Meet Skai’s Forward Deployed Engineer: Your Partner for Building With Agentic AI

This number should stop anyone who’s building AI products in their tracks: according to MIT NANDA’s The GenAI Divide: State of AI in Business 2025, 95 percent of enterprise GenAI projects showed no measurable business return, despite $30 to $40 billion in enterprise investment. Only 5 percent of pilots ever turned into something a team used day after day.

My take is simple: that number is about the support that teams receive while they learn to use the product and who’s in the room with them as they do. That’s the reason for FDEs, and it’s why you should care about who gets assigned to your account as much as which platform you signed up for.

I’m a Forward Deployed Engineer, or FDE, working on our new agent-native marketing OS, Skai Studio (currently in beta). The title is relatively new in our industry, though consulting firms and large language model companies have been running FDE teams for a while now. In practice, my team builds alongside you inside your actual workflow until you can run it without us.

What are 5 reasons to work with an FDE?

1. The odds are just better with a partner. MIT NANDA found that AI deployments built with an outside partner succeed roughly twice as often as internal-only builds, 67 percent versus 33 percent. Those odds make a strong case for having someone build alongside you.
2. We’ll work through your friction instead of leaving you with it. Security reviews, legacy systems, and internal approvals are what slow big companies down after a great demo. I’ve sat through enough of those to have an opinion: having someone whose job is to work through them with you can keep the pilot moving forward.
3. You’ll find out what you actually need beyond what you assumed at the start. Most clients can’t fully describe their own workflow until they say it out loud. I keep asking questions in that first conversation until the real picture shows up.
4. You get a workflow that’s actually ready to use. A training session leaves you with a tool and a lot of questions. By contrast, we build the actual Studio setup next to you until you can run it on your own.
5. What we ship together is built to last. We build for your brand’s practices and your team’s judgment calls, and send anything Studio can’t do yet straight to our product team so they turn what we’ve found into improvements.

Those aren’t just my opinions. The broader research points in the same direction:

• McKinsey’s 2025 State of AI survey found that only 23 percent of organizations have scaled an AI agent into everyday use, even after running a pilot.
• WRITER’s 2026 Enterprise AI Adoption survey found that 97 percent of companies deployed AI agents in the past year, but only 29 percent see significant ROI from them.
• Gartner projects that more than 40 percent of agentic AI projects will be canceled by the end of 2027 over rising costs, unclear business value, and weak risk controls.
• Budget allocation: roughly 70 percent of GenAI spend goes to visible, front-facing work like sales and marketing, while high-value back-office use cases stay underfunded. If you’re a client-side marketing team or an agency running back-office work, that split is exactly why I work across both.

Not My First Rodeo… or My First Build

I’ve been at Skai for eight years and counting. I was the first hire on the Skai Labs team, and for six years before this title existed, I built custom implementations for clients, the same work an FDE does now, just without the title. When our reporting MCP shipped, I built my own actions MCP on top of it because I wanted to act on the data in addition to reading it. I did the same thing for our Help Desk, and I proactively built both because they solved real problems I was dealing with.

You shouldn’t have to teach your FDE your own product while you’re also trying to learn Studio. I already know it. That’s the difference between working with Studio’s FDEs and working with whoever happens to be available.

What does working with a Studio FDE actually look like?

1. Before we talk. You write down what you actually do day to day, weekly, and monthly: the reports you pull and the questions you ask. A little prework translates into big results.
2. Discovery. Instead of asking you to fill out a form, we get on a live call for a conversation. This is the part we care about most, because you probably can’t fully describe your own workflow until you say it out loud. I keep asking until the full picture shows up, then (you guessed it) use a custom-built MCP to translate your manual processes into Studio workflows.
3. Build. We document our discovery, turn it into a priority list, and build with you in the room in real time. Some of it ships right away, while some of it becomes feedback we send to our product team.
4. Handoff. We stay embedded until you can run the workflow on your own, deliver learnings to your Client Success (CS) team at Skai, and then hand over the day-to-day. All the work is still yours.

Why this is bigger than one role

Any large language model can technically run an ad campaign if you wire up the right access. Running that campaign according to years of your brand’s practices, your agency’s judgment calls, and the maintenance that keeps a system working as publishers change the rules underneath you requires more than access alone. Studio is built to carry that. Getting you to that value quickly, instead of leaving you to spend months rebuilding it yourself, is the whole reason a role like mine exists.

If you’re weighing build versus buy right now, the ability of AI to do the work is only part of the decision. You also need to consider who sits inside the work with you while you figure out what “done” looks like. Access is the starting point. An embedded FDE is what turns that access into value you can actually point to.

Ready to learn more? Contact us to request a demo.

---

Frequently Asked Questions