Your software stack is a generation behind. Here's the test

Most of what you're paying for in your software stack is about to be worth a fraction of what you pay for it.

Have a quick look at your stack. The CRM, the accounting platform, the project tool, the BI dashboard, the marketing system, the HR thing nobody uses. Most founder-led businesses I work with are sitting on 15 to 30 active subscriptions, which is its own conversation, but the sharper question is what those things are actually doing for you, and whether what they do is still where the value lives.

I've been watching B2B software evolve for about 25 years, mostly from the inside as a builder and operator, and the arc is cleaner than people give it credit for.

There have been four eras. We are standing right at the start of the fourth.

The first era was Process. The 90s and early 2000s, the great push to get workflows off paper and into systems, accounting, ERP, the early CRMs. Run the process, store the data, don't lose anything. MYOB, SAP, Salesforce in its early form. For its moment, revolutionary, but the win was operational rather than strategic. You stopped losing things, which is a lower bar than it sounds.

The second era was Data. Mid-2000s through the early 2010s, the cloud migration and the SaaS boom, when "single source of truth" became a phrase nobody could attend a board meeting without hearing. The real win was visibility. For the first time in business history you could see what was happening across your business in something close to real time, which sounds obvious now but wasn't.

The third era was Insights. The 2010s. Tableau, Power BI, Looker, the whole BI stack, and the promise quietly shifted from "here's your data" to "here's what your data means". Dashboards proliferated, BI became its own software category worth roughly $72 billion globally, and every other SaaS product on the market started bolting on an analytics layer because that was where the perceived value had moved.

This is where most businesses are still living. And frankly drowning.

The cruel joke of the insights era is that more information doesn't equal more clarity, quite often the opposite, and you end up with twelve dashboards, three of which are out of date, none of which are surfacing the one thing you actually need to do something about. You can't manage what you can't see, but it turns out you also can't manage what you can see if it's been smeared across a dozen tools that don't talk to each other.

The fourth era is Execution. The question has moved one final time, from what happened and what does this mean to what should I do, and can you just do it for me. The technology unlock was the LLM, because for the first time the reasoning step that used to require a human looking at the dashboard became cheap enough and reliable enough to be done by the software itself, and the consequence of that has been the fastest repositioning of an entire software category I've seen in my working life.

Bain's 2025 tech report puts it well, that AI is moving "from thought to action". Gartner is forecasting that by 2028, 33% of enterprise software applications will include agentic AI, up from less than 1% in 2024. Deloitte's most recent survey found 57% of large companies are putting between 21% and 50% of their digital transformation budget into AI automation. ServiceNow paid $2.85 billion in March 2025 for an AI agent platform, Salesforce launched Agentforce, SAP launched Joule Agents, and Workday, Oracle and Adobe have all rebuilt around agents in the last 18 months.

Every major SaaS vendor on the planet is repositioning around the same thing.

Bain's blunt assessment is that within three years most routine digital tasks will move from "human plus app" to "AI agent plus API", which means the app stops being the value and the reasoning sitting on top of it becomes the value instead.

Which leads to the question worth sitting with for your own business. How much of what you pay for stops at the dashboard? If most of your stack is data-and-insights tooling, which it almost certainly is if you're honest about it, you are paying premium prices for a layer that's about to be commoditised. The execution layer, the bit that recommends, decides, prioritises, and ideally just does the next action, is where the value has moved.

I should declare an interest. I'm building in this space with Highlander, our strategy-to-execution platform, so I'm not exactly a neutral observer. But the trend is bigger than any one product, and it's the entire direction of enterprise software for the next decade.

The practical action this week is small enough to do tonight. Pull up your software list and for each subscription ask one question. Does this thing tell me what to do, or just show me what happened? If the honest answer for more than half your stack is the second one, you're a generation behind where the market is about to be, and you'll be paying premium prices for it for as long as you let yourself.

Cheers,
Josh