Tools and infra to support the broad adoption of AI in the enterprise. Part of our Request for Startups series.
By Ryan Hoover and Vedika Jain on July 11, 2023
We’ve been investing in superpower tools that unlock new abilities and help people achieve more in their jobs since Weekend Fund started in 2017.
LLMs are not just a superpower tool, they’re infrastructure to create superpower tools. They’re programmable with words – hence, “natural language”. If you can speak (or type), you can use these tools with far less training than traditionally required. This language-driven interface is making existing superpower tools like Zapier Natural Language Actions and Intercom’s Fin far more accessible and creating a new category of AI-native superpower tools.
We’re seeing consumers rapidly adopt new AI tools – e.g. OpenAI’s ChatGPT amassed 100M users within its first few months of launch – but stakes are higher in the enterprise. They’re even higher in regulated industries like automotive, fintech, and healthcare. Today, enterprises lack the privacy, security, compliance, data integration, workflow orchestration and access/permissioning tooling to unleash this new technology… for now.
We’re interesting in funding startups building tooling for the broad adoption of AI in the enterprise - the “enterprization of AI”
Enterprises will demand tooling to adopt AI quickly, safely, and reliably. They’ll need application builders that allow them to ship AI-powered products (and internal tools) powered by their propertiary data. They’ll need a unified data integration layer that gives LLMs access to the most relevant parts of their data. They’ll need workflow orchestration infrastructure that lets employees automate the most repetitive parts of their jobs themselves. They’ll need better ways of handling privacy, security and compliance. The right employees need to be able to use the right data requiring LLM-native permissioning and access control infrastructure. When things go wrong, companies need to be able to accurately detect, diagnose, and correct model failures increasing demand for monitoring.
We’re seeing an explosion of early-stage startups in the space already.
Enterprise-grade application builders: StackAI allows you to build and deploy AI applications in minutes, LastMileAI allows you to build multi-modal AI-powered apps across text, image and audio and Olympus allows you to build AI apps personalized with user data.
Data access, integration and management infra: Usable Machines allows you to connect any AI model to your company’s data safely, and Oxen helps manage machine learning datasets.
Privacy and compliance tooling: Commonbase helps companies deploy secure and compliant LLM integrations, DyanmoFL is building infrastructure for regulation-compliant AI, and Gretel helps companies develop models AI models using synthetic data.
Security and permissioning/access control tooling: Cadea allows businesses to deploy enterprise-level AI tools with data security and role based access control, Preo is an enterprise security solution that protects against data loss and other threats introduced by generative AI, and Credal allows enterprises to securely adopt AI enforcing existing access policies.
Quality and performance tooling: OpenLayer is building model evaluation infrastructure and guardrails for AI and Guardrails adds guarantees to the outputs of LLMs.
The upcoming wave in enterprise adoption of AI creates new opportunities to build infrastructure.
We’re especially interested in companies that are building:
We’ve made one unannounced investment in this category and keen to do more. If you’re building something in this space, and bold enough to go where no man has gone before, drop us a note at team@weekend.fund.
P.S. We’re working on a larger market map for this category. Subscribe at Signature Block to get it directly in your inbox.
P.P.S. Read more about what we want to invest in here.
