🎙️ Upside Chat with Andre Antonelli, CEO of Machina Sports, a Leading AI Agent Platform for Sports.

Today we have the honor of interviewing André Antonelli, CEO of Machina Sports, a leading AI agent platform for sports.

Machina Sports is an AI infrastructure platform designed specifically for the sports industry, enabling teams, media companies, and sportsbooks to build and deploy intelligent AI agents that automate content creation, fan engagement, and data analysis workflows. Their platform combines structured sports data (like match statistics) with unstructured sources (such as news and fan sentiment), processes it using large language models and vector databases, and outputs personalized, multi-language content — including match previews, quizzes, and betting insights. Through a developer-friendly SDK and customizable agent templates, Machina Sports empowers organizations to scale AI-driven content generation without needing to build and maintain complex infrastructure themselves.

Here is a video that explains how Machina Sports AI platform works:

Video: Machina Sports

Here is a screenshot of Machina Sports’s developer studio to create AI agents:

Here is a screenshot of Machina Sports’s analytics tool within developer studio to track its AI agents:

📝Show Notes: Through this interview, we touched on his background, his company and product, the benefits for sports organizations and teams to use his product. We also talked about his competitive advantage, business model, and his plans for the next 12 months.

Andre will be speaking on our AI based fan engagement panel at our Upside US Sports tech summit on October 23 at a World’s class sports stadium in Sunrise FL.

You can read the full transcript of the podcast interview with Andre located at the top of this blog post.

Here are the quotes from the interview with Andre:

---

Q1. Tell me about your background.

"I'm a computer scientist by training — I studied in Brazil and was born there. Since 2017, I've been based in the San Francisco Bay Area working for major companies like Nestlé and Gilead, primarily focusing on emerging technologies. I’ve always been interested in how tech like AI, IoT, and even satellite data can be applied in impactful ways — whether it was for agriculture at Nestlé or innovation projects across industries. I also immersed myself in the startup ecosystem here, especially in the last few years. Toward the end of 2023, I started attending a lot of AI events and hackathons in San Francisco, and that’s actually where I met Fernando, our CTO. From there, we started collaborating on what would eventually become Machina Sports. So while I come from a technical background, my focus has always been at the intersection of tech, product, and innovation — and more recently, doing a lot of hands-on development with AI myself."

---

Q2. Tell me about your company and product.

"Machina Sports is an AI agent platform purpose-built for sports. What we’ve developed is a foundational infrastructure that allows sports organizations to easily build and deploy intelligent agents. These agents can process and synthesize structured data — like stats from matches or leagues — but also bring in unstructured data like news articles, fan comments, and sentiment from platforms like YouTube. We then use vector databases and large language models to contextualize that data and output a range of content, from match previews to real-time quizzes and fan engagement tools. So rather than having a one-size-fits-all AI product, we provide a flexible software development kit and open-source templates that customers can customize and integrate directly into their own apps, websites, or workflows. We’re really focused on enabling teams to scale and personalize content creation while reducing the operational lift."

---

Q3. What are the benefits for your customers to use your product?

"The biggest benefit is speed and simplicity. Instead of having to assemble a whole AI team, figure out how to maintain vector databases, and integrate with multiple real-time data sources, our customers can use our infrastructure to go much further, much faster. We’ve already mapped out a variety of use cases with early partners and have open-source agent templates that make it easy to get started. On top of that, because we’re packaging everything — from the retrieval pipeline to the real-time semantic enrichment — customers don’t have to worry about keeping up with the latest model updates or maintaining their own AI stack. That alone saves a huge amount of time and resources. And because our platform is designed for developers, it integrates seamlessly with their existing systems through APIs, giving them full control and flexibility."

---

Q4. How does DAZN use Machina Sports’ AI platform today?

“DAZN is one of our early customers. They have over a thousand live events each month. They used to manually create ‘run of show’ documents with quizzes and trivia. Now, our AI agents automate that process at scale. We integrated their existing data, added news and fan sentiment, and built a semantic layer on top of each match. The AI now generates quizzes and posts in 12 languages, one to two hours before each event. They still use a human in the loop for validation, but the content is created automatically.”

Q5. How does one sportsbook client use Machina Sports’ AI platform today?

“We have one sportsbook that we are working with — still private project — but we are working with them for SEO. Today, the cost to generating these articles with their editorial team is 30 to 60 dollars per article. And with our stack it's gonna be like less than a dollar. Right? So 98% savings. They also bring more information... integrated with the odds provider from this specific sportsbook. And now it's creating articles, match previews at scale for multiple languages for all the competitions they wanna track.”

“The AI goes through the event statistics, goes through the news, to the odds, and generates an article of match preview — t’s more for new users to explain how you should bet, what type of information you should look at when you’re making a betting decision.”

Q6. What do your customers like the most about your product?

"Most of our customers tell us they’re blown away by how easy it is to get started. A lot of them have tried to build something similar internally — like RAG systems or ways to integrate LLMs with sports statistics — and they hit walls quickly. When they try our platform, they immediately see how much faster and easier it is. We’ve made the technical part accessible, even if the platform is still early and leans technical today. That said, we’re working to bring more of the backend functionality into the frontend and moving toward a low-code or even no-code experience, so eventually more people on the business side — not just devs — can build and experiment without friction."

---

Q7. Tell me about competitive advantage and business model.

"Our biggest competitive advantage is that we’ve curated a full AI stack specifically for sports. Unlike generic AI platforms where you have to assemble all the pieces yourself, we provide a turnkey system — real-time data integration, semantic vector search, low latency embedding, all optimized for sports workflows. It’s like what Google did for web search or Airbnb did for travel — we’re doing that for sports AI agents. And what’s exciting is that AI in sports is not disruptive in a way that replaces people or events. It’s additive. The game still happens live, but we help enhance the fan experience and business operations by unlocking richer, more personalized content around it."

"In terms of business model, it’s usage-based — very similar to OpenAI. You pay per million tokens. For enterprise clients, we offer a monthly package where we also support custom integrations and agent development. For startups and smaller teams, we have a pay-as-you-go model — currently priced at $80 per million tokens — which includes the full infrastructure, not just the token cost. So it's flexible, scalable, and lets clients build on their terms."

---

Q8. What are your plans for the next 12 months?

"We’re currently raising our seed round, but we’re also laser-focused on product. The backend is solid — our SDK is live and developers are already using it to build agents via API. The next step is making the front-end experience more user-friendly. We’re working on visual workflow builders and interface upgrades that will let users drag and drop their way to creating powerful agents without needing to write code. On the technical side, we’re also training a sports-specific language model — a fine-tuned open-source LLM — to make our outputs even more accurate and relevant. That also ties into our work on embedding models, so our semantic search results will be sharper. And from a market standpoint, we’re already working in soccer, and expanding into NFL and NBA use cases as well. As long as we have access to data, our platform can support any sport."

---

Q9. What is the best way to reach out to you?

"You can reach me directly by email at andre@machina.gg — that’s usually the fastest way. I’m also very active on LinkedIn and respond to messages there. And of course, if you want to see the product in action, just head to our website and click on ‘Request a Demo.’ That’ll let you book a time and we’ll walk you through the process of how to build and deploy your first sports AI agent."

You may also like: