As the healthcare environment continues to rapidly change and with the increasing role of digitization, Johnson & Johnson is transforming its business by embedding Intelligent Automation into the fabric of the organization. The opportunity to reimagine work is significant and has the potential to empower and equip employees with new skills, capabilities, and the freedom to focus on more fulfilling, engaging and purpose-driven work.
By using a common chatbot framework along with the Azure Bot Services, employees without technical training or experience can now build their own bots to serve their teams and customers at a fraction of the time and cost it took to develop previous chatbot projects. Ten chatbots have already been deployed and dozens more are currently in development. Expanding Johnson & Johnson’s use of bots and adding intelligent automation into its practices is keeping Johnson & Johnson on track to achieve its ambitious plans for digital transformation.
As Douglas Lima, Johnson & Johnson’s IT Lead of Intelligent Automation in Latin America, listened to staff in the legal department discuss how much time they spent answering the same questions for Johnson & Johnson employees, day after day, he knew the solution: The team needed a chatbot.
Creating a new chatbot typically required Lima to sit through close to 20 meetings while his team wrote more than 70,000 lines of code. But this time would be different. He turned to a new tool—appropriately named Genie the Genius—built with the Microsoft Azure-based chatbot framework. Aravinda Boyapati, Intelligent Automation Leader at Johnson & Johnson Technology (JJT), and his team had recently created Genie to deliver chatbots in days instead of weeks and at a fraction of the previous cost.
Lima used the Genie framework and a visual-based bot builder to create Lexi, a chatbot to serve as the department’s front-line customer service contact. Over the next few months, Lima saw the number of Lexi users consistently increase. Employees were not just visiting the chatbot once but were making it part of their regular workflow. Because of the positive experience and service Lexi provided them, employees kept returning to the bot for answers to their legal questions.
Lima says Genie the Genius was the perfect solution for the legal department’s dilemma. After Lexi’s deployment, it fell to the legal team to keep their data up to date by adding new questions and changing the appropriate data replies to make sure users continue to receive the right information. Previously, Lima would have spent days training employees how to maintain their data. Even though the legal department employees had little or no coding experience, this time the training took only 30 minutes and then the team was off and running.
Lima’s and the legal department’s experience isn’t unique. Business units across Johnson & Johnson need real-time information to share internally and externally so that employees and customers can get the answers they need to make informed decisions. Over the past three years, Johnson & Johnson has focused on transforming business units across the globe through intelligent automation to make Lexi’s success a reality.
Innovating together: A botathon event shows proof of concept
Johnson & Johnson is headquartered in New Brunswick, New Jersey but has operations around the world. Chatbots can be a key way to transform how the business operates globally, as they offer the ability to improve employee and customer experience, effectiveness, and efficiency by giving more than 130,000 employees access to real-time data. However, because the company must follow the regulations of global regulatory agencies, developers had to include several extra steps related to security and code validation, which increased the cost of bot creation. With an average cost of $116,000 for a custom chatbot, the process was too expensive to launch on a widespread scale.
In 2019, Johnson & Johnson collaborated with Microsoft during a multiday botathon event to begin exploring how the company could more cost effectively use chatbots to share information. During the event, JJT developers and leaders worked with Microsoft product owners to design a platform and outline the processes needed to meet their goals. When Boyapati and his team returned to the office, they were armed with key pieces of the solution, including a Johnson & Johnson chatbot store and key metrics to help train the chatbots to deliver more accurate information.
Boyapati says the highlight of the botathon was developing a bot builder that enabled teams like Johnson & Johnson’s Latin American legal team to create no-code/low-code chatbots like Lexi with little training. Because the bot and dialogue builder use a visual interface, employees don't need to know JavaScript or .NET. Instead, they use their own business and subject knowledge to design how conversations should flow and then use drag-and-drop tools to create the dialogue. “The botathon with Microsoft was the true turning point for us to ideate, shape, and deliver the Genie the Genius project,” says Arti Patel, Director of Intelligent Automation at Johnson & Johnson.
Creating digital human capabilities
After the botathon, Boyapati pulled the business teams together for a meeting and asked departments to bring an idea for a chatbot to help serve their customers. During the meeting, Boyapati reviewed platforms available to use for chatbot development. “Because we were already using Azure, including ML, Databricks, and IoT, it was a natural choice to continue using the Microsoft Azure platform without making additional investments,” says Mrunal Saraiya, Sr Director and Intelligent Automation Leader at Johnson & Johnson.
In addition to security and functionality, Boyapati wanted Genie to have a personality, so customers would have a positive experience and get the information they needed. “A person talking to the chatbot should feel the same way as when talking to a human, which is what we strive for with digital human capabilities,” Boyapati explains. “You want the chatbot to show empathy and feel like you're getting help from a real person instead of talking to a bot.”
Developing secure and conversational chatbots
Boyapati’s team got to work developing Genie the Genius, a robust and flexible end-to-end, enterprise-grade chatbot platform, using Azure Bot Services. They also incorporated Language Understanding (LUIS), a natural language understanding AI service, as well as QnA Maker, a cloud-based API that helps create conversational question-and-answer layers over existing data. “Microsoft really takes customer feedback seriously and brings those suggestions into next-generation chatbot technologies,” Boyapati says.
Throughout the development process, security was a top priority and concern. To meet the stringent security requirements of regulatory authorities, each Genie chatbot runs on its own service, meaning it doesn’t share resources with other chatbots. If a breach or attack should happen, the security response is designed to contain the breach to the affected tool. Additionally, JJT uses the latest security standards and controls, such as SSL, authentication, and authorization.
Expanding chatbots across the organization
Within six months, JJT deployed 10 chatbots into production. An additional 50 chatbots are currently in development for both internal and external use. Departments worldwide now employ chatbots for a wide range of uses such as addressing customer returns, answering product-related questions, and checking the status of change requests and change orders in ERP systems.
Boyapati notes that instead of spending time writing code, his team now supports employees that want to become citizen developers to help launch their initiatives. He sees firsthand how empowered employees can identify and solve their own business challenges. “With our new Genie platform and low code/no code capability, we shortened the time from 145 days from start to finish to 33 days,” says Boyapati. “This shortened time frame is a large part of the reason that we reduced the cost per chatbot from $116,000 to $26,000.”
When the Information Security Risk Management team began building a chatbot to reduce the number of emails they receive daily, Boyapati helped onboard the team onto the platform. Within days, they learned the skills they needed, then developed the chatbot on their own. “We have many chatbot initiatives going across each department right now,” Boyapati notes. “We are keen to leverage the chatbot as a common mechanism to speed up and improve overall operations.”
Many Johnson & Johnson employees use multiple chatbots throughout their workday, and JJT needed a way to make it easy to access those chatbots and get information from them. Instead of having to launch a chatbot from the web or a mobile device, employees can access the chatbots to get insights and data while in Microsoft Teams. Because employees already use Teams to attend meetings, access files, and communicate both internally and externally, they spend less time looking for the tools and prefer to turn to the chatbot rather than an email or phone call.
Building the future of intelligent automation
Genie the Genius is just the beginning for Johnson & Johnson. The expected deployment of 50 chatbots by the end of 2022 has the company swimming in ideas. And by the end of 2023 they plan to have 100 total chatbots helping with internal and external operations. What's more, the chatbots will ultimately be working together with other forms of intelligent automation.
Johnson & Johnson expects to create $500 million in value as it expands its end-to-end business automation initiatives. Boyapati explains if an employee uses Lexi to get legal information instead of spending an hour with a legal team member, the company saves two employee-hours: one for the internal customer and one in the legal department. With 130,000 employees, those hours can add up quickly.
Boyapati notes that he views chatbots as more than just a tool or technology. They are a fundamental shift in how people share and receive information. He says that customers tell him daily that they don’t want to spend a lot of time browsing or navigating. They want information at their fingertips through either voice or chat. “I strongly believe chatbots are changing the face of how we interact with our customers and partners, both internally and externally,” he says.
“Microsoft really takes customer feedback seriously and brings those suggestions into next-generation chatbot technologies.”
Aravinda Boyapati, Intelligent Automation Leader, Johnson & Johnson Technology
Follow Microsoft