AWS offers glimpse of what CIOs can do with Amazon Q Business

Amazon Web Services gave IT leaders a real-world look at what enterprises are doing with its Amazon Q Business generative AI assistant, showcasing implementations from customers such as Zoom and HS Brands at an event last week in Boston.

Among the use cases highlighted was the ability to make enterprise AI agents accessible to external services, which was piloted by the NFL during last month’s draft, said Natalie Hirsch, senior worldwide specialist on the Amazon Q Business team.

Amazon Q Business Workflow, which the company announced in December, was also highlighted as “coming soon” to help automate tasks and integrate with various business applications, and Hirsch intimated that Q Business may have agentic capabilities in the future.

“What we’re hearing from customers is they want to automate more and more,” she said. “They want an agent-like experience that can help users save more time with those repetitive tasks and a feature that is coming really soon is the ability to build workflows based on documentation or recordings.”

HS Brands to build on early Q Business success

Mystery shopping service HS Brands Global is one company deftly deploying AWS generative AI tools to enhance its service quality, save time, and bag more clients, in conjunction with Ironside Group.

HS Brands, which spends roughly $250,000 on proofreading annually, partnered with Lexington, Mass.-based analytics and AI consultancy Ironside Group to develop a replacement proofreading app using LLM and AI features in Amazon Q Business. Amazon Q, introduced in November 2023, brings advanced generative AI technology to Amazon QuickSight business intelligence service for the cloud. Amazon also offers Q as part of its Q Developer coding assistant.

The Las Vegas-based HS Brands, which measures customer experience and provides brand protection by evaluating roughly 50,000 shops for vendors annually, anticipates the Ironside Q solution will save the company more than 85% of its annual proofreading costs and identify more inconsistencies and errors than humans to improve the quality of its reporting and feedback, said HS Brands founder and CEO Tommy Mills.

“Mystery shopping is a funny little industry. We’re considered a piece of market research without really doing a lot of market research. We are the people that verify things,” Mills explained. “We send somebody in acting as a customer, and then they report upon what they saw. We usually don’t give a lot of opinions. We just say this is what happened. And then it’s very statistical after that.”

Mills launched a web-based mystery shopping platform called Sassy in 2005 on AWS and grew it into a global service provider in 12 countries. He later sold the platform and concentrates on delivering AI-driven analytics for real-time customer insights.

Client Buffalo Wild Wings hired HS Brands to determine whether the quick-service restaurant’s burger sliders are built the same way in every location. Amazon retail is another one of HS Brands customers.

“Amazon hires us to go onto the site, order particular products from particular providers, have them shipped to people, and we’re measuring everything [during] that process,” Mills said. “Does it arrive? Is the box in the condition it’s supposed to be?”

As part of the evaluation process, HS Brands also takes photos, uploads them, and measures as many aspects of the product or service deliverable. HS Brands clients range from one pizza empire that wants to ensure its pepperoni pizza slices are made with precisely 8.5 to 11 slices of meat, to big casinos that require a mystery shopper to spend three days documenting the entire experience from shopping, to eating, to gambling, to parking.

Mills said he turned to Ironside Group to apply advanced analytics to its Sassy platform. As part of the solution, data is moved into AWS QuickSight, which gives account managers of HS Brands’ customers a vast array of dashboards to view data collected. The solution reduces those account managers’ time burden for analysis while improving the quality of HS Brands’ service, which not only helps its customers with their businesses but helps HS Brands get paid faster and grow their customer base, Mills said.

HS Brands and Ironside worked together to develop the Q Business proofreading application, which in addition to saving money, enabled the company to redirect its proofreaders toward identifying other ways to apply AI techniques to company workflows and grow the business more.

“AI is exploding everywhere, and like everyone, we’ve got to get involved. This is going to be a major, major jump forward for us,” Mills said, adding that, although HS Brands is in the early stages of AI use, he plans to apply Q to QuickSight to enable account managers to grow and manage more accounts. “It’s going to save money. It’s going to allow me to bid differently. It’s allowed me to walk into the door with an account and show them that we’re playing at a different level and that we’re going to be more accurate, faster, and cheaper. It checks off a lot of boxes for HS Brands. “

With Amazon Q in QuickSight, business analysts have access to a generative BI assistant they can use to build BI dashboards and visualizations and to conduct complex calculations in minutes, for example.

Toward agentic AI

Amazon also pointed to other customers such as Abstracts, which formulates flavors and fragrances sold to a variety of industries. Abstracts uses QuickSights and Q to consolidate and query data.

“Their challenge was that they had infrastructure with data-trapping silos across all these different SaaS applications, and it’s very difficult to combine data and have a holistic view of the business. And they also wanted to monetize the research data,” said Lucinda Linde, principal data scientist at Ironside, adding that the company centralized its data so that it could be accessed by the QuickSight solution it eventually built.

“One of the key aspects of this was to use the generative AI within QuickSight, such that you don’t have to have AI [expertise], you don’t have experts in SQL, to be able to do the sort of query of the data” that they wanted to perform, Ironside’s Linde added.

Amazon claims this saved Abstracts quite bit of manual effort, made their employees more efficient, and “reduced the burden on the analysts and subject matter experts,” Linde said.  

Michael Davis, an analyst at GAI Insights who attended the event, said that use of enterprise gen AI assistants in business workflows is particularly useful for startups and SMBs to aid in the delivery of their core business value.

HS Brands is “a good example of how a firm is pioneering the use of gen AI to make work more efficient, save time and money, and free up people to focus on higher-value work,” Davis said. Empowering the account management team at HS Brands to have better conversations with clients without as much admin work and using gen AI to assist with proofreading are good examples of this, he added.

Amazon Q offers more than 40 connectors to allow customers to integrate it with many business applications. Another useful product is AWS Index, a data store used by AWS Resource Explorer to store data about all the company’s AWS resources. Users can create an index in AWS Region to launch AWS Resource Explorer.

Zoom, another major AWS customer, built its custom AI Companion within Zoom Studio to enable customers to create meeting-summary templates that connect to external sources, data sources, and services from third-party applications.

Michael Bellezza, AWS GTM alliance lead at Zoom, said the integration of Amazon Q Business and the Amazon Q index are major productivity boosters.

Zoom is currently piloting its custom AI Companion with a large customer for incident management, vastly reducing the time for data entry. “This is where the real work begins. What used to take a support engineer about four hours to do is automatically done once that meeting is wrapped,” Bellezza says. “Custom agents are coming soon.”