Suzan uses First Care mobile app in her phone and the app suddenly detected a abrupt stop or shift in the speed of the vehicle using smart phone accelerometer and the GPS. The insurance company tries to reach Suzan on her mobile to check if she is safe, in case of no response they call 911 and share the GPS coordinates to send emergency services to the site. If she did respond to the call, they would provide her instructions on the next steps.

The accident was automatically detected by machine learning that uses various attributes like change in speed of car, GPS location, local weather conditions, traffic conditions and so on, you can imagine an app like WAZE that can do this.

To confirm that Suzan’s mobile was not stolen the AI based system will check the stolen records and processed Suzan voice using natural language processing and matched against her audio recordings. Knowing she has gone through an accident; Suzan’s voice must be distressed and hence an added authentication is added using combination of voice and text. Suzan is then instructed to walk around the car and take photos for the damage and send to First Care for accident estimates. The images are scanned using computer vision to classify the type of damage and estimates are done using algorithms based on machine learning. The system guides the first care representative to the nearby tow truck facilities which can be sent to the site and Frist Care also checks with Suzan if she requires a rental car.

This automated response that sense the dispatches is driven by an expert system that has learned the rules over time may be guided by human operators. The truck comes by and takes the car to the body shop.

The rental car company picks up Suzan once the car is in the repair shop. The repair shop initiates the repair job and communicates the status periodically with First Care. First Care runs an agent-based software business process management system, to notify Suzan of the relevant updates which she receives on her mobile and email. Suzan can track the progress of the repair and status updates from the notifications received.

The success of this case is very much dependent on data coming from multiple sources like – repair shop, tow trucking company, Suzan’s agent and from First Care itself. The data is fetched from different sources, consolidated, and massaged to be presented to Suzan. This would require integration with multiple systems which can be on a secure cloud and can be accessed using distributed API’s or RPA.

This customer experience case is surely a very important one and may look simple but it requires the business to account for all the what-if scenarios. Building this type of customer experience requires a solid architecture and the following must be considered: –

• Actors involved in the complete process and their roles.
• Clear and concise business process definition.
• What technology components to be used and where to use them.
• Data coming from different sources like – sensors, mobile accelerometers etc.
• Data storage and management.
• Interaction with different business partners like – Rental car, Tow trucking company, insurance agents etc.
• Easy, simple, and robust architecture which is capable to handle numerous variations of a scenario that can happen.

To conclude, businesses must be on top of their game to provide the best of the customer experience by including the high-end technologies available today. Artificial Intelligence is a broad subject and cannot be just applied without thoughts and in isolation. For AI to work at its peak, there are lots of dependence on the data, business partners involved and the leadership vision to take their businesses to the next level.

In conclusion, AI cannot simply be deployed in isolation, it needs architecture.