Today’s episode of the Future Ear Radio podcast features Bruce Rasa, CEO & Co-Founder of AgVoice. Bruce’s background of growing up on farm in Missouri, mixed with his passion for computers, culminated in the founding of AgVoice, where he’s aiming to fuse the world of agriculture with cutting-edge technology. As you’ll hear throughout our discussion, there’s serious potential when it comes to arming our farmers (the largest cohort of global workers) with new technology, particularly voice technology as it appeals to the nature of so many agriculture jobs that are not hands-free.
We kick things off with a background of how AgVoice started and what the company is ultimately striving to do with its technology. The Atlanta-based company’s mission is to increase the efficiency of a wide-variety of farmers and ag-based workers by offloading some of the more tedious work, such as data entry, to AgVoice’s custom-built voice system that’s capable of capturing data on-the-go, in real-time.
As Bruce tells me, there are tons of different types of ag jobs that are “labor-intensive, hands-busy” and require two people – one to gather the metrics and one to record them (often with pen & paper). For example, farmers working in a corn field who are inspecting their crops for a variety of metrics on the corn (tenderness, sweetness, ripeness, etc), have to call out to their partner the scores of each corn metric. According to Bruce, some of these workers will conduct 300-400 inspections per day and at the end of each day, the pair of workers then needs to transcribe each score and log it into their data system.
Now, imagine equipping those two workers with a phone that is running AgVoice and a hearable device to communicate with the system. They’d each be able to run their own inspection, logging everything to AgVoice, and not having to transcribe their notes at the end of the day. As one inspector told Bruce, “This is a thousand times better than what we were using before.”
So, how exactly does AgVoice make this all work?
Well, for starters, one of the biggest learnings that I have had through my research into the voice technology space, is that improvements in natural language processing (NLP) is at the heart of the innovation that voice-based technology is undergoing. This was something I learned from Jeff Adams in episode 30. Interestingly enough, Jeff and the team at Cobalt Speech were sought out by the AgVoice team to help build the custom NLP systems that AgVoice is deploying. It turns out that if you lead the team that built Alexa (as Jeff did), your services become pretty in-demand!
So, by building very domain-specific NLP systems, the AgVoice team is able to build a voice recognition system that is designed entirely around the language used by the agronomists, cattle workers, farmers, and so on. Hyper-specific lexicons derived from the language used by these workers, in addition to NLP becoming considerably more accurate, has allowed for AgVoice’s system to achieve high 90’s and even perfect accuracy in the data generated by the voice input.
Ok, so the system is capable of capturing the data accurately. But, what about the fact that this requires an internet connection to run and we’re talking about a type of worker often working in some of the most remote parts of the planet with no cellular service?
This is where the conversation begins to get really fun, as we start to explore how we might see the internet connectivity dilemma solved for these more remote parts of the world. One way that we’ll likely see this tackled is by much of the processing moving from the cloud to the edge. Edge computing represents on-device processing, meaning that it does not necessarily require an internet connection. The limitation here however is that for high-powered tasks, like running AgVoice on one’s device, it takes a lot of processing power, which the edge might not be capable of supporting (yet).
Enter, SpaceX. As a huge space nerd that can’t get enough science fiction, I’m fascinated by all the cool things that SpaceX is pioneering. Whether it’s landing re-usable rockets and privatizing space, or re-imaging the space suit, I’m all for it. So, when Bruce started to share what SpaceX’s satellite business, Starlink, would enable for AgVoice, my ears perked up. 5G internet connectivity to virtually any plot of land in the world… this allows anyone to run AgVoice on a (cheap) phone connected to Starlink, and suddenly the notion of cellular connectivity seems like an afterthought.
Enabling the global work force of agriculture-type workers with the internet on a cheap Android device is going to have profound effects. AgVoice is just one example of what’s being made possible.
This was one of the most fun conversations I’ve had on the podcast. As I point out at one point during the discussion, its these custom solutions that intrigue me the most about the voice technology space. Alexa and Google might grab the lion’s share of the headlines and are the household names, but I find the companies like AgVoice, Cobalt Speech, Canary Speech, Nuance, Orbita and so on, to be equally, if not more fascinating. Based on the topics discussed on this episode, I’d bet that many farmers and agriculture workers would agree.
– Thanks for Reading –