Commerce.AI started looking into the power of AI data insights in 2018. In the last two years, the software company’s solutions have exploded in utility, in no small part due to its adoption of Microsoft Azure OpenAI Service and Azure Cognitive Services. Layering its own technology on top of these, Commerce.AI unearths new kinds of customer feedback from unstructured video, audio, and photographic data. The company’s technology has already begun driving 30 to 50 percent increases in productivity for its customers, with even greater benefits yet to come.
“With OpenAI models, the answers to our questions are ready the moment we ask them. What we’ve done at Commerce.AI using Azure OpenAI Service is supercharge our customers’ ability to take action based on those insights through automation.”
Andy Pandharikar, Founder and Chief Executive Officer, Commerce.AI
Unstructured data interpreted by AI
As of mid-2023, the world creates quintillions of bytes of data every day. Commerce.AI estimates that 90 percent of this data is unstructured, originating from sources like audio calls, video, and images. Historically, this data has been nearly impossible to fit into the structured rows and columns of the average spreadsheet. But for the last five years, Commerce.AI has used advanced AI to extract intelligence from it, uncover opportunities, and automate workflows. The company recently embraced Azure OpenAI Service and Azure Cognitive Services to develop even more helpful solutions for its customers.
More than simple transcription and description, the services Commerce.AI provides have long included sentiment analysis, entity identification and extraction, and trend analysis. “Businesses spend a lot of time and effort creating case studies and focus groups as they determine what their customers want,” says Andy Pandharikar, Founder and Chief Executive Officer of Commerce.AI. “Yet so often, they lack the sophisticated tools required to derive the answers to those very same questions from customer support scenarios.” The company’s journey began as an effort to create purpose-built, advanced analytics solutions capable of revealing these insights. “OpenAI’s generative models caught everyone by surprise,” adds Pandharikar. “The question then became, ‘What can we build alongside it?’ That’s why we began exploring more Azure services.”
Adopting multiple Microsoft solutions and assuring that they effectively interoperate with Commerce.AI’s own technology was surprisingly simple, according to Pandharikar. “We’d had our own sentiment module, for instance,” he says, “but switching over to sentiment analysis with Azure Cognitive Services was incredibly easy.” The company has already begun layering its own solutions, including custom sentiment analysis models, on top of Azure and OpenAI technologies.
Using AI solutions to deliver immense productivity gains
Commerce.AI brought AI technologies from both Microsoft and OpenAI together with its own to help companies improve productivity in their customer support centers. This Commerce.AI system begins when an agent receives a call from a customer and Azure Cognitive Services uses a custom, industry-specific Commerce.AI model to transcribe the interaction in real time with a high level of accuracy. If necessary, Azure Cognitive Services will also translate the speech into multiple languages as the call takes place. After the call ends, the system applies OpenAI technology through Azure OpenAI to automatically create a summary that includes customer contact information and topics of conversation, with embedded sentiment analysis from Azure Cognitive Services. During this step, the system automatically selects next steps and follow-up action items from pregenerated options. Then the customer service agent exports the information to Microsoft Dynamics 365 in one quick step.
Customers using this Commerce.AI system report an average increase in productivity of 30 percent within their support centers, chiefly based on the total elimination of their agents’ need to write up post-call notes. The monetary savings associated with this increase in productivity can also add up rapidly. “We have a pharmaceutical customer that employs people with PhDs in its support center to assure accurate descriptions of its products and their intended uses,” says Pandharikar. “Each interaction that these employees handle can cost the company as much as USD8,000, so it becomes imperative that that precious time is spent with the customer and not extended for note-taking afterward.”
Other customers’ productivity has increased by as much as 50 percent. Employees at these organizations spend much of their time conducting interviews. Each interview traditionally required two agents to complete—one to ask questions and one to take notes. By adopting the Commerce.AI system, these organizations can retrain their note-taking staff members as interviewers and effectively double the productivity of their staff. “The first step in delivering the immense benefit we’ve been able to deliver for our customers is the contextual interoperation of staff and technology,” says Pandharikar. “The second step is to extract the kind of actionable information from these customer interactions that can then lead to new opportunities or the expansion of their business.”
The breadth of potential
Even before Commerce.AI began introducing Azure AI capabilities to its solutions, the company was uncovering timely opportunities for its customers. For example, by cleaning and parsing a cookie company’s product reviews, Commerce.AI was able to identify the keto dietary movement, then still in its infancy, as a growing and important opportunity. The data was then fed into Power BI dashboards, and the customer developed a keto-friendly cookie in the following weeks. Since introducing Azure OpenAI to its products, Commerce.AI now views this process as comparatively slow. “Using our new capabilities, we can bring insights to the surface in an instant,” says Pandharikar. “With the scalability of Azure and the power of generative AI, we’ve built a data ingestion pipeline capable of steadily training on and deriving value from unstructured, 24-hour customer data streams.”
On top of these data streams sits auto-MATE, a new solution developed by Commerce.AI specifically to work with OpenAI GPT models and Azure OpenAI. If a business wants to better understand its customers, auto-MATE can deliver a database of customer insights derived from internal and external workload-related sources, which the business can then query in depth. For product insights, these sources include external reviews and forums and internal support center calls and chats. The solution has more than 170 connectors, meaning that it can deliver similar insights from complex solutions like Microsoft Teams and other communications platforms. For video and audio calls, Commerce.AI uses Real-Time Messaging Protocol audio hooks to ingest the call as it happens. The solution can then transcribe the audio, translate it, and even play it back in the listener’s native language in real time.
For something like job screenings, the solution summarizes and analyzes each interview. “In job placement applications, we have auto-MATE extract interview questions and summarize answers by candidate,” explains Pandharikar. “The summary also links to each interview’s transcript if hiring managers want to review the original responses. So, you can imagine the productivity gains when we push that entire database into a customer’s system.”
Commerce.AI is already using auto-MATE to complete more than 200 tasks for various customers. Throughout the solution’s development, the company worked closely with Microsoft to align with Microsoft’s principles and practices for responsible AI. “We look to Microsoft as a leader in the development of frameworks for responsible AI,” says Pandharikar. “The Microsoft brand being tied to products like Azure OpenAI Service gives us the confidence that generative AI is coming to market with proper guardrails in place.” As part of the Microsoft for Startups program, Commerce.AI has access to multiple Microsoft support systems, and the company’s employees touch base weekly with a diverse team of Microsoft subject matter experts before it deploys Azure and OpenAI technologies within customer tenants. “As we bring OpenAI technology to more rigorously regulated sectors, we value even more the support available to us from Microsoft,” he adds.
Containerized apps for regulated sectors
A couple years ago, as Commerce.AI began vetting potential service providers for the services and framework that would support its solutions, regulatory compliance and security were major decision-making factors. “We realized that Azure has some of the strongest, most easily demonstrable security out there,” says Pandharikar. “We then came to the conclusion that it has the best HIPAA compliance and documentation, and we doubled down on Microsoft as our service provider.”
Some of Commerce.AI’s customers operate in regulated sectors, including those who need to prove Health Insurance Portability and Accountability Act (HIPAA) compliance. For these organizations, Commerce.AI containerizes its applications with Azure Kubernetes Service (AKS), a managed Kubernetes service that helps increase security and optimize resource allocation through unified management and governance. “We can improve our compliance tracking with AKS and increase security by making sure our data stays within the proper tenancy,” says Pandharikar. “Using AKS also helps us expand our customer base because we’re better able to facilitate the needs of more regulated sectors.”
Interconnected AI workloads will drive the next wave of innovation
Commerce.AI is now looking into how AI solutions can deliver specific outcomes under the supervision of human experts. “With OpenAI models, the answers to our questions are ready the moment we ask them,” says Pandharikar. “What we’ve done at Commerce.AI using Azure OpenAI Service is supercharge our customers’ ability to take action based on those insights through automation.”
Now, if a company discovers that its customers have shifted their priorities, it will be able to use a Commerce.AI solution to quickly alter the descriptions of its relevant products to meet the emergent need. To use the cookie example, imagine that a baked goods company had a product that met Keto requirements but was not described as such. The moment the trend was identified, the company could use its Commerce.AI solution to rewrite and upload the cookie’s description wherever it is sold online, immediately responding to the trend. “Interconnecting and automating workflows is the next frontier,” says Pandharikar, “and with Azure OpenAI Service, we have a powerful resource at our side.”
Commerce.AI’s customers are already lauding its AI solutions as delivering up to 50 percent productivity gains, dramatically increasing efficiency, and delivering new insights and the power to act on them instantaneously. Some also view the introduction of AI assistance to their workflows as an opportunity to transfer employees to roles where they can make a greater difference. “The whole concept of our value is enabling our customers to do more with less,” says Pandharikar. “A new wave of interest is building around technologies like Azure OpenAI Service and Cognitive Services, and we’re very happy to be riding that wave.”
Soon, Pandharikar expects to see companies extracting information from unstructured data and using those insights to automate the creation of new ad campaigns—or potentially even launching new products based on Commerce.AI intelligence. “The cost of using generative AI in business will continue driving downward,” predicts Pandharikar. “Microsoft is constantly coming up with new solutions like Azure OpenAI Service, and as the applications of those solutions broaden, so will the customer base of companies who make the most of them.”
Find out more about Commerce.AI on Twitter, Facebook, and LinkedIn.
“With the scalability of Azure and the power of generative AI, we’ve built a data ingestion pipeline capable of steadily training on and deriving value from unstructured, 24-hour customer data streams.”
Andy Pandharikar, Founder and Chief Executive Officer, Commerce.AI
Follow Microsoft