Amith Parameshwara

Every AI lab is hiring "Forward-Deployed Engineers." It's the most hyped role of 2026, and it still won't fix the full AI-led transformation in firms. Mercenaries win battles; they do not build nations. The "FDE" is having its moment. Google Cloud's Thomas Kurian announced hundreds of new FDE hires this month. OpenAI just launched an "OpenAI Deployment Company" built around them. Box CEO Aaron Levie calls it the most in-demand job in tech. Palantir Technologies, where the title was coined in 2011, must be quietly amused. 1️⃣ This isn't new. IT services have shipped engineers to client sites for four decades, solution architects, delivery leads, and embedded consultants. The technologist who codes by day and translates business needs by evening is a familiar figure. What's changed is the marketing, more than the operating model. 2️⃣ What is genuinely different? Today's FDE is an elite generalist, equal parts product thinker, applied researcher, and business interpreter. They aren't scaling staffing pyramids; they're collapsing them. One person now does what a discovery workshop, a business analyst, and three developers once did. That compression is real, and it is consequential. 3️⃣ Why this still won't close the last mile? AI projects don't fail in the demo. They fail six months later, when the model is technically working but no one in the business actually uses its output. The bottleneck isn't engineering. It's workflow design, incentives, accountability, and trust. Worse still, everything works with the AI solution, and employees are using the results. Yet the cumulative tangible results are not reflected in the firm’s P&L and balance sheets. This won’t happen unless the organisation uses AI for deeper transformation. None of which a brilliant outsider can rewire on a 12-week engagement. The last mile is owned by the client, not the vendor. 4️⃣ The dual shock to IT services. Demand for FDE-style engagements will rise sharply. But IT services firms will, for the first time, face direct competition from product and platform companies fielding their own elite talent. ----------------------------------------------------- I write at the intersection of AI, Strategy, and Economic theory

@JayaGup10 This is because this is one of the few pieces that analyse the deeper impact of AI on organisations. Good job @JayaGup10

Wow didn’t think this would travel this far and people would put in their 2 weeks after reading this but no better time to do so!!!

Anthropic’s new AI services company is not detrimental to consulting firms. It is almost the opposite. It's a signal to the market that consulting is more needed than ever to support organisations in their AI-driven transformation. Most people have been wrongly prophesying that AI is a consulting-killer. This is because of 1) an optimistic but wrong view of how complex/simple organisations are; 2) another wrong view of how easy/messy the transformations are; 3) limited understanding of the micro and macro impact of AI on firms, industries, and economies. Consulting firms will surely undergo their own transformation, but they will thrive.

This is Part 1 of three. Part 2 looks at what AI is actually doing to the structure of work, and the three very different futures unfolding inside every enterprise right now. Full essay at linkedin.com/feed/update/ur…

So the diagnosis: Tenure is not a moat. The most dangerous AI strategies are the ones that look like they are working. The part of the prize that compounds sits where the income statement cannot see. Time is not the moat. Depth is. And depth is built, not aged.

I've spent five years studying how AI translates into firm performance across hundreds of organisations. The pattern is uncomfortable enough that most leadership teams refuse to see it even when shown the numbers: Time spent with AI is not what makes AI valuable.