YC W24 Demo Day - The implications for enterprise tech

Over the past two days, Y Combinator held its 38th Demo Day, which continues to be one of the most exciting sets of days in tech. We had the privilege to participate, and we congratulate each of the YC W24 batch founders on their companies’ launches!

As with previous batches, we include in this post our summary takeaways–including both favorite quotes and insightful trends from the founders’ presentations. Y Combinator founders reflect the best of the best at the earliest stages of innovation and startup formation.

Of note, we define enterprise tech broadly, across developer, SMB, large enterprise, and government customer segments. 

I. Favorite Quotes

Over the two days and 243 presentations, we had numerous favorite moments. Some of these are unsolved problem statements about the world. Some of these are contrarian trends. Some are inspiring facts about the founders. Our takeaway quotes from the W24 batch were:

“The current analytics stack just doesn’t cut it for AI.” Dawn

“LLMs will be as ubiquitous as Javascript.” Vectorview

“Ingestion is one of the biggest bottlenecks for deploying enterprise AI applications.” Reducto

“The open source AI stack is a mess. You have to use at least 10 different tools just to get started.” OpenFoundry

“The shift to AI is just like the shift to cloud 20 years ago. Every company is going to build on it.” OpenFoundry

“Most impactful LLM apps require retrieval, and most teams are building their own stack from scratch. Existing solutions like Langchain only complicate the process.” SciPhi

“I've been programming for over 13 years since the age of 11.” Preternatural AI

“Jupyter notebook will get replaced as the de facto data tool just as MySQL got replaced by Postgres as the de facto database.” Pretzel AI

“Most company data is unstructured, and existing data platforms cannot handle it.” Trellis

“80% of corporate data is locked away in unusable formats today.” OmniAI

“Flutter is the fastest growing cross-platform framework today with 5 million developers globally, growing at 33% per year.” Celest

“Fixing bugs is the hard and valuable part about building an AI software engineer, not generating code.” Tusk

“Today, every engineer needs to be an observability expert, and AI code generation is making observability even harder.” OneGrep

“Retool got stuck building a product that is halfway between code and no code. Now with AI, it is possible to build internal apps with complex state management with minimal effort.” Creo

“Nine in ten businesses are using or looking to use computer vision.” Dragoneye

“Our goal is to build a foundational model for humans in video.” sync labs

“When I was head of growth at Twitter, we changed our push notification from ‘Elon tweeted’ to ‘Elon just tweeted’ and we got 800K additional users. But such tiny changes would take us three months to launch.” just words

“We are building AI-first Salesforce. Instead of a system of records, we are building a system of actions.” Octolane AI

“We will 100x the amount of video created by companies.” Yarn

“Local sales run on voicemails.” Terrakotta

“We've been building products together since fourth grade.” DryMerge

“Adobe earns more than $2.7 billion every year on the PDF ecosystem, and they have been ignoring developers for decades.” Onedoc

“The bottleneck for robotics foundation models is the lack of data.” yondu

“We’ve personally pentested over 50 LLM applications, and we found cybersecurity risks in all of them.” PromptArmor

“We're at the cusp of a $40 billion market opportunity that parallels the cloud security market.” PromptArmor

“~50% of government contracts today have two or fewer bidders.” Hazel

“We are building a foundation model for the brain.” Piramidal

II. Company Themes and Descriptions

For this batch, 75% (182 of 243) of the companies were tagged by Y Combinator as B2B. By comparison, the S23 batch was 70% B2B (153 of 220), the W23 batch was 70% B2B (191 of 273), the S22 batch was 51% B2B (120 of 236), and the W22 batch was 43% B2B (175 of 403). Enterprise tech continues to dominate the early stage innovation at YC. 

In this section, we review 139 of the 243 W24 companies, organized into 13 themes: AI tools, data, devtech, workflow automation, salestech, martech, healthtech, fintech, hrtech, robotics, cybersecurity, other verticals, and other. 

1) AI Tools

Garry Tan, CEO of Y Combinator, commented in his opening remarks that 50% of the W24 batch use LLMs in some way. A major theme in the batch was AI tools–the infrastructure required to build the AI wave. Specifically, there were several companies each in generative AI retrieval, evaluation, and deployment. 

• Retrieval: Quivr provides an open source RAG framework to deploy chat assistants. Trieve provides infrastructure for search teams that combines search language models with fine-tuning tools. SciPhi provides a platform for developing, deploying, and optimizing RAG systems. Preternatural AI provides client-side ETL and RAG framework for Swift developers to integrate AI search and assistants into apps.
• Evaluation: phospho provides open source text analytics for LLM apps. Ragas provides open source evaluation for LLM apps. Relari provides an LLM app testing and simulation platform. Vectorview builds custom evaluation tasks for FMs and LLM agents. MAIHEM provides AI agents that test AI products such as chatbots and voicebots.
• Deployment: nCompass provides AI model hosting and acceleration platform, starting with voice AI and chatbots. Tensorfuse, the "OpenShift for LLM apps", helps enterprises deploy LLM pipelines in their clouds. Preloop helps data scientists translate experimental scripts into production services.
• Analytics: dawn provides analytics for AI products to define user cohorts, understand feature usage, and flag malicious activity.
• Applications: Keywords AI provides DevOps platform to build LLM applications.
• Hallucinations: Reprompt enables non-engineers to tune models in production and fix hallucinations without code.
• Ingestion: Reducto provides an ingestion API to convert complex documents into chunks for LLMs.
• Interpretability: Guide Labs provides interpretable FMs that provide accuracy and human-understandable explainability.
• Memory: Lantern provides postgres vector database to improve performance and reduce cost. Zep AI provides long-term memory for AI assistants to recall, understand, and extract data from chat histories.
• Open Source: OpenFoundry provides a developer platform for building on open source AI.

2) Data

Data was another major theme in YC W24. Several of the companies emphasized the problem of unstructured data, which represents 80% of corporate data. Several other companies looked to upend the dominance of Jupyter notebooks in an AI-first world. 

• Unstructured Data: Roe AI provides a data warehouse for unstructured data. Trellis provides Snowflake for unstructured data. OmniAI provides a platform for unstructured data.
• Notebook: Tile provides an AI-first SQL notebook for data apps. Pretzel AI provides an AI-first alternative to Jupyter for notebooks.
• Analytics: Quary provides an in-browser analytics engineering platform. Upsolve AI enables businesses in industries such as marketing and financial services to offer analytics to customers. kater.ai enables executives to ask data questions in natural language.
• Compliance: Titan provides an open source platform for data compliance, helping to identify and close gaps in data access.
• Fine tuning: Datacurve provides complex, high scale data for fine tuning LLMs.
• Integration: Buster provides connection infrastructure for LLMs and databases. Stacksync helps engineers and revenue teams sync data between CRMs and databases.
• Open Source: InQuery provides an open source alternative to Snowflake.

3) DevTech

In addition to AI tools and data, developer tech was the third major trend at YC. This year’s batch included numerous startups using AI for code generation, code search, code testing, code debugging, and code documentation. In addition, several embarked on building the AI-first version of Retool for custom internal tool development. 

• Code Generation: Ellipsis converts technical instructions into code. Double provides AI coding copilot. ion design converts Figma designs into React code. Marblism generates apps from single prompts. Pythagora provides an open source developer tool to build apps by talking with users.
• Code Discovery: Greptile provides API to search codebases in natural language and build internal devtools. Assembly enables developers to search not only codebases but also context sources such as Slack, Linear, and Notion.
• Testing: camelQA tests mobile apps through AI and computer vision. Momentic tests software and eliminates the need for QA departments.
• Debugging: Fume provides an AI software developer to solve complex tasks such as debugging. Tusk provides AI agents to fix customers' bugs. Decipher AI identifies and fixes production issues by analyzing runtime product behavior with AI.
• Documentation: Agentic Labs provides an AI editor to write technical design documents. Driver AI writes interactive technical documentation.
• Tools: Creo is an “AI-powered Retool” to build complex internal apps. Reform, the "Retool for logistics", enables internal custom tools for logistics companies.
• CI: Blacksmith provides serverless cloud infrastructure for CI workloads.
• Cloud: Pump.co, the "Costco for cloud", uses group buying and AI to reduce cloud costs. Ubicloud provides open source cloud infrastructure.
• Cross-Platform: Celest is the "Vercel of Flutter".
• Observability: OneGrep provides AI observability copilot for software engineers.
• Task Management: Hatchet provides an orchestration and visibility platform for background tasks, including rate limiting, concurrency, and observability.
• Visual: Dragoneye helps developers build computer vision models. renderlet, the "Unity for visual apps," enables developers to build 2D and 3D visual apps across platforms. Retell AI provides an API for conversational voice AI.
• Website: Million provides APM that finds and fixes slow website code.

4) Workflow Automation

A next generation of workflow automation may be ahead of us, i.e. an AI-first approach to RPA. Several companies in this year’s batch provided “AI assistants” and “AI interns” to automate tasks ranging from repetitive browser work to complex knowledge work.

• AgentHub provides a no-code AI platform for workflow automation.
• Brainbase provides AI workflow automation, including HR, sales, and customer support built-in templates.
• CloudCruise provides AI-powered RPA software.
• Paradigm provides AI intern to automate repetitive tasks such as lead generation, email outreach, and data collection.
• Speck provides AI assistant for web automation in tasks such as LinkedIn outreach, CRM updates, and data extraction.
• Basepilot provides AI automation of knowledge work by highly paid workers.
• DryMerge automates workflows in plain English.

5) SalesTech

Within salestech, the W24 batch included numerous companies building the AI copilot for sales. Specific tasks ranged from outbound demand generation to price quoting. Separately, other companies looked to build the “AI-first ZoomInfo”, the “Google for B2B data”, and the “AI-first Salesforce.”

• AI Copilot: Artisan AI provides AI sales rep for outbound demand generation and management. Amber AI provides AI copilot for account managers. Veles provides AI quoting copilot for SaaS sales.
• Lead Data: Openmart, the "AI alternative to ZoomInfo", provides local business data with websites, reviews, and socials. Firebender, the "Google for B2B data", helps sales teams find leads with LLMs.
• CRM: Octolane AI is building an “AI-first Salesforce”.
• Relationships: Vector develops shared experience graphs to turn relationships into introductions.
• Video: Yarn enables sales and marketing teams to create studio quality video with AI.
• Voicemail: Terrakotta provides an AI voice platform to leave personalized voicemails.

6) MarTech

Over the past 18 months, customer support has been frequently mentioned as an initial area for LLM projects in our enterprise CIO research. Numerous startups in W24 focused on AI agents to automate or support customer support functions. Other martech W24 founders focused on lip-sync translation, copy management, and video creation. 

• Customer Support: Opencall.ai provides voice AI call centers for SMBs. Duckie provides an AI tech support engineer to handle tickets. OpenCopilot enables developers to build AI text-to-action assistants in products. Senso provides an AI-powered knowledge base for customer support agents.
• Translation: sync labs provides API for real-time lip-sync for language translation in videos.
• WhatsApp: Kiosk, the "Mailchimp for WhatsApp", provides a marketing platform for WhatsApp.
• Copy: just words provides infrastructure to manage, translate, and optimize copy for emails, push notifications, websites, and apps.
• Video: Magic Hour provides a professional AI video creation tool.

7) HealthTech

Among industry verticals, healthtech was one of the two largest at YC W24. Founders covered areas such as behavioral health, neurology, home health, drug discovery, and life science research, among others. In addition, several companies provided “AI receptionists” and “AI assistants” for medical front office staff and for clinicians. 

• Support: Arini provides an AI receptionist for dentists to answer calls and schedule appointments. Anaphero provides voice AI to automate clinical phone tasks such as scheduling, billing, intake, and support. Hemingway provides an AI receptionist for dentists to book appointments, manage cancellations, and verify insurance. Somn provides an AI assistant for clinics to answer calls, book appointments, and call patients.
• Behavioral Health: Attunement provides a behavioral health remote monitoring platform.
• Drug Discovery: Junction Bioscience provides an AI platform for molecular discovery by translating physical atom movements into chemical code. Drug Discovery: Tamarind Bio provides a computational biology platform for drug discovery.
• Home Health: Andy AI provides clinical documentation for home health nurses.
• Insurance: Blume Benefits provides AI copilot for health insurance brokers.
• Life Science: Metofico provides no code data analysis platform for life science research.
• Neurology: Piramidal provides a foundation model for the brain that is trained in-house on large scale EEG data.
• Pharmaceutical: Argon AI provides AI for pharma intelligence, integrating proprietary data with public and third-party data.

8) FinTech

Fintech was the other major industry vertical for YC W24. Companies included the “Plaid for tax documents” and the “Rippling for financial compliance.” Numerous companies focused on AI agents for investment banks and investment firms. 

• Accounting: Taiki, the "Plaid for tax documents", provides an API to extract forms and data from financial and payroll providers."
• Brokerage: Powder provides a sales platform for wealth management advisors.
• Compliance: Greenboard is the "Rippling for financial compliance and operations". Shiboleth automates consumer lending compliance for financial institutions.
• Investment: OffDeal provides an AI-native investment bank for buying and selling SMBs. Clarum provides AI-powered investment due diligence. Model ML provides an AI platform for private equity firms' and investment banks' research and due diligence. Investment Firms: ProSights provides AI analyst for deal tracking and new company alerting for investment firms.

9) HRTech

Several W24 companies provided an “AI recruiter” or an “AI HR person” for SMBs or large enterprises. 

• Apriora provides an AI recruiter to identify and interview talent.
• DianaHR provides an AI HR agent for SMBs.
• Parasale provides A recruiter to find and engage with talent.

10) Robotics

A few companies in the W24 batch focused on robotics–including labeled data and foundation models specific to the industry. 

• Adagy Robotics, the "Scale AI for robotics", provides labeled data to create a robotics foundation model.
• Yondu provides robotics foundation models for intelligent navigation.

11) Cybersecurity

Several W24 companies were in cybersecurity. The companies included both the use of AI to improve security posture and the development of tools to address risks in AI products. 

• PromptArmor provides an API to detect and respond to LLM cybersecurity risks.
• Tracecat provides an open source AI security alert automation platform.
• CodeAnt AI provides an AI platform to address bad code and security vulnerabilities.

12) Other Verticals

Three other industry verticals with a few YC W24 companies were property tech, legaltech, and manufacturing.

• PropTech: HeyPurple provides an AI leasing agent to fill rental vacancies by creating listings, handling communications, and scheduling property tours. HostAI is the OS for vacation rental management. InspectMind AI writes construction inspection reports with AI review of inspection videos. Konstructly is the OS for the construction industry.
• LegalTech: Leya provides AI assistant for lawyers, connecting internal documents with legal data from every jurisdiction. Abel provides AI e-discovery document review for litigation teams.
• Manufacturing: Draftaid provides AI platforms to convert 3D models into manufacturing CAD drawings. Yoneda Labs provides the foundation model for chemical manufacturing. Dime provides the OS for factories to generate production plans and track floor operations.
• Auto: Toma provides AI phone automation for automotive dealerships.
• Electronics: atopile provides tools for code-based electronic circuit board design. SiLogy provides chip developers a collaborative test and debut platform.
• Government: Hazel provides an AI-enabled marketplace for government contracts.
• Retail: BetterBasket provides AI-powered pricing for grocers. Zaymo builds interactive emails for ecommerce brands.

13) Other

Other enterprise startups at YC W24 focused on productivity (collaboration, communication, and presentation), procurement, and logistics. 

• Collaboration: Circleback provides an AI meeting productivity suite including summaries and workflows. Patchwork provides a team communication platform that replaces Slack with intelligent feeds.
• Presentation: Alai creates high quality presentations with AI. Carousel Technologies provides an AI powerpoint toolkit for junior bankers and consultants. Lucite automates presentations for finance professionals.
• Procurement: Forge provides procurement software for the hardware industry. Kontractify provides an AI platform to run RFPs, manage workflows, evaluate proposals, and facilitate reverse auctions.
• Logistics: RetailReady provides a supply chain compliance platform that replaces instruction manuals with a tablet application. Carma provides a marketplace for same-day commercial fleet vehicle repair.
• IT Support: Risotto provides AI IT help desk support in Slack.
• CFD: Navier AI provides real-time computational fluid dynamics (CFD) simulations.
• Document: Onedoc provides an API for PDF document workflows.
• Research: Superagent provides open source AI agent for web research, with use cases such as sales, marketing, and product management.

Conclusion

Garry Tan projected that 14 of this year’s batch would become unicorns, at least based on prior cohorts’ percentages. For this year, AI tools, data, and developer platforms were the major themes, and healthtech and fintech were the major industry verticals. With that said, we found a remarkably wide range of new applications, enabled and differentiated by LLMs, to reimagine enterprise functional areas and industry verticals. 

To give one sneak peek into next week’s Enterprise Tech 30 report… many of the new ET30 companies this year are in gen AI infrastructure, and many of the returning ET30 companies are in data infrastructure. ET30 companies represent the most promising across early, mid, late, and giga stages. The YC companies represent the company formation stage, and the YC W24 batch tells us that the gen AI applications wave is emerging in enterprise tech, leveraging the gen AI infrastructure and data infrastructure over the past few years.

Congratulations again to the entire YC W24 batch. We’re excited to see what’s next for each of your companies and to collaborate with each of your founding teams across Wing’s activities.

Upcoming Wing Summits:

Wing AI: Fireside Chat with Clem Delangue

Wing Publications:

Wing Briefing

Notorious PLG and AI Substack

Data Gravity Substack

Tanay’s Newsletter

We would like to thank Wing Partners Chris Zeoli and Tanay Jaipuria and Wing Consultant Clayton Ramsey for their contributions to this research post.

‍