AI Moats Move From Models to Workflows as Schneider Pays $22.6B

Schneider Electric's agreement to acquire PTC for approximately $22.6 billion in equity value exposes where artificial-intelligence vendors may build their next durable advantage: inside the customer's existing workflow, according to strategic adviser Itay Sagie.

What does the Schneider-PTC deal reveal?

The transaction hands Schneider software used to design, manufacture and service physical products, embedding the buyer in decisions about how factories, plants and machines get specified and maintained over decades. PTC sits at the desk of the engineer, not the consumer.

Synopsys, by contrast, has opted for partnership over purchase. Its collaboration with OpenAI pairs frontier model capability with established chip-design tools, supported by licensing and revenue sharing between the two firms.

Sagie's reading: both moves reward ownership of the environment in which AI performs useful work. Companies already trusted in complex industries carry the customer relationships, domain expertise and approval processes that AI developers need to monetize raw model output. The model alone does not sell itself.

For founders, the implication is to build a position inside a specific workflow before a larger competitor does. For incumbents with that seat already, it is a bargaining chip to defend or license.

How does ElevenLabs make the workflow case?

ElevenLabs reported its AI agents now handle more than 15 million conversations weekly — roughly 780 million a year — covering refunds, insurance renewals and healthcare bookings. The company separately announced an employee tender offer at a $22 billion valuation.

The volume matters more than the headline figure. Once an agent connects to internal systems, follows permissions and reliably closes tasks, replacing it requires migration, retraining and operational risk. Sagie argues that the switching cost, not the underlying model, is the durable asset.

A caveat belongs here: integration alone offers limited protection. Customers must depend on the product and measure the value it delivers, and a better model from a competitor could still break the lock-in if the workflow is shallow.

Why did ServiceNow buy Moveworks?

ServiceNow's acquisition of Moveworks folds an AI assistant and enterprise search technology into its workflow automation stack. At close, Moveworks reported 5.5 million employee users and roughly 250 customers that already used both companies' products.

The strategic opportunity is to route employee requests directly into the systems that resolve them across IT, HR and finance functions. Sagie frames the move as a template: identify a target workflow, understand the barriers, then buy or partner to enter it.

Diligence must test how deeply customers depend on the product, whether those relationships survive a change of control, and whether the combined offering delivers measurable improvement. An attractive target brings technology, customer relationships, integrations and a daily operational role that a standalone tool cannot match.

What the moat argument leaves unsaid

Sagie's framework treats workflow ownership as a defensive asset, but he does not address what happens if regulators, customers or AI providers themselves shift the rules. Nor does he price the risk that incumbents with embedded software grow slower than model-native challengers in fast-moving segments.

The biggest open question is whether integration depth or model capability drives purchase decisions three to five years from now. The evidence cited favors workflow, but the sample is small and the deals are recent.

Schneider's $22.6 billion for PTC, ServiceNow's undisclosed check for Moveworks and ElevenLabs' $22 billion employee tender together back the thesis that embedding beats modeling. Investors and boards should track second-year retention numbers, not launch press releases, before declaring the moat built.