🔴 AI: The Battle to Disrupt Professional Services Begins

With one trillion dollars poured into AI, between infrastructure and startups, what comes next?

The new debate is about how to allocate that capital: keep investing in infrastructure (models, tools, compute), or start funding AI-native companies, built from the ground up around AI?

Sequoia puts the question bluntly: did we make a mistake by trying to replace cloud and SaaS with AI, when the real opportunity lies in disrupting professional services, a market estimated at $20 trillion?

The question is a strategic one, because the AI infrastructure model, while still central, has become structurally loss-making.

The United States, with Microsoft in the lead, is entering a cost-cutting phase. That shift could come at Europe's expense, just as Europe has invested massively, not least at Emmanuel Macron's urging.

But to what end?

AI may now be part of our daily lives and our workflows, but the next stage, agentic AI, will take longer to emerge.

And yet, as I discussed in a previous newsletter, the technical infrastructure is already in place: Anthropic has launched MCP; Google offers A2A, a protocol for agent-to-agent communication.
Two building blocks that will finally let heterogeneous systems work together.

But stacking up technology does not make a business model.

With a trillion dollars invested (more than the entire SaaS industry's revenue in 2024),
what we have, for now, is a technology that is still barely integrated into business processes.

Sequoia reckons the opportunities created by AI are at least ten times larger than those of cloud computing.
AI does not merely replace tools or software: it also goes after services and labor budgets.

The next battle will be fought over business use cases, not just infrastructure.

There are legitimate questions to ask.

The world of computing remains rooted in deterministic logic.
Nobody wants to plug systems that hallucinate into critical processes.
The hype has died down. Nobody in the corporate world can get away with invoking AI innovation indiscriminately anymore.

Today, two broad approaches coexist:

• AI-augmented models, where humans remain central and AI plays the role of copilot.
• AI-native systems, where humans and intelligent agents collaborate seamlessly.
  Picture organizations where 10 people run 1,000 agents capable of producing equivalent work.

It is this second approach that is increasingly drawing American VC funds.
They are now targeting the AI-native market, with the ambition of rebuilding professional services around AI.
The point is no longer to automate a task, but to rethink the organization of the company from top to bottom.

This opens the door to a new generation of players able to compete with consulting firms, law firms and private equity funds.
In the jargon, these are called vertical agents.

The trend has accelerated with two recent announcements: Duolingo and Shopify going AI-first.

The rest is for subscribers