The Autopilot era for enterprise software isn’t coming. It’s already here.

The line between software and services is blurring – and that’s probably a good thing. A lot of modern enterprise software will start looking more and more like an Autopilot.

“A copilot sells the tool. An autopilot sells the work.”

Here are 6 thoughts worth connecting on this topic:

SaaS → Services as Software. The Internet gave us Software-as-a-Service, or SaaS – pay a subscription instead of a stale license. Modern AI adds a twist: Services as Software. AI agents deliver completed services (the outcome), not just tools humans use to reach one.

So Palantir was right? For years, people accused AI software firm Palantir of being a professional services firm in disguise, with their ‘forward deployed engineers’ (FDEs) seen as glorified staff augmentation. Ironically, every major AI and data company is now standing up field deployed engineer teams, because AI automation that actually works is… hard. And FDEs dramatically increase successful outcomes.

Traditional PS firms adapting. McKinsey is reshuffling its workforce with 25% fewer back-office staff and 25% more high-touch client-facing experts. Accenture is discretely filling revenue challenges with M&A while they rethink their business. Meanwhile, Claude is building the board deck due tomorrow.

The next trillion-dollar company (per a recent Sequoia article) may be a software firm masquerading as a services firm. AI handles heavy-lift intelligence work, while humans focus on high-touch and ‘judgment’ work. “A company might spend $10K a year for QuickBooks and $120K on an accountant to close the books. The next legendary company will just close the books.”

The Complexity Cartel under threat. Mainstream enterprise software has normalized vanilla platforms that require long, painful, 7-figure implementations. As AI systems deliver work output, the pressure is on for legacy systems that can’t get anything done without a small army of consultants.

AI’s killer app isn’t better search: it’s writing code. Large Language Models (LLMs) are fundamentally good at two things: languages and reasoning. Since software is built from programming languages and logic, LLMs are prolific coding machines. This makes rebuilding enterprise software from scratch far more accessible.

More importantly, it’s shifting the market. Businesses that once defaulted to off-the-shelf software suites can now cost-effectively build custom apps and agents tailored to their exact needs. That’s a real competitive threat to established software vendors. To stay relevant, off-the-shelf products can no longer coast on convenience – they need to embed deep domain expertise and an ecosystem edge that custom-built solutions can’t easily replicate.

AND… they are expected to do more of the work – period.

That’s why a lot of modern enterprise software will look more and more like an Autopilot than a Copilot.

And for us humans, the world is still in short supply of creativity, judgement, and consultative subject matter expertise!

Stop co-piloting. Get a true Autopilot. See Maxa in action