Business Overview: Abbott is a prominent global healthcare tech giant known for pioneering breakthrough products in diagnostics, medical devices, nutrition, and branded generic pharmaceuticals. With a workforce of over 103,000 dedicated employees, the company operates in more than 160 countries, striving to make a lasting impact on global health.
Committed to their vision of providing access to healthcare, community engagement, and science education, our client has been driving positive economic, social, and environmental impacts through their business practices. Despite their remarkable achievements, our client faced critical challenges in effectively supporting their vast and distributed workforce. The need to assist and engage their field personnel, resolve queries promptly, and optimize support executive’s time were paramount concerns. They aimed to elevate employee experience and satisfaction through personalized engagement while increasing user interaction with innovative solutions.
The Challenge: Abbott’s workforce, consisting of thousands of executives responsible for sales and other vital roles, required constant access to up-to-date information about the company’s brands, competitor brands, and their individual performance. However, the existing methods of retrieving information, either through the company portal or via phone calls to the head office, proved to be cumbersome and inefficient.
Field personnel faced significant challenges, including:
- Time-Consuming Query Resolution: The process of accessing relevant data from multiple backend systems was laborious and time-consuming, leading to delays in query resolution.
- Inefficient Support Mechanism: Support executives spent a considerable amount of time handling routine queries, diverting them from more strategic tasks.
- Inconsistent User Experience: The lack of a unified support platform resulted in inconsistent user experiences, affecting workforce engagement and productivity.
- Slow Onboarding: New personnel struggled with lengthy ramp-up times due to the difficulties in accessing critical information swiftly.
To tackle these obstacles and optimize their field force support framework, the healthcare giant sought a comprehensive and intelligent solution that could empower their distributed workforce, streamline support processes, and significantly improve workforce engagement and experiences.
The Solution – A virtual assistant called, Maya
In response to this challenge, SmartBots stepped forward with an innovative new virtual assistant to support and engage Abbott’s pan-India sales, and her name is Maya.
Maya acts as a personal assistant to the employees, providing sales operations support, and keeping them ready for the day, providing access to contextual information at their fingertips.
Maya is powered by AWS Language Understanding service – Amazon Lex
SmartBots team has custom designed the architecture of Maya to fit into the existing enterprise system of Abbott. The design has two key components
Maya package
Smartbots NLU Engine (for language understanding)
Maya Architecture
| Sl No | AWS Service | Usage |
|---|---|---|
| 1 |
Amazon API Gateway |
Scalable and secure API endpoint to access NLU Engine |
| 2 | AWS Lambda |
Lambda is used in two instances:
In both the instances Lambda holds the logic developed by Smartbots team. Lambda ensures continuous availability and scalability of the service. |
| 3 | Amazon Cognito | Used for user management and role and policy definition (configured in IAM). |
| 4 | Amazon Lex | Language model definition |
| 5 | Amazon Comprehend | For sentiment understanding |
| 6 | Amazon DynamoDB | For session management and response mapping |
Smartbots NLU Engine
The NLU engine has three parts
Conversation Manager
Language Model
Business Logic
Conversation Manager is the entry point to the NLU engine which handles the flow in and out of the NLU engine. The logic resides in AWS Lambda service.
Language Model is where all the intents and entities of the bot are defined. Maya’s language model is built using Amazon LEX. For the given user input, LEX identifies the right intent and extracts the entities from the user input.
Business Logic layer handles response generation.
Results The implementation of Maya yielded encouraging results for the company:
- Improved Query Handling: Maya now answers 32% of the monthly queries raised by the workforce with a 74% success rate, reducing wait times for resolution.
- Enhanced Workforce Efficacy: By swiftly accessing key information from multiple backend sources, Maya increased the response rate and efficacy of the field force.
- Increased Productivity: Mayaaddressed close to 100,000 queries, leading to a 50% gain in employee productivity.
- Streamlined Onboarding: New personnel experienced reduced ramp-up time due to the quick and easy access to information provided by the Maya.
- Resolved over 90% of queries successfully
- Reduced the average query resolution time by 50%
- Enhanced Employee Engagement: Maya contributed to improved employee engagement and experiences, with 80% of the field force rating it as helpful.
Highlights of Maya:
- A powerful AI-powered chatbot that can answer questions, resolve issues, and provide support to employees in real time.
- It is integrated with the company’s legacy systems, so it can access and process information from a variety of sources.
- Has been a huge success for the company, improving employee engagement and satisfaction, and reducing the number of support tickets that are escalated to human agents.