Finance4U

Investing in Tennr | Andreessen Horowitz


An enduring promise of new technologies is to find ways to increase human productivity. Software and robotics, for example, have always presented immense potential to automate repetitive mundane tasks that humans perform, freeing them up to do more valuable work. In the last few decades, we’ve made meaningful progress toward this goal, and the first generation of robotic process automation created some massive companies, such as UiPath.

Yet, we never made it past the most simple tasks. Moreover, these automations were helplessly brittle, often requiring a new configuration if any of the underlying tools or steps changed. They weren’t living up to their potential and, as a result, there remains a huge category of repetitive processes that humans still perform manually. We can’t help but think solving the next layer in complexity would create even bigger companies!

Watching advancements in LLMs and AI reasoning unfold over the last few years, we think there is about to be an enormous unlock in what’s possible with automation. There’s a vast subset of currently manual tasks that computers can’t perform because that data is unstructured, or the steps involve just enough ambiguity that human reasoning and judgment is required — for example, determining the correct next step based on how a patient has completed an intake form. It turns out LLMs are especially good at overcoming these hurdles, and they will set new standards for what software is capable of automating. 

We first met Tennr two years ago, and we believe they’re the right team to tackle this problem. Today, we are excited to announce we’ve led their Series A.

It’s clear that founders Trey Holterman, Diego Baugh, and Tyler Johnson are craftsmen with a vision to redefine the status quo. In 2022, as LLMs took the world by storm, the market was trying to navigate the right way to build solutions on top of the many models becoming available. Trey kept the team focused and spent months obsessing over novel applications of new open source models. Diego designed and built an elegant product to make these applications practical and usable for non-technical teams. Tyler (who was the fastest promotion to engineering director at his previous firm) scaled the systems data pipelines so that training and inference could work at scalable economics for the average business. Together, they turned sophisticated ML research into differentiated, practical applications for non-technical business users. 

The result is a powerful horizontal product called Tennr, which is anchored around understanding unstructured inputs and doing the work that comes after. It can apply AI reasoning and decision-making to perform end-to-end workflows that are much more complex than what legacy solutions can support. Customers in healthcare, financial services, and logistics have been particularly successful using Tennr to perform processes they previously had to do manually. 

When we spoke to Tennr’s customers, it was clear we were observing a true step-function change in how organizations think about their workflows and processes. We’re quick to think of time and cost savings, but the impact is much larger. For example, one customer pointed out that the ability to automate processes around new customer onboarding is a revenue-accelerating outcome because it enables them to serve more customers. Another pointed out that their staff can now focus their time on higher-value work. We were hearing soundbites reflecting the true potential of process automation that actually works: business owners changing assumptions on their business models. As one customer put it, “We are salivating at the opportunities for what this means for our business.”

We’re excited to partner with the entire Tennr team on their journey. Their focus, determination, and grit to advance how software can make humans more productive is unparalleled. We have no doubt they’ll continue to push the boundaries of what’s possible.

Share This :

Follow Us

Podcast

Subscribe Our Newsletter