Last week, SparkCognition co-hosted an enlightening learning session with the Conexxus Innovation Research Committee (IRC). The committee members participated in practical demos and deep-dive discussions of high-value use cases for convenience stores and gas stations delivered by our computer-vision-enabled solution to improve business intelligence, safety, and security for retailers: SparkCognition Visual AI Advisor.
Conexxus is a non-profit, member-driven technology organization dedicated to developing and implementing standards, innovation, and advocacy for the convenience store and retail fueling market. IRC members lead the organization’s effort to identify and research emerging technologies and related innovations that promise future impact on their industry. The Conexxus IRC team organized a two-day “field trip” to SparkCognition headquarters and HyperWerx following conversations that began at the 2022 NACS Show in Las Vegas in early October.
Demystifying computer vision: From possibility to profitability
The first day kicked off with a “Computer Vision 101” presentation by SparkCognition’s Chief Science Officer, Dr. Bruce Porter. Professor Porter provided a level-set on what computer vision is, where and how the technology is being used now, and how industries would scale the considerable advantages of visual AI solutions. Real-world examples were shown to IRC of practical uses of Visual AI Advisor in the field, where first movers successfully navigated from possibility to profitability.
IRC chair Doug New had helpfully provided video clips from one of his company’s convenience stores before the meeting for example purposes. SparkCognition demonstrated those same video streams running Visual AI Advisor so that the committee could observe our product serving common retail use cases. That demo galvanized the room, as members offered up numerous additional use cases in which Visual AI Advisor could address vexing challenges traditional camera and DVR systems struggle with today.
This topic of conversation continued in earnest the rest of the afternoon at SparkCognition headquarters. The two parties compared notes on how retailers could leverage visual AI technology in their stores, why it represents such a paradigm shift for retail operators, what roles it would impact the most, how to involve partners, and more. Among the key takeaways: the breadth and depth of opportunities to deploy visual AI software on current camera equipment to improve customer experience, mitigate retail labor challenges, and get better value out of IoT and data representation.
Demonstrating Visual AI Advisor: Real-world, real-time value at a local gas station
Early the next day, IRC members met at a bustling local gas station where Visual AI Advisor is deployed on the operator’s existing camera infrastructure. SparkCognition CTO Sridhar Sudarsan presented the store’s dashboard view of data analytics, including metrics such as:
- Dwell time within the store by time (who’s staying in the store 15 min or less, half-hour, etc.)
- Demographics (interestingly, the store attracted more women than men)
- Endcap analysis (how many people are picking up items on a given endcap display, capturing time spent vs. another part of the store, etc.)
- Turnaround time at checkout (how many people are in the queue, how long is checkout at the counter, flagging store managers when service time is too high)
- Forecourt counts (how many cars vs. trucks came in)
- Forecourt to store behavior (how many people are fueling and coming in, etc)
- Efficiency trends (how long people are staying at the pump)
- All the alerts that are configured to send to store managers (and how/which CAPA alerts are dispositioned)
- Heat maps of activity in the store
Sudarsan explained how all of the high-value use cases and analytics were implemented in less than one week, thanks to Visual AI Advisor’s easy-to-deploy, no-code/low-code deployment model. Several IRC members asked follow-up questions to make sure they heard that right!
On to HyperWerx: More practical use cases and deep-dive discussions
IRC and SparkCognition then traveled to HyperWerx for even more demos, discussion, and a working lunch. Chief Designer Milton Lopez and the team at HyperWerx led Conexxus IRC members through a series of real-time demonstrations of Visual AI Advisor use cases for retail, showcasing how cameras running our visual AI software can automatically recognize dozens of safety, security, and business analytics scenarios inside and outside the buildings.
Inside a mock store environment complete with shelf inventory and checkout areas, Lopez walked through how Visual AI Advisor keeps track of the exact number of wine bottles on a shelf to combat pilferage, monitor stockouts, and flag goods placed on the wrong shelf. He then transitioned into queuing activity analytics and customer ID verification. Before the demo portion of the day was done, the IRC team saw use cases where Visual AI Advisor alerted on weapon detection, person in distress, near-miss safety hazards, unauthorized access, and much more.
After lunch, we were privileged to listen in as Conexxus’ Executive Director Gray Taylor led the energized IRC team through a brainstorming and information-processing session. Members discussed what use cases stood out as most impactful to their stores, where they would start, and how they would position the ROI of visual AI technology to stakeholders in their business. From our point of view, it was incredibly insightful to hear these nationally-recognized leaders in convenience stores and fueling stations methodically work through the foreseeable scaling process of computer-vision solutions across their industry and in their own stores.
“Game-changing for retailers”
Reflecting on the experience, Doug New said: “It’s been a learning adventure for me. I really didn’t know what to expect when we first started talking to Sridhar and some of the others. But what was most interesting was what we saw [at the convenience store], the insights, the dashboard that was shown to us for the video that was applied to their internal camera system in the store, and the insights they were able to glean from that. I think that’s remarkable and game-changing for retailers.”
When asked what insights he came away most excited about, he said: “I don’t know if I can pick one because there seem to be several, a lot we identified, I think yesterday in our meeting, 50 or so use cases, things that people in the room thought were important. Most of them came back to labor—augmenting labor. We’re not the only industry having difficulty hiring people and finding people. So instead of filling all the open positions, this kind of technology can help augment the existing staff and still get the same kind of things done.”
IRC member Ed Collupy talked about redefining the value of cameras in retail: “To me, the excitement is really around how existing infrastructure and ecosystems, like cameras that are underutilized, can be brought to life and become a tool … So for the convenience and petro industry, we’ve always come together as groups. [It’s about] the networking of working together and sharing ideas. [This] really provides an extra opportunity of hearing from folks that are on the leading edge in developing technologies that can be put to use—immediately—just like we saw this morning at the store.” He continued: “The added bonus of being able to come to a place like SparkCognition has really been a plus to see how you’re putting things to life and how then it can be put to use for the convenience retailer very quickly.”
IRC member Jeanette Lee spoke about the value of sharing these experiences with her industry colleagues exploring emerging technology, offering: “We have to keep changing. The world is evolving. Consumers want something different. We want something different. And together, we need to elevate these services with technology and AI with the infrastructure we have.”
When all was said and done, both SparkCognition and IRC learned greatly from each other during the two-day learning session. We want to thank Conexxus and the Innovation Research Committee for the opportunity to co-host this event and applaud their work moving the convenience store and retail fueling market towards difference-making solutions like SparkCognition Visual AI Advisor.