Matthew Fitzpatrick, a seasoned veteran in operations and growth strategy, recently assumed the prestigious role of CEO at Invisible Technologies. Fitzpatrick’s expansive professional journey includes leading QuantumBlack Labs at McKinsey, where his expertise in scaling AI products and leading large engineering teams shone through. As the new CEO of Invisible, he brings those skills with him to focus on operationalizing AI in a balance of automation and human expertise, aiming to transform how enterprise workflows function.
But what sets Invisible Technologies apart in the world of automation? Unlike a conventional automation company, Invisible doesn’t just replace human labour with digital agents – it creates bespoke workflows where both digital agents and human employees collaborate effectively. The company’s unique “work-as-a-service” approach enables businesses to outsource complex duties (such as data enrichment, customer support, and back-office operations) to them, freeing up time and resources to zero in on strategic business growth.
For Fitzpatrick, the allure of Invisible lay in the potential to scale artificial intelligence effectively. His time at McKinsey, helping clients create AI products, has indeed informed his current mission at Invisible: “At Invisible, I get to help them operationalize it,” Fitzpatrick states. The idea of Reinforcement Learning from Human Feedback (RLHF) and its crucial role in creating reliable generative AI systems remains central to his beliefs.
Invisible Technologies wholeheartedly supports the complete AI value chain, from data cleaning to custom evaluations, shaping their mission around the intertwining of human intelligence and artificial intelligence to fully unlock enterprise potential.
But Fitzpatrick also highlights an essential aspect of successful AI implementation: it’s more than just technology – it’s a transformation. “The winners in AI are those who master the ‘last mile,’” Fitzpatrick insists, alluding to the challenging transition from experimentation to production. At his new company, there is a keen focus on structured processes, aiding clients’ shift from pilot projects to implementing scalable and quantifiable solutions.
Seemingly, 2024 was a year for AI trials. But Fitzpatrick insists that 2025 is all about appreciating the returns on those trials. Enterprises that are already witnessing tangible ROI are aligning their AI initiatives with business KPIs and improving the data quality. “They’re not just experimenting—they’re scaling with purpose,” he notes.
Looking further into the future, Fitzpatrick predicts a rising demand for high-precision data labeling as foundational model providers like AWS, Microsoft, and Cohere take on more complex verticals. Invisible meets this demand with an elite pool of experts—only 1% of applicants are accepted, with 30% holding advanced degrees. This expertise plays a pivotal role in training models that can provide nuanced, context-aware feedback.
Invisible Technologies is also leveraging what Fitzpatrick defines as “agentic AI” — systems that plan, make decisions, and act within certain boundaries and function more like coworkers than tools. In areas like customer support and insurance claims, these AI systems reduce manual effort, improve consistency, and are designed not to replace humans but to augment them.
As we look to the future of AI and human collaboration, Fitzpatrick envisions AI becoming the infrastructure supporting human expertise in fundamentally critical sectors like finance, healthcare, and government. “AI won’t replace experts—it will empower them,” he insists.
Take a look at the original interview with Matthew Fitzpatrick at Unite.AI.
This website uses cookies.