PromptQL, the artificial intelligence unicorn valued at over $1 billion, is launching an unconventional consulting practice that puts its own AI engineers directly in front of Fortune 500 decision-makers at $900 per hour — a move that could disrupt traditional consulting firms struggling to deliver on AI transformation promises.
The San Francisco-based company, which has quietly signed seven-figure deals with some of the world's largest corporations in recent months, announced this week it will offer "AI Investment Assessment" services designed to address what recent MIT research identified as a 95% failure rate among enterprise AI deployments.
Unlike traditional consulting firms that dispatch MBAs with limited technical experience, PromptQL's service deploys the actual engineers who built two billion-dollar products in three years. The consulting offering represents a stark departure from the typical vendor-customer relationship, positioning PromptQL as both a technology provider and strategic advisor.
"Every data leader I talk to — when I ask them what they measure, they don't measure accuracy and adoption," said Tanmai Gopal, PromptQL's co-founder and CEO, in an exclusive interview with VentureBeat. "They're measuring data prep goals. They're measuring things like, 'In this quarter, we'll have all of our data centralized into this data warehouse.' You know what they're not measuring? Whether anyone will actually use it."
Why enterprise AI systems are "confidently wrong" and costing companies millions
PromptQL's entry into consulting comes as enterprises grapple with what the company calls the "confidently wrong" problem — AI systems that deliver incorrect answers with unwavering certainty. This fundamental issue, Gopal argues, has created a universal "verification tax" where business users must fact-check every AI response, eliminating productivity gains.
"The biggest problem is not being able to just accurately answer the question," Gopal explained. "The big problem is AI pretends to be accurate even when it's not. So AI is confidently wrong. That's the problem."
The company's core innovation centers on teaching AI systems to signal uncertainty and learn from feedback — capabilities that distinguish it from traditional large language models that hallucinate information without acknowledgement. This approach has enabled PromptQL to achieve what it claims is near-perfect accuracy for enterprise clients across analysis and automation use cases.
From open source success to billion-dollar AI platform: the Hasura-to-PromptQL journey
PromptQL emerged from Hasura, the open-source data access platform that achieved unicorn status after raising venture funding from Lightspeed Ventures, Greenoaks, Vertex Ventures, and Nexus Venture Partners. Hasura's GraphQL engine reached over one billion downloads and is used by 50% of Fortune 100 companies.
"We went from 2 million to 100 million downloads in the first year," Gopal said, describing Hasura's explosive growth. That success positioned the team to tackle what they identified as the next major challenge: enabling AI to accurately access and reason about enterprise data.
Two years ago, the team spun up a research lab that became PromptQL, recruiting talent from Google Search, Microsoft Research, and Intuit Research. The timing proved prescient as enterprises began confronting the limitations of generic AI tools.
"We realized that apps are no longer the most important things that need access to data," Gopal said. "The future is AI talking to data—AI accessing data on the user's behalf."
Seven-figure deals with Fortune 500 giants signal enterprise AI market shift
PromptQL's client roster includes what Gopal describes as "the world's largest networking company, doing about like, you know, 50, 60 billion revenue, your world's largest fast food chain, the two largest grocery food delivery companies in the world." While the company cannot yet publicly name these clients due to ongoing legal clearances, it has been closing seven-figure deals in three to five months for a product still technically in beta.
The scale of deployments reflects enterprise appetite for reliable AI systems. For high-transaction-volume companies, PromptQL processes close to a petabyte of data. For large sales organizations, deployments reach 25,000 users across multiple subsidiaries and acquired companies.
"We're doing seven-figure deals in three to five months for a product that's still in beta," Gopal said. "The momentum is so hot that PromptQL is now our main focus, and we let the GraphQL API product run on autopilot."
How PromptQL's architecture solves enterprise AI's biggest technical challenges
PromptQL's architecture addresses enterprise AI challenges through three core innovations: an agentic semantic layer that captures evolving business context, including "tribal knowledge" that exists only in employees' heads; a domain-specific language that separates query planning from execution to avoid hallucinations; and a distributed query engine that accesses data across systems without requiring centralization.
The company's approach contrasts sharply with retrieval-augmented generation (RAG) systems that treat all queries identically. Instead, PromptQL uses intent-driven routing to optimize different types of questions, achieving response times 26 to 90 times faster for simple queries while maintaining higher accuracy for complex analyses.
"Instead of generating answers, we generate plans in a domain-specific language unique to your business," Gopal explained. "Those plans compile to deterministic actions with runtime validations and policy checks."
AI engineers vs. McKinsey MBAs: how PromptQL plans to disrupt $200 billion consulting market
The consulting service launch targets what PromptQL sees as a fundamental mismatch between traditional consulting approaches and AI implementation realities. While McKinsey, Deloitte, and other Big Four firms charge millions for AI transformation strategies, their success rates remain disappointingly low.
The contrast is stark: while traditional consulting giants command millions for AI transformation strategies that rarely deliver measurable results, PromptQL's engineering-led approach is already generating significant cost savings for early clients despite being a fraction of the price.
The $900 hourly rate, while premium, represents a fraction of typical Big Four engagement costs while providing direct access to engineers who have successfully deployed production AI systems. Early clients report "millions in savings" from replacing overconfident AI systems with reliable alternatives.
The 95% AI failure rate: why most enterprise deployments never make it past pilot stage
PromptQL's consulting launch occurs against a backdrop of widespread AI deployment struggles. A recent MIT study indicates that 95% of enterprise AI pilots fail to deliver measurable ROI, with value concentrating in a small minority of integrated, learning systems.
The failure rate stems partly from treating AI as traditional software rather than recognizing its unique requirements for continuous learning and feedback. PromptQL's emphasis on measurement and evaluation—through what it calls "GATs" (GenAI Assessment Tests) — addresses this gap by providing concrete metrics for AI performance and business impact.
"Before you fund another 'AI for X' pilot, ask: Will it tell me when it's unsure—and why?" Gopal advises executives. "Does it learn from the correction I just gave it? Will the next user avoid the same trap without re-prompting?"
The new AI consulting playbook: how technical expertise could reshape strategic advisory services
PromptQL's move signals a broader transformation in how AI companies approach enterprise relationships. Rather than simply selling technology, leading AI firms are positioning themselves as strategic partners who can guide implementation from conception through deployment.
This shift reflects growing recognition that successful AI adoption requires more than technical solutions — it demands organizational change, process redesign, and cultural adaptation that traditional consultants often struggle to deliver alongside technical depth.
The consulting service also creates a powerful feedback loop for PromptQL's core platform. Each engagement generates insights about enterprise AI challenges that can be incorporated into product development, potentially accelerating the gap between PromptQL and competitors who lack direct implementation experience.
As more AI unicorns observe PromptQL's success, the consulting industry may face its own "confidently wrong" moment—discovering that technical expertise, not just strategic frameworks, determines whether enterprise AI initiatives succeed or join the 95% that never deliver value.
For an industry built on the premise that intelligence beats implementation, PromptQL's bet is simple: when it comes to AI, the engineers who build the systems understand the business better than the consultants who've never coded a line.