If your business bought AI tools this year and nothing much changed, you are not alone, and you are not doing it wrong. Even the companies selling the software have quietly reached the same conclusion: a login and a license don’t transform anything by themselves. In the past two weeks, Amazon and Microsoft each committed real money and real headcount to proving that point, and a $1.14 billion IT services deal shows where the demand is actually landing.

Two Trillion-Dollar Companies Decide Selling Software Isn’t Enough

On June 30, Amazon Web Services announced a new internal group of AI-focused “forward-deployed engineers,” backed by $1 billion in internal resources. These are not salespeople or support staff. They are engineers who embed directly inside a client’s business for a fixed engagement, build the specific AI agents that company needs, and then hand over both the finished system and the know-how to run it. AWS framed the goal explicitly around outcomes rather than tools: customers should leave with working systems and the internal skills to keep innovating on their own, not just another subscription.

Two days later, on July 2, Microsoft followed with something bigger. It launched Microsoft Frontier Company, a $2.5 billion commitment staffed by roughly 6,000 engineering and industry experts, aimed at the same problem: getting AI actually working inside large organizations instead of sitting half-deployed. Microsoft’s commercial business chief, Judson Althoff, was careful to distance the effort from the now-familiar “forward-deployed engineer” label, calling it something bigger and more outcome-driven. Early partners cited in the announcement include the London Stock Exchange Group, Unilever, Land O’Lakes, and, notably, Accenture, one of the world’s largest consulting firms.

That last detail is worth sitting with. Two of the biggest cloud and software companies on earth are now building embedded, hands-on implementation teams that look a lot like consulting firms. And they’re doing it alongside, not instead of, actual consulting firms. OpenAI and Anthropic made similar moves earlier this year, each launching their own implementation joint ventures backed by private equity partners, valued at $4 billion and $1.5 billion respectively. The pattern across all four companies is the same: platforms alone don’t close deals or deliver results anymore. Hands-on execution does.

Where the Actual Money Is Going: A $1.14 Billion Case Study

While the platform giants were building their implementation arms, HCLTech quietly closed one of the largest technology contracts of the year. On July 3, the Indian IT services firm announced a $1.14 billion, 5.5-year agreement with an unnamed Europe-based Fortune Global 50 company to overhaul its digital workplace and enterprise network using AI. The goal, per HCLTech, is straightforward: better employee experience, tighter operational efficiency, and support for a broader transformation strategy that the client evidently could not execute alone.

This is the part that matters most for a mid-sized business owner watching from the outside. A Fortune Global 50 company, with essentially unlimited internal engineering resources and its pick of every AI vendor on the market, still chose to hand a multi-year, nine-figure mandate to an outside implementation partner. If a company that size needs hands-on help to make its technology strategy real, it is worth asking honestly whether your organization’s current approach, buy the tool, hand it to IT, hope for the best, was ever going to work.

Why the “Sell It and Walk Away” Model Stopped Working

For years, the software industry ran on a simple formula: build a good product, sell a license, let the customer figure out deployment. That formula assumed customers had the internal expertise, time, and organizational appetite to do the hard part themselves. Generative AI and agentic systems broke that assumption. These tools are powerful, but they are also new enough that most internal IT teams haven’t built the muscle memory to deploy them safely, integrate them with existing systems, and retrain workflows around them without outside guidance.

The forward-deployed engineer model exists precisely because the gap between “we bought the AI tool” and “the AI tool changed how we work” turned out to be wider than almost anyone expected. Some industry estimates suggest the large majority of AI pilots never make it into production use at all. That is not a technology failure. It is an implementation failure, and implementation is exactly the kind of work that gets shortchanged when a purchase decision is treated as the finish line rather than the starting point.

What This Means If You’re Evaluating a Software or AI Investment

Before signing off on any new platform, custom build, or AI initiative this year, it’s worth asking a prospective partner a few pointed questions. First: who is actually doing the implementation work, and are they embedded with your team or handing you documentation and a support ticket queue? Second: what does success look like in terms your finance team would recognize, not just “the system is live,” but a measurable change in cost, speed, or output? Third: what happens after the contract ends. Does your team walk away with real capability, or does the project quietly stall the moment the vendor’s engineers move on to the next client?

These aren’t abstract concerns. They are the exact questions that separate technology purchases that pay for themselves from the ones that end up as unused line items in next year’s budget review.

The Takeaway

When Amazon, Microsoft, OpenAI, and Anthropic all decide, within months of each other, that the winning move is to build embedded implementation teams rather than simply sell better software, that’s a signal worth taking seriously. The tools have gotten good enough. The bottleneck now is execution, and execution is a people-and-process problem before it’s a technology one. Businesses that treat their next software or AI investment as a partnership, not a purchase, are the ones most likely to see it actually pay off.

If you’re weighing a custom build, a platform rollout, or an AI initiative and want a second opinion on how to structure it for real adoption rather than a shelved pilot, we’re happy to talk it through. Reach us at info@kodevox.com or visit our contact page to start the conversation.

Sources and further reading:

• Amazon launches new $1 billion FDE org, following OpenAI and Anthropic — TechCrunch
• Microsoft launches its own AI deployment company with $2.5 billion commitment — TechCrunch
• HCLTech wins $1.14 billion AI transformation deal with European client — Domain-b.com
• Anthropic and OpenAI are both launching joint ventures for enterprise AI services — TechCrunch

— The Kode Vox Team