Artificial intelligence has been hailed as the future of business, but reality paints a different picture. According to a recent Massachusetts Institute of Technology (MIT) report, nearly 95% of companies investing in AI aren’t generating profits.
The problem isn’t a lack of ambition — it’s misunderstanding how AI truly works. Jason Droege, CEO of Scale AI, believes the issue stems from inflated expectations and misconceptions about AI’s real-world application.
“There’s been this general promise that you can just plug in the model and everything will work,” Droege explained. “The reality is a little bit different.”
Rethinking AI’s Role in Business
Scale AI built its reputation by supporting large-scale machine learning systems with high-quality labeled data. These massive datasets are what allow large language models (LLMs) to distinguish between something as simple as a photo of a cat or a fish. For years, tech giants relied on Scale for precisely this — structured data that teaches AI to recognize, classify, and respond accurately.
Instagram | alexanddeer | Meta acquired 49% of Scale AI, prompting founder Alexandr Wang and senior executives to join their team.
This business model caught the attention of Meta, which acquired a 49% stake in Scale AI in June for $14.3 billion, valuing the startup at $29 billion. Following the deal, Scale’s founder Alexandr Wang and several senior leaders joined Meta. However, the acquisition raised eyebrows across the AI industry. Competitors like OpenAI and Google reportedly reduced their collaboration with Scale due to concerns over Meta’s influence.
Despite this, Scale’s data labeling division continues to grow every month. Now under Droege’s leadership — who first joined as Chief Strategy Officer — the company is expanding its mission. Beyond supplying organized data, Scale now helps businesses develop their own AI systems to streamline operations, automate repetitive tasks, and reduce inefficiencies.
“I think companies thought it was easier than it actually is,” Droege said. “But there is a ton of value when you get it right.”
Why Most Companies Miss the Mark
While executives across industries talk about AI’s potential to enhance productivity, most aren’t seeing measurable results. MIT’s findings suggest many corporations are chasing the wrong problems with AI. Droege argues that businesses often treat AI like a magic wand capable of solving everything.
“AI isn’t a fix-all,” he said. “It’s most useful when people are slow, inconsistent, or prone to errors.”
The right approach involves identifying tasks that benefit from automation, such as:
1. Reviewing and summarizing lengthy documents
2. Processing insurance claims
3. Compiling medical histories for physicians
Scale has already helped organizations design tools that handle these challenges. For example, its AI systems assist doctors by summarizing patient information before consultations and help insurers evaluate claims faster and more accurately.
Human Expertise Still Matters
Despite AI’s growing sophistication, Droege emphasizes the importance of human oversight. “If a healthcare organization builds an AI tool to assist doctors, you’d want your most senior medical professionals giving feedback, identifying errors, and refining the system,” he said.
This collaborative process takes time — often weeks or months — but results in tools that truly assist rather than replace human experts. Government agencies are already adopting similar strategies, using AI to pre-screen building permit applications before human review, significantly cutting down approval times.
Turning Potential into Profit
Instagram | indianstartupnews | The AI application market is intensely competitive, with Amazon and Microsoft fighting for dominance.
Skeptics argue it could take years before AI investments generate consistent profits. Gil Luria, head of technology research at DA Davidson, noted that implementing enterprise-wide AI tools is a long game: “This is going to take years for large companies to implement AI tools that are broadly useful and generate revenue and save expenses.” Yet, he added that when done right, these systems will deliver “tremendous” financial impact.
Competition in the AI application space is fierce, with major players like Amazon and Microsoft vying for dominance. Luria described Scale’s evolution from a pioneer in data labeling to one of thousands now competing to build business-focused AI tools. Still, Droege sees the demand for specialized, problem-oriented AI solutions as far from saturated.
The Path Forward for AI Integration
MIT’s research reinforces Droege’s stance: companies that succeed with AI often collaborate with external experts rather than trying to develop tools entirely in-house. Understanding which business functions AI can realistically improve is key.
Droege remains optimistic about Scale’s dual focus — data infrastructure and applied AI. “Our applications business is already generating hundreds of millions in revenue,” he said. “And our data operations have grown every month since the Meta deal. It’s a large business for us, and we’re very happy with it.”
AI isn’t failing companies — companies are failing AI by expecting instant results. The path to profitability requires strategic problem selection, strong data foundations, and ongoing human input. As Droege and his team at Scale AI demonstrate, success comes from building systems that work in tandem with people, not in place of them.
For now, AI’s true potential lies not in grand promises but in precise, targeted execution — and that’s where the next wave of innovation will thrive.
