Friends+Family DirectAI Monthly Update (September 2023)
Hey all!
Keeping you up to date with what’s new at DirectAI. To reiterate, we’re enabling anyone to build computer vision models that classify, detect, and track objects in images & videos, without training data. If you missed last month’s update, check it out here!
Last Month’s Rundown
We brought on a fall intern - please join us in welcoming Avi Upadhyayula! He just closed out his fourth week and is already shipping great code.
We didn’t sign a new customer or grow revenue this month. More info on this in the “learnings” section. We did meet with a bunch of leads and build demos for them! We’re working on anomaly detection in industrial contexts, object detection for cleaning robots, boat detection in high-res satellite images, and plant segmentation from infrared camera feeds (to name a few). We’re continuing to solidify our partnership with Ant Media - see their announcement about DirectAI being Product of the Month.
We successfully transitioned all of our production GPUs from AWS to Azure without an outage. What’s more, we brought our API’s uptime to > 99.95%.
What We Learned
On Revenue Growth (or lack thereof)
Learning: Enterprise sales is hard! We have lots of improving to do when it comes to qualifying leads early, having well-timed pricing conversations, and validating the pain of a given problem via constrained trial periods or LOIs. That said, enterprise sales cycles are multi-week and we kickstarted the majority of existing conversations in early September. Lagging indicators are lagging!
On Traction
Learning: After last month’s shift to revenue growth as an explicit focus, we optimized for shipping demos to prospective customers. To the converse of the previous learning, we overshot our goal and are feeling proud of our rate of learning. We’re building specifically for explicit asks rather than assumed needs. We’ve largely stopped pitching DirectAI as a solution in search of a problem.
We validated market desire (and built demos) for edge-deployment of DirectAI’s models on tiny GPUs and semantic segmentation of objects.
Learning: These are both distinct departures from DirectAI’s existing offering that required some fundamental upgrades of our tech. Upon closing a relevant deal, the former would require on-premise deployment, which is generally considered less scalable than a cloud-based solution. We’re currently being pulled in a number of directions simultaneously. This makes us less focused but it’s rooted in explicit asks by prospective customers attached to a willingness to pay.
What’s Up Next
More talking to customers. More demo building. Honing in on a tiny & well-defined customer segment that we can monopolize as a stepping stone to becoming the de-facto solution for all Computer Vision problems.
We’re deep in the weeds on some new tech and we’re getting strong initial responses from prospective clients. Stay tuned!
Asks
Do you know anyone at the following companies? Please help me get in touch with them!
They come from a variety of industries so I’ll pass along forwardable blurbs if you reply! Our shared calendly is here!
Kindly,
Ben
P.S. We’re really excited about GPT-4V’s release!