According to Volish corporate veep Charles Lamanna, business apps are relics from the mainframe era that will soon be replaced by “agent-native” platforms.

Lamanna told a Madrona podcast that web-based enterprise software hasn’t evolved much since the 1980s and likened current systems to legacy mainframes, still chugging along but increasingly pointless. He thinks the future belongs to AI agents that don’t follow rigid workflows or forms but instead adapt dynamically to users and goals, chatting in plain language and navigating vast data sets through vector databases.

He reckons the whole shift will be codified within the next 18 months and widely adopted by the end of the decade. That has raised more than a few eyebrows in an industry known for moving at a glacial pace when real money is involved.

Microsoft MVP Rocky Lhotka said the prediction is “very forward-looking and optimistic,” noting that industries like construction or logistics don’t tend to throw away working tools, people or processes just because Vole fancies something shinier.

The vision comes with structural changes too. Lamanna says departments will merge, workers will become generalists, and AI agents will be part of every team. If you need customer data, you won’t be logging into a CRM, you’ll ask your agent. The CRM might still exist, but it will be buried under a conversational layer that makes the app itself irrelevant.

Richard Campbell argues that this will redefine software. He reckons if large language models can interpret your comms and sales data, they can functionally be your CRM.

Not everyone’s convinced. Lhotka pointed out that AI isn’t deterministic, but accounting and supply chains are. You can’t have your agent guessing how much concrete to pour or what to put in a truck. And if most companies end up using the same agents, Lhotka warns innovation could grind to a halt because LLMs can’t invent, they can only remix.

Lamanna claims industry support is strong. He cites growing adoption of open protocols like MCP and A2A, which he says have seen the kind of alignment not seen since the days of HTML and HTTP. That is the kind of grand claim that tends to come before a standards war, not after.

Analysts like Brad Shimmin from Futurum Group are sceptical. He admits the idea of ditching Excel and dashboards for chat-based interfaces is tempting but warned that replacing app ecosystems entirely would be an operational and governance nightmare.

Microsoft’s long game appears to be killing off its software before someone else does. Whether that future arrives by 2030 or drags on another decade, it’s clear that Redmond wants to be first in line to declare SaaS officially obsolete, even if it’s still raking in monthly fees long after the funeral.