Athena is your AI-powered remote hire 🎛

Plus: CEO Brendon Geils on how Athena is automating complex workflows for businesses...

January 17, 2025

CV Deep Dive

Today, we’re talking with Brendon Geils, Founder and CEO of Athena Intelligence.

Athena Intelligence is an enterprise AI analyst designed to function like a remote hire. Built on top of Athena’s Olympus platform, Athena can take on distinct employee roles such as paralegal, intelligence analyst, or market researcher, and operates across tools including  Slack, email, spreadsheets , and more. Athena’s goal is to enable organizations to better automate workflows, improve productive efficiency, and streamline manually-intensive and data-heavy tasks. .

Today, Athena has already seen adoption across industries like legal, market intelligence, and corporate development, with customers using it for tasks ranging from document intake to real-time decision support. Its flexible deployment options, including fully on-prem setups with air-gapped networks, make it a standout solution for businesses with stringent security and compliance requirements.

In this conversation, Brendon shares how Athena was built, the challenges of developing AI agents for enterprise use, and his vision for how AI can transform the way businesses operate.

Let’s dive in ⚡️

Read time: 8 mins

Our Chat with Brendon 💬

Brendon - welcome to Cerebral Valley! First off, give us a bit about your background and what led you to start Athena? 

Hey there! I’m Brendon Geils, originally from the Chicagoland area. My formal background is in electrical engineering. I’ve worked as an electrical engineer at a few firms, most notably SpaceX down in Cape Canaveral - this was during the early days when we were launching and landing rockets with about a 50% success rate. I worked on 5 launches, and we landed 3 of them - it was an exciting moment for the industry.

Most of my professional career has been with Fortune 500 enterprise software companies like Palantir, Scale AI, and Praxis. My roles were typically focused on go-to-market and deployment functions. At Palantir, I was based in the Middle East and then did a lot of work in DC with their defense and intelligence business.

During that time, I got to see how both governments and Fortune 500 companies were building out their data infrastructure, which evolved into data science infrastructure and later AI infrastructure. I also dealt with the non-technical challenges that come with that—educating large enterprises, addressing security and compliance concerns, and deploying into human workflows with redundancy, consistency, and reliability. That experience became a great testing ground for Athena.

How would you describe Athena to an AI engineer or enterprise who’s slightly less familiar with what you do? 

Athena is an enterprise AI analyst, or AI agent as they’re increasingly called. Today, Athena works across various organizations as a paralegal, an intel analyst, a market research analyst, and even an appraiser for a real estate tech firm. It takes on different roles in different companies, much like a new technical hire out of university would adapt to entry-level roles.

We think of Athena as essentially an artificial employee. COVID has been great for us in that it normalized the process of hiring, training, and getting value from a remote hire. Athena as an artificial human performs those same functions remotely.

Our early users are companies like Anheuser-Busch's product innovation  teams. For example, when they commercialize new products —let’s say they want to launch a new Corona flavor in Brazil while their South African division did  something similar with different packaging—they Athena to help answer questions like how well the new product might perform.

This type of work usually involves a team of people specialized in product, marketing, sales, and distribution, who ask various ad hoc questions of their data. Athena essentially acts as another counterpart in this process, supporting global innovation teams to quickly support  or reallocate resources to other new products .

Are there any specific use-cases that you’d like to highlight? How are you measuring the impact you're having on your key users so far? 

I’d say today the largest use case is document intake. Most firms have a process, often via email, where they need to handle large documents. For example, a law firm might receive a 50-60 page document and need to extract details for a loan. For another firm, it might involve pulling out invoicing details.

Traditionally, this process requires running the documents through specific software tools that need configuration. Since Athena is an agent, users can define these tasks in real time, send them over via email, and Athena takes the first stab at a workflow—something that previously would have required manual setup.

As far as impact, there are two main buckets we’ve seen in procurement. The first is a cost savings exercise—how many human hours can we reduce or remove? Each company typically has its own method for measuring this, so we rely on them to determine how they want to measure ROI.

The second bucket is revenue generation. This involves offering a service that the company couldn’t provide before, for whatever reason. By implementing Athena, they’re able to open up a new revenue channel within their existing business.

We've seen a lot of excitement around AI agents and the idea that autonomous systems will be able to complete a lot of different tasks. How are you thinking about integrating AI agents into Athena itself? 

We’ve been articulating this vision for a long time. Up until about six months ago, we were selling Athena as a preconfigured agent that could do a bunch of tasks. At the end of the day, it came pre-made. What we’re starting to see now—and what we expect more market pull for—is the ability to buy Athena and then build multiple Athenas.

For example, we have one customer who has written down 52 agents they believe they can fully hand over to an autonomous system like this by 2025. Depending on where a company is in their life cycle with agents, they might be buying an agent that does one small task in a part of their business or something even earlier, like a copilot. Then there are other groups coming to us saying, “Here are the 52 agents we want. Which ones can you guys do? Which ones do we need to help build?”

This is where swarm architectures and multi-agent architectures come into play. We try our best not to call these “agents” because of the marketing baggage, but sometimes we have to. Honestly, calling it a remote employee is more aligned with how you should think about it—like hiring, onboarding, and training.

If you think of it as a remote hire, you set expectations around everything a remote hire does. That avoids a lot of the issues people face, like logging into a system and it not working on the first click. Then you ask: did you give it access to the data? Does it have context? Does it even know the workflow? There’s an education process that gets missed when you call it an “agent,” but calling it a remote hire sets clearer expectations for how it operates.

You have Athena, which is your AI employee - and then Olympus, which is the name of the underlying platform. Talk us through the distinction there and what the Olympus side of things is all about? 

In the early days, we marketed and sold the software as a tool rather than an agent because it was more difficult to sell the idea of an agent at that time. Now, the distinction is much clearer. All of the agents, including Athena, sit on top of a platform we call Olympus.

For example, we have a firm with 13 agents, and Athena is just one of them. They think of Olympus as a kind of scratchpad with applications like a spreadsheet tool, a document tool, a Python or Jupyter notebook, and about 12 other tools. Athena is essentially one user of the platform and is very good at using it, but you can also develop additional agents on top of Olympus. Athena is just the first of many.

Olympus itself is what everything sits on top of, and it integrates directly into the business. The main differentiation is in how they’re deployed. Olympus is deployed into customer environments as a standalone piece of software, running continuously. Meanwhile, Athena might operate as an ambient system that activates when it receives an email and turns off when it’s done. Olympus, on the other hand, is an always-on system.

Take us under the hood of Athena’s stack - how are you thinking about merging multiple models and fine-tuning to get the results you’re looking for? 

We’re probably the most agnostic company ever, which is really exciting. Today, we have separate runtimes for different models. Out of 25 models, we currently deploy eight model providers, and those 25 models operate on top of two runtimes. For example, LangGraph would be one runtime, and Live Kit would be another. These are the systems taking in all the information and performing agentic architecture, reasoning, and pass-through.

In practice, this system is voice-enabled, chat-enabled, and can join video calls. It records, participates in videos, and operates just like a human across Slack, email, text, audio, and video. Over time, it’s becoming more proficient in each of these modes.

Right now, for example, voice is a little further behind chat in terms of capability, but it’s continuing to improve.

What has been the hardest technical challenge around building Athena into the platform it is today?

Our hardest technical challenge is that we allow the system to be completely on-prem. From the beginning, we made the decision to let customers use third-party services, including tooling and models, but ensure the entire system can deploy into a private cloud environment within fully air-gapped networks.

This means that while we can use third-party APIs, we don’t allow them to be a core part of our system’s architecture. Everything we build and deploy has to work in a private cloud environment.

The challenge comes down to taking something that works as a cool API from a new company and making it functional in a private cloud with an air-gapped network.

How do you plan on Athena progressing over the next 6-12 months? Anything specific on your product roadmap that your existing customers are excited about? 

We’ve seen a ton of demand from the legal sector, so we’re looking to double down on that. There’s also been strong interest from CPG and retail-style customers, which we’re excited about. Market intelligence and corporate development teams have also found a lot of value in the software. Right now, we’re focused on going deeper with the customer segments we already have.

One thing I’m particularly excited about is getting the system hired without anyone realizing it’s software. Today, we go to market by submitting fake resumes. While people figure out it’s not a human fairly quickly, we’ve managed to have the agent complete full interviews and even get hired through that process.

My goal for 2025 is to have the system fully hired—where no one realizes it’s software throughout the entire process!

Lastly, tell us a little bit about the team and culture at Athena. Are you hiring, and what do you look for in prospective team members that are joining?

Today, we’re a nine-person team, all in-person and based in New York. We’re committed to keeping it that way—we don’t do remote hires because we really value having everyone in the office together.

Our team is made up of platform engineers, product engineers, and forward deployed engineers. Platform engineers focus on building the core infrastructure, product engineers work on the applied AI side to develop features, and forward deployed engineers sit with customers to implement and refine solutions.

We’re hiring across all three roles. We’ve had a lot of success with engineers or engineering consultants who want to go deeper into the technical side, as well as research-focused individuals looking to work on applied AI and infrastructure.

What’s exciting is that this is already in production with real companies, many of which are also based in New York. In fact, the office I’m sitting in right now is one of our customers. After we moved in, they knocked on our door and said, “Hey, this sounds pretty cool.” Now we’re helping them have a full-time analyst working for them through our system.

Conclusion

To stay up to date on the latest with Athena, learn more about them here.

Read our past few Deep Dives below:

If you would like us to ‘Deep Dive’ a founder, team or product launch, please reply to this email ([email protected]) or DM us on Twitter or LinkedIn.