20VC: Enterprises Will Not Adopt AI without Forward-Deployed Engineers | Who Wins the Data Labelling Race: How Does it Shake Out? | How Synthetic Data Threatens the Future of Human-Generated Data with Matt Fitzpatrick, CEO of Invisible Technologies - The Twenty Minute VC Recap
Podcast: The Twenty Minute VC
Published: 2025-12-31
Duration: 1 hr 24 min
Guests: Matt Fitzpatrick
Summary
Matt Fitzpatrick argues that enterprise AI adoption hinges on forward-deployed engineers, not just technology. He also explores the future of data labeling and warns against over-reliance on synthetic data.
What Happened
Matt Fitzpatrick, CEO of Invisible Technologies, discusses the challenges enterprises face in adopting AI. He asserts that merely having AI models isn't sufficient; enterprises need robust data infrastructure, workflow redesign, and trust. Fitzpatrick emphasizes the necessity of forward-deployed engineers (FDEs) for successful AI implementation, a strategy acknowledged by companies like Palantir a decade ago.
Invisible Technologies has raised over $100 million under Fitzpatrick's leadership, focusing on AI training platforms. They utilize a modular architecture, which includes components like Neuron and Axon, to cater to a variety of sectors. The company has achieved significant enterprise engagement, confirming 12 deals in just 45 days.
Matt Fitzpatrick challenges the viability of AI talent marketplaces, suggesting that specialized data is becoming more crucial as opposed to generic labeling tasks. He predicts that the market for human-generated data will continue to grow due to its necessity in complex reasoning tasks, despite the rise of synthetic data.
The episode also covers the evolution of the AI landscape, with Fitzpatrick noting that externally driven AI projects tend to be more effective than those developed internally. He anticipates that many AI projects will be canceled by 2027 due to a lack of proper execution strategies.
Fitzpatrick reflects on his company's transition from a fully remote model to having physical offices, recognizing the benefits of in-person collaboration for building culture and enhancing productivity. He also stresses the importance of capital allocation and discusses how Invisible Technologies is investing in enterprise-specific growth rather than focusing solely on AI training.
The episode concludes with a discussion on the rapid changes in AI technology, where Fitzpatrick notes that strategic planning should incorporate flexibility due to the fast-evolving nature of the field. He also highlights AI's potential positive impact on sectors like healthcare and education, emphasizing its role in improving risk identification and providing educational access.
Key Insights
- Enterprises require forward-deployed engineers (FDEs) to successfully implement AI, a strategy used by companies like Palantir for over a decade. This approach ensures robust data infrastructure and workflow redesign, which are crucial for AI adoption.
- Invisible Technologies has secured over $100 million in funding and confirmed 12 enterprise deals in 45 days, leveraging a modular architecture with components like Neuron and Axon to serve diverse sectors.
- The market for human-generated data is expected to grow despite the rise of synthetic data, as complex reasoning tasks still rely heavily on specialized, human-labeled data.
- Externally driven AI projects are generally more effective than internal ones, with many AI initiatives predicted to be canceled by 2027 due to poor execution strategies.