Philip Morris reached 1.8m customers and improved key KPIs by up to 70%
Channels
Call Center, Rakuten Viber
Industry
Retail

Philip Morris International Inc. is a Swiss-American multinational tobacco manufacturing company, with products sold in over 180 countries
Background
In 2020, when the pandemic hit the world market, Philip Morris stores were closed. At the same time, the number of calls from customers to the contact center increased dramatically.
Philip Morris faced two main problems:
- The company needed to provide customers with a tool to quickly resolve their inquiries from the comfort of their own homes. They also wanted to reduce customer service costs and offload call center agents.
- The pandemic also affected the volume of requests to buy/test devices. Therefore, Philip Morris also needed a tool to handle such requests without agent interaction to minimize financial and human costs.

Goals
Automate the process of handling FAQs
Reduce average handle time
Automate the processing and verification of requests
Minimize financial and human costs
Solution
One of the main goals of the project was to automate communication with customers and reduce the involvement of phone agents while increasing the level of customer service
During the development process, the team encountered various challenges. For example, initially, clients preferred completely different channels of communication — a hotline and a visit to the selling point.
But with the COVID-19 lockdown, customers had no choice but to turn to online communication channels. At the same time, the team even had to release a chatbot a month earlier to reestablish customer support processes.
The chatbot Mark
The chatbot Mark was launched in February 2020. To determine the functionality of the MVP, the company analyzed customer requests and identified the top 5 topics
The team then gradually introduced new scenarios: the ability to track order status and checkout directly within the chatbot and more. They deliberately made the client's journey more complicated because of specific issues related to business regulation and some legal aspects.
5topics
- Device diagnostic
- Online order status
- Trade-in service
- Device registration
- Referral program
The callbot Christina
- 1A user fills out the form on the website
- 2A hotline agent offers a delivery date or a convenient store location
- 3The voice bot contacts the user regarding their order
Next, the development of the callbot Christina began. Its main task was to cover the process of checking the already existing requests to buy/test devices, where clients did not come to the meeting/did not accept the order. Following a successful pilot of the callbot, the company decided to scale the solution in internal and external processes.
Now Christina covers the call center functionality, making first calls, processing requests to purchase the device and collecting customers’ data
Has Christina lived up to the PMI's expectations?
The voice bot's Reach Rate was even slightly higher than the agent's Reach Rate.
The voice bot makes the cost per contact 7 times cheaper than a hotline agent does
Reach Rate of
the voice bot
Reach rate of a
hotline agent
Device purchase
Paid trial
65%
75%
67%
69%
Results
up to 30%
Drop in the average handle time
+17%
Growth of automation of incoming requests through text channels
-27%
Drop in the cost of incoming contacts
x7times
More profitable than a hotline agent
As a result, Philip Morris has provided its clients with convenient tools. The company intends to continue to strengthen its digital presence and expand conversational agents’ functionality.
Customer satisfaction has increased with the quicker service. While Mark and Christina create a positive impression
Let’s have a conversation