Customers needing an FDE (median)
20→27→32%
2024 → 2025 → 2026 projected. The attach rate for hands-on deployment is climbing every year.
Vision
Vision & Mission
A world where buying AI software and successfully running it are the same decision — not two separate, unequally staffed problems.
Grapnel exists to close the gap between AI vendors' sales velocity and their delivery capacity — by supplying vetted, production-grade forward deployed engineers on an elastic basis, so a Series A vendor with eleven enterprise logos can deploy like a company ten times its size, without building a services org it never wanted.
The Gap
AI software is being bought faster than it can be deployed.
The deployment burden per customer is rising, not falling, even as the models get better.
Use FDEs as a hybrid role
44%
Bridging product and delivery — customising for customers and feeding insight back to product.
Accenture AI & data practitioners
77,000
At FY25 close, targeting 80,000 by FY26 — and still only serving the top of the market.
FDE job postings, YoY growth
1,165%
2026 vs. prior year. Demand for the role is outrunning every other hiring category in AI software.
Source: ICONIQ, “State of AI: The Execution Era of AI,” Jan 2026 · Accenture FY2025 10-K · Bloomberry analysis of 1,000+ FDE postings, YoY through Oct 2025.
Why the Last Mile Matters
Why the last mile doesn't compress.
Context is per-customer
Every deployment meets different data, different systems, different exceptions and a different definition of “correct”. None of it ships as configuration.
Agents act on live systems
An agentic product writes to the customer's stack. The integration and permission surface is far wider than a SaaS product that only renders a screen.
Buyers now buy outcomes
Pricing has moved from seats to results. Whoever owns the result owns delivery — and that is now the vendor.
Every AI product company is forced into a services business it never planned for — at exactly the moment it's trying to show product margin to investors.