kfk

At least in Enterprise IT, where I work, the challenge is not how you scale technology, but how you scale trust. Big IT problems are always trust problems. That messy CRM systems everyone complains about? The messiness is likely due to low organizational trust, departments working in silos, and disagreements on definitions (what is a customer?).

LLMs can scale technology, but they are terrible with scaling trust. You know what scales trust in Enterprise? Good leaders, good culture, good conversations, good meetings, good narratives, good insights, good decisions. Exactly the things that scaled trust 50 years ago, and exactly the things LLMs cannot directly impact.

But now see what is happening. Rather than doing the hard work of gaining trust and consent, employees are using LLMs to build their localized agents. Of course they are happy! But the Enterprise, as a whole, is not better off, the real problems have not been addressed, the consensus has not been reached.

You might be right but you are missing a lot of nuance here. For instance, yes, Italy is not thinking about growth, true. But Poland? Poland is all about growth, they just made the list of the richest 20 countries in the world.

The real problem here is that EU as an economic block is much less integrated than people think. Pensions? Not integrated. Health insurance? Partially integrated. Exit taxes? A complete mess. Languages? Try speaking English or German or French in Spain. Etc.

EU has demonstrated that you can have local identities (I feel more Neapolitan than "Italian", for instance) and one economic block. Unfortunately, the economic block integration is not as deep as you might expect.

This is all fun and games when you work with toy data samples. But most organizations are more complex, they have to match invoices from SAP with opportunities in Hubspot; or they have to consider that little sales territory exception for the sales guy in Munich to calculate the proper commission projection; or they have custom tables in Salesforce with 0 documentation; or... you get my point.

Not all context is documented, and some context has to even be changed because it doesn't make sense.

I find AI very useful, but I think a lot of this AI SQL products are misleading.

Can you make chinotto?

how do you handle intent orchestration? I see you have workflows, but imagine this is used in combination with other MCP servers, how do you make sure the prompt is sent to the right MCP server and that the right tool or chain of tools gets executed?

How would you market such a business in 2026? I am from an Italian region where farmer grow many special coltures, and I was always a bit surprised why they don’t try selling on the internet. I ended up convincing myself it is not a viable business model.

I met many programmers during the boom years of software that straight out refused to develop any type of soft or managerial skills. Forget that, they even refused to maintain good relationships with decision makers (and I did this too, but only once in my carrier), left jobs in bad ways, focused on chasing salary increases every 6 months.

And here is the problem. If you have been chasing "easy" salary increases, working only on the comfortable stuff like developing tech skills, you should have seen this coming. It's very, very, very hard to maintain sharp coding skills decade after decade. Even if the job market was good, the reality is that you will eventually end up with a set of tech skills that a kid 20 years younger than you, with no family and so being able to live on lower salary, probably has too.

Because in the industries I have experience in IT has typically very little to do with the type of efficiency that matters in manufacturing. For instance, moving a factory from Germany to Romania is going to give you faster and more predictable savings than, say, optimizing production with IT and automation (obviously take this example with a grain of salt and the usual "it depends").

The theory was that Industry 4.0 should somehow change that, at least in Europe, by adding more automation and tech to production. I have seen only 1 factory built that way and, let me say, it was a complete mess IT wise. There is also the issue that the same people that were building factories in the 90s are now also building the "new" factories.

I run a small IT consultancy in EU mostly for companies with factories, so I have some biases. I have a few issues with the way IT leadership in EU has been running things the past 20 or so years:

- Absolutely 0 care for having a de-risked supply chain. In fact, IT leaders are extremely happy to have fewer and fewer suppliers, I think it is even one of their goals! And look at it now, what to do when 70% of your company runs on Microsoft and this happens?

- Buy always, no matter what the process is, just buy more tools. So what could have been 1 Python script now is a 5 years contract with yet another US supplier, all data stored in proprietary formats locked under complex APIs of course

- Bundle everything, and I mean everything, in the ERP. Make the ERP so big and complicated that adding anything to it requires tens, or even hundreds, of thousands of euros and multiple months of "development" (as an aside, did you know SAP ABAP code is stored in a database???)

- Lock yourself up completely with whatever Cloud provider you decide to use. Using AWS? Let's do everything with lambdas! Because, who cares about being Cloud agnostic and de-risking your AWS investment?

- Never invest in your internal tech talent, always go after shiny new tech solutions that deliver at best 20% of what they promise, while good motivated employees could have delivered 80% at a fraction of that cost

- Never push back on business asking for specific tools just because the vendors of such tools are amazing at marketing. 90% of manufacturing companies could replace Salesforce with much simpler tools (who knows, maybe even EU based?) and save millions. But no, let's go after brands and never consider the actual 1) business process we are trying to improve; 2) the reference architecture; 3) the underlying data we want to do CRUD on

The re-thinking of the tech stack is not a US tariffs issue, it is an IT leadership problem, and a serious one. The overwhelming lack of understanding of simple risk management strategies has gotten us here, EU companies should never, ever, have put themselves in this position in the first place.

I wonder how much of this drop can be simply explained as having reached a certain industry maturity. I assume at some point most of the fundamentals are in place, so we need less people?

You can also just use google sheets if the size of your data allows it.

The challenge with Microsoft in Europe is that it is so convenient, it doesn’t make business sense to consider alternatives. See how easily MS won market share over Slack (MS Teams) and PowerBI (Tableau / Qlik). They have such a big bundle of services that any single player has to be either amazingly good, or specialized, to win maybe 1-2% of market share. The only way I see Europe doing something about it is antitrust laws that break the bundles.

I struggle to make the VC math work on Substack. They take a 10% cut on paid subscriptions. They need 1 billion in paid subscriptions to have a $100m revenue, which at a hopefully 30% ebit would put them in the $0.5-1 billion evaluation (maybe). Now, at an average of $200 of paid subscribtions, it would take 5m paid users to achieve a 1 billion revenue.

If we assume a very generous paid / non paid user ratio of 10%, it means Substack needs 50m total users to have 5m paid users.

You tell me, that with 50m users, Substack would be content to make “only” $100m per year?

The most successful use pf generative AI is a chatbot… once you can use a model with natural language you need less rich UIs. I think UIs will become simpler and value will be in the proprietary data you can manage to grab to adapt existing models to some valuable use cases.

Not a direct answer, but I am targeting data marts from ERPs and other Enterprise applications like CRMs. I think data marts (data warehouses) are very valuable but they are too expensive and hard to build, so an AI that could generate the sql for the marts directly from the apps could be very valuable. What do you think?