The models are ready, but the workforce isn’t.
Deloitte’s State of AI 2026 report says the share of companies with 40% or more of their AI projects in production is set to double in six months. But the same report flags an “AI skills gap” as the biggest barrier to making it real. That means the technology isn’t the problem. It hasn’t been for a while.
The effort really to see and really to represent is no idle business in face of the constant force that makes for muddlement.
Henry James, Prefaces: What Maisie Knew
The limitation is the number of people who can take what AI can do and reimagine how a business operates. And that is why the forward-deployed engineer is suddenly the hottest role in tech. This role is not a researcher or prompt engineer. It is someone who is embedded with customers to learn their actual workflows and constraints. Then builds something to work with those hard constraints and implement a new workflow.
This person is part engineer, part consultant, and a partner the business trusts. Salesforce is hiring 1,000 of them. Palantir, Ramp, and Cohere are all following suit.
Anthropic and OpenAI are building billion-dollar joint ventures around the same idea. The race isn’t to build better models. Or at least not quite as much. It’s to close the gap between what the current models can do and what companies can actually deploy.
Are you building that capability inside your organization? Or waiting for someone else to bring it to you?