The "AI" You're Buying Might Just Be a Very Expensive Staffing Agency in a Lab Coat

The "AI" You're Buying Might Just Be a Very Expensive Staffing Agency in a Lab Coat

What if the AI platform you were about to pay for wasn't really AI at all, but a small army of engineers quietly doing the work behind the scenes, billed to you as software? According to Anurag Gurtu, co-founder and CEO of Airrived, that is not a hypothetical. It is how a growing number of enterprise AI companies are operating right now, and the gap between what they promise and what they deliver has billion-dollar consequences for anyone buying, investing in, or building with AI tools.

The practice Gurtu describes centers on what the industry calls Forward Deployed Engineers, or FDEs. These are highly compensated technical specialists who are embedded inside customer organizations to manually build the workflows, integrations, and automations that the AI vendor's platform was supposed to handle autonomously. Job postings for FDEs have exploded, with Lightcast data reported by Fortune showing an increase of more than 1,000% from January to August 2026. Indeed data tracked postings rising from 643 in April 2025 to 5,330 in April 2026. The pay is not modest: Anthropic lists base salaries between $280,000 and $320,000, with some total compensation packages reportedly reaching $400,000, and OpenAI lists base pay between $185,000 and $300,000, plus equity.

The financial architecture underneath this model is where it gets particularly important for buyers to understand. Gurtu lays out the pattern with precision: a vendor sells a customer an "AI platform," the platform does not function autonomously, engineers move in to make it work, and the customer is billed for software. That revenue gets booked as Annual Recurring Revenue, or ARR, which inflates the company's valuation for the next fundraising round. OpenAI itself pledged up to $1.5 billion into a deployment joint venture valued at $10 billion, called DeployCo, with more than $4 billion in total investment from private equity firms including TPG, Bain Capital, Advent, Brookfield, and Goanna. OpenAI guaranteed those private equity backers an annual return of 17.5% over five years, which Gurtu points out is a services-business structure, not a software growth thesis.

For small and mid-size business owners evaluating AI tools right now, this matters enormously. The enterprise AI market is being actively shaped by companies that have learned to use the language of autonomy and scale while quietly depending on expensive human labor to deliver results. When those costs are hidden inside a software subscription price, the buyer has no way to know whether they are paying for a scalable platform or an outsourced consulting team with a sleek interface. Gurtu offers a sharp diagnostic: ask the vendor directly whether the product would continue delivering its promised value if the embedded engineers were removed tomorrow. If the honest answer is no, you are not buying software. You are renting people.

The gross margin signal is equally revealing. Gurtu draws a clear distinction between two companies with identical $20 million in ARR. One requires 10 engineers to serve its customer base. The other requires 100. Both pitch decks say "AI software company." The difference, invisible until someone examines costs carefully, is that true software gets cheaper to deliver as it scales. People-powered ARR does not. It grows the org chart instead of the margin. For a small business owner being pitched on an AI platform, the practical question is: does the vendor's pricing model assume you will need ongoing technical hand-holding, or does the product genuinely operate and improve without constant staffing behind it?

The actionable move this week is to request a gross margin disclosure or a cost-per-customer breakdown from any AI vendor you are actively evaluating. Specifically, ask how the product performs without dedicated onboarding engineers present, and whether the first-year implementation cost is included in the software price or billed separately. The vendors building real, scalable AI tools will answer that question directly. The ones running a staffing operation in software clothing will redirect you to another demo.

Knowing the difference between true AI automation and an expensive team of engineers dressed up as a platform is not just due diligence. It is the foundation of a smart AI marketing and operations strategy built to scale.

Originally inspired by: The FDE Bubble: Silicon Valley's Most Expensive Game Of Pretend (https://news.crunchbase.com/venture/fde-bubble-enterprise-agentic-ai-gurtu-airrived/)

Your turn

What is your traffic actually doing?

Send us your details and we will come back with a short, specific read on what your traffic, your pages and your pipeline are doing today — and the first three things we would change. A real person reads every submission, and you get the read whether or not we ever work together.