Bot for a million: how we automated the Dodo Pizza contact center
Telephony bot
DoDo pizza is one of the largest catering chains in Russia
Unload the Dodo Pizza contact center using a voice bot.
Client
Task
Background
In 2019, DODO and EORA launched a pilot porject; a voice bot used to answer topic specific DODO pizza calls
The customer was satisfied with the service and decided to expand the range of requests that the bot could handle.
At first, the bot only accepted calls from users who placed an order
To make it "learn" other topics faster, we began to transfer all calls to the bot.
What calls does the bot handle
01/
Bot already knows
Bot examines
courier late complaint
product complaint
change of delivery method
order cancellation
minimum order amount
order preparation time
schedule
the presence of a children's room
the address
Additional functions
02/
Bot informs customer whether puzzaria is open or closed during holidays
If a number calls that recently ordered, the bot will first provide the order status
Rather than being inputted maunally, the message is no synced to the calender
Then the bot asks if the caller needs anymore assistance (Ex. provide information about the work schedule)
Interesting fact
We used 1000 most common phrases and 1000 lines of dialect from a random sample to help the bot recognize subject matter
Development process
03/
Launched a chatbot in pilot mode
Collected statistics on his work
Launched a chatbot in «combat» mode
Analyzed the «bot-man» dialogues
Added new dialogue scripts
Debugged, launched the bot
Step 01
Step 02
Step 03
Step 04
Step 05
Step 06
We switched to our own services to fast-track the process of development and debugging
Further fine-tuning of the service was completed using the zDialog platform from EORA. For example use of information only option holidays
Tools
04/
We used the following services in our work:
A simple and convenient platform for developing chat bots, voice assistants and contact center automation systems.
The AI service for chatbots analytics provides a detailed assessment of their effectiveness.
zDialog
OneDash
Yandex SpeechKit
Speech recognition and synthesis technology from Yandex
Project team
05/
Project manager
Script editor
Technical Director
Developer
Analyst
Developer
Sergey Ponomarenko
Sergey Verentsov
Alexey Sherchenkov
Anna Michnik
Anton Kapridov
Tatiana Popova
Emotions from the project
06/
«The project is cool, because, firstly, the customer is big and he has a lot of data. Secondly, we can immediately show the result: the person who calls Dodo is talking to our bot».
Sergey Verentsov
Project manager
Statistics
07/
calls per month
processes the bot in «combat» mode
accuracy of recognition of the topic of treatment within the pilot
rubles per month
saving by the bot in «combat» mode
~450 000
85%
> 1 000 000
Quote
«Latest data: your bot recognizes the topic 85% of the time. I don't know how you did it, but this is an awesome result!»
Alexey Medvedovsky
Leader of the IT direction of the contact center «Dodo Pizza»
Plans
08/
We want to teach the boy teh ability to identify a complaint and switch the caller to operators in charge of handling complaints.
Soon the bot will be able to report delivery status using DODO's application when a call is made to the call center.
The range of topics that the bot can workout will continue to expand. We are currently approving them.
Other projects for «Dodo»
09/
Automation of the «Dodo Pizza» contact center. Pilot
Robot for analyzing reviews in the «Dodo» mobile application
Contact us
PRODUCTS
SECTIONS
STUDIOS
Russia, Innopolis, st. Universitetskaya, 7
2023 © EORA. All right reserved