When it comes to artificial intelligence, fintech company Supernova Technology is in rare air.
Supernova is part of the 13.8% of American tech companies that use AI in their businesses, according to a recent survey from the United States Census Bureau. Yes, despite all the chatter about the revolutionary potential of AI, many companies are still working out how to implement it in their daily work.
For Supernova, the answer was to build its own large language model (LLM) and let its team get creative with it. William Reinhard, a software engineer II, said a hackathon led to the creation of a new product that extracts financial data from PDF documents. Reinhard told Built In that the LLM is also used to improve communication between teams with vastly different focuses, such as engineering and investing.
“The latest AI technology builds bridges between groups with different domain expertise by turning complex concepts into simple analogies and metaphors,” said Reinhard.
Supernova is a prime example of how early adopters are using AI, with a focus on generating real value for customers and enabling teams to work more efficiently. That said, it’s far from the only Chicago tech company taking this approach. Continue reading to learn more about how Supernova and two other Chicago tech companies walk the walk and use AI in their daily work.
Supernova Technology builds software that automates securities-based lending.
How is Supernova embracing AI in its operations and making it an integral part of your company culture?
My favorite part about working for Supernova is the company’s focus on learning and exploring the newest technology to help solve user problems, and our embrace of new AI technology is a great example. This summer we hosted our first AI hackathon, challenging participants to meet our business needs with AI. My team’s ideas led me to drive the development of Prism, our new enterprise data extraction solution. The product expansion team took the innovative technology from the hackathon to a marketable product in just a few months.
Prism blends optical character recognition capabilities with a Supernova-built large language model to read PDF documents, extract and edit key information in a streamlined web interface and output results to a simple spreadsheet. We are continually adding functionality as we go to market, including new document types and integration mediums, and are excited to launch. We know how tedious data entry can be for our platform users and have built a product to solve that problem.
In what ways does Supernova leverage its resources and expertise to implement AI in different areas of your work, and what specific AI technologies or applications excite you the most?
As a software engineer on the product expansion team, my responsibility is to take an existing business or technology need, evaluate new technologies that offer a solution and build a proof of concept. Leveraging our deep domain expertise in the banking- and securities-based lending industries, we use AI to develop tools that have a real-world impact on our organization and partner institutions.
One technology that really excites me is LangChain, a tool we use to aid in the code development process. LangChain supplies a framework to prompt an LLM to answer questions that, when aggregated, answer a higher-level question. Developing a workflow centered around breaking a big question down into a series of simpler, consecutive questions enables the model to be context-aware, thereby increasing accuracy.
After the model processes each question, the user can refine the output, further minimizing the number of contextual assumptions made by the model. Keeping the end user in mind, maximizing accuracy and streamlining root cause analysis squarely aligns Supernova’s AI initiatives with our partners in the world of finance.
One technology that really excites me is LangChain, which supplies a framework to prompt an LLM to answer questions that, when aggregated, answer a higher-level question.”
What are the benefits of incorporating AI into Supernova’s work, and are there any challenges or considerations that need to be addressed when implementing AI?
While AI continues to speed task completion and decision-making, it can also help to solve one of the key problems every organization faces: communication between teams. For instance, recently my team had to configure our network stack at our data center. To supplement our knowledge, we asked our LLM to explain a networking protocol, which it described in the relatable context of two individuals having a conversation. Similarly, LLMs have helped software engineers understand portfolio risk analysis strategies.
Data security becomes a consideration as more people use AI tools. Public models use the data they are fed to improve continuously, which can expose the data in unexpected ways. Supernova’s clients trust us with their data because of our commitment to security, leading us to train and host our own private models. Tools like Prism are a more secure alternative to free, public models when uploading sensitive information.
SDI Presence is an IT consulting firm and managed services provider.
How is SDI embracing AI in its operations and making it an integral part of your company culture?
On-premises and hybrid managed services are a significant part of our business, with most of our professional services consultants supporting those services for our customers. We deployed artificial intelligence for operations as part of our managed services, enabling us to collect data, monitor and preemptively resolve issues.
We can integrate the ticketing system in ServiceNow with AIOps. Amazon Connect, an omnichannel contact center solution operating out of our SDI Innovation center located in the South Side of Chicago, is rich with several AI capabilities such as automated call routing, natural language for chatbots, transcription, translation and sentiment analysis.
For example, the AIOps infrastructure reduced the number of duplicate alarms and extraneous tickets by 50% for a major Midwest city. It reduced the time to action by 45%, with the tier-one team resolving 70% of incidents. AI services are so inherently embedded in many of our solutions and partner products used in our day-to-day operations that they sometimes are not touted as AI.
In what ways does SDI leverage its resources and expertise to implement AI in different areas of your work, and what specific AI technologies or applications excite you the most?
Our partnerships give us access to services we can assemble like LEGO bricks to build something new and exciting for our customers. Simple solutions like ingesting paper documents, extracting data from them — including handwritten text — and storing them in a searchable database are effective when we realize that paper documents still exist in public and private sector organizations.
Similarly, AI services used to analyze images or videos to detect incidents or threat elements — object recognition and not necessarily facial recognition — provide deterrence or allow faster response times to physical threats. This is a critical use case in the public safety vertical. One exciting use case in the utilities vertical or any vertical with manufacturing, distribution or field services management is the use of AI for preventive maintenance and response operations related to the condition of equipment or distribution lines or leakage.
Our partnerships give us access to services we can assemble like LEGO bricks to build something new.”
What are the benefits of incorporating AI into SDI’s work, and are there any challenges or considerations that need to be addressed when implementing AI?
As with any technology, there should be guardrails on its use or adoption. There is considerable debate on AI and implementing policies at all levels, both public and private. The White House has published a blueprint for an AI bill of rights, which it describes as “a set of five principles and associated practices to help guide the design, use and deployment of automated systems to protect the rights of the American public in the age of artificial intelligence.” Overall, as long as AI is used ethically and for the betterment of humanity, its benefits far outweigh its threats.
dscout is a video-based research platform that helps companies better understand consumers.
How is dscout embracing AI in its operations and making it an integral part of your company culture?
Well-known brands already come to us to conduct their research on using AI in their products as well as to answer questions on the technology’s implications for the research industry. We believe we have a responsibility to bring a research-forward approach to the adoption of this technology, and we have a dedicated team doing just that through their own primary research.
The team is focused on understanding customer pain points and how AI can help while keeping transparency, traceability and security in mind to ensure AI-generated content is clearly labeled and that our zero data retention policy is adhered to.
At our recent Co-Lab event, we brought nearly 100 dscout customers together to discuss, learn and challenge ourselves on the risks and opportunities in recent AI advances. We covered AI literacy, ethical challenges and embracing change. In addition, we encourage employees across teams to experiment with the technology to develop a better understanding of the risks and benefits in their specific focus areas.
We believe we have a responsibility to bring a research-forward approach to the adoption of this technology.”
In what ways does dscout leverage its resources and expertise to implement AI in different areas of your work, and what specific AI technologies or applications excite you the most?
We are in a unique position as we serve the user experience research community, which is deeply concerned with ethics in technology. Because generative AI technology is so new, rapidly evolving and powerful, we believe that AI needs UXR to fulfill its promise. Conversely, UXR needs AI in order to succeed, build influence and keep business human as development cycles become shorter and shorter. Thus, our resources and expertise are crucial to that research-forward approach and help not only ourselves but other curious companies navigate this new frontier.
For UXR, the applications of AI are nearly endless. A few of the most exciting examples are its potential to improve the user experience, reduce the time needed to turn data into business insights and increase qualitative, data-backed decision-making. Knowing that out-of-the-box AI often takes many iterations to get the desired result, prompt engineering and fine-tuning to achieve hyper-focused goals will be critical to help us use AI as a research assistant to follow our philosophy of working “with you, not in place of you.” dscout provides users with an AI sous chef in the hypothetical user research kitchen.
What are the benefits of incorporating AI into dscout’s work, and are there any challenges or considerations that need to be addressed when implementing AI?
The potential benefits are endless. The ability for nearly anyone to improve both their quality of work and their throughput is a compelling enough reason to start experimenting. But there are key considerations that illustrate a need to slow down in order to speed up, to not rush work out the door just to say we did.
Critical issues of AI literacy, ethics, data privacy, bias, human-AI relationship dynamics, potential shifts in role responsibilities and more warrant deep, thoughtful research to ensure the use of AI keeps humans in mind. It’s a tool to solve pain points and facilitate desired outcomes.
dscout is addressing these challenges openly, proactively and jointly with our customers. And with our expertise and clear mission to “make a more human tomorrow,” we are well-positioned to support customers, employees and consumers alike in all aspects of the challenges at hand.
