u00a0<\/span><\/p>\n

Which brings me back to that room<\/h2>\n

If judgment is the asset, a company that\u00a0can’t\u00a0make a decision\u00a0doesn’t\u00a0have one to capture.<\/span>\u00a0<\/span><\/p>\n

Across my client conversations this month, the problem is almost never ambition. It’s that the same foundational questions keep getting re-asked instead of answered. What do we do first? Who owns the number? What do we mean by operating model? Nobody refuses to decide. The decision just never gets surfaced, routed and closed inside the window the market gives you.<\/span>\u00a0<\/span><\/p>\n

One more thing\u00a0I’ve\u00a0noticed since about April. The first half of the year had real energy behind big swings and moonshots. Lately\u00a0I’ve\u00a0felt a flip back to “let’s just pick one small use case and get started.” I\u00a0can’t\u00a0tell whether people got overwhelmed or whether\u00a0it’s\u00a0the wider economy. Either way,\u00a0small things\u00a0cost about the same effort as big ones, and\u00a0the worst outcome is a pile of\u00a0small things\u00a0that\u00a0don’t\u00a0land, followed by everyone deciding AI\u00a0doesn’t\u00a0work.<\/span>\u00a0<\/span><\/p>\n

In case you missed it<\/h2>\n