Humans vs Machines? The future of customer service · Humans vs Machines? The future of customer...

Daniela Braga, Founder and CEOJoao Freitas, CTO

daniela@definedcrowd.com; joao@definedcrowd.com

Humans vs Machines?

The future of customer service

Seattle - Lisbon

5/16/2017

definedcrowd confidential 2

definedcrowd confidential 3

definedcrowd confidential 4

Consumers don’t like to deal with IVR

49%

0.0%

10.0%

20.0%

30.0%

40.0%

50.0%

60.0%

Better for me, saves metime.

Better for the company,saves them money.

Better for both of us. Not better for thecompany or me.

Customer Opinions on IVR, Interactions LCC (408 questionnaires), in SpeechTek 2012

definedcrowd confidential 5

Main IVR issues

• Same technology of 20 years ago – command and control

• Inability to understand beyond the commands

• List of consumers’ issues cannot fit in the limited number of

menus

• Forces people to repeat themselves within a complex system

of options and waste their time listening to irrelevant options

definedcrowd confidential 6

Human to Human Interaction

definedcrowd confidential 7

How is technology evolving

IVR

Virtual Assistants

definedcrowd confidential 8

The AI revolution

definedcrowd confidential 9

AI revolution

Call centers will be enhanced by AI

Cars, public transportation will

be self-driven

A lot of shopping attendants will be replaced by robots

and bots

First stage of medical diagnosis will be done by AI

Music will be done by AI

AI will revolutionize education and

learning experience

Robots will help elderly at home and replace a lot of the

caregivers’ jobs

House tasks will be done by machines

definedcrowd confidential 10

How do machines learn

=+

definedcrowd confidential 11

Machines need structured data to learn

90% of the generated data is unstructured

definedcrowd confidential 12

Our platform enables data scientists to collect, enrich and structure high quality

training data for AI and ML applications.

definedcrowd confidential 13

AI data workflows + humans-in-the-loop + machine learning

definedcrowd confidential 14

How can we improve customer care?

Audio

Text

Metadata

16

extract meaning

definedcrowd confidential 17

Can a machine be reactive?

definedcrowd confidential 18

Automation Example

Client Database

Inbound Call

Speech Recognition

Paralinguistic Events

Text Processing

(NLP)

Generate Response

5/16/2017

definedcrowd confidential 19

5/16/2017

definedcrowd confidential 20

5/16/2017

definedcrowd confidential 21

5/16/2017

definedcrowd confidential 22

5/16/2017

definedcrowd confidential 23

5/16/2017

definedcrowd confidential 24

5/16/2017

definedcrowd confidential 25

5/16/2017

definedcrowd confidential 26

5/16/2017

definedcrowd confidential 27

5/16/2017

definedcrowd confidential 28

5/16/2017

definedcrowd confidential 29

5/16/2017

definedcrowd confidential 30

5/16/2017

definedcrowd confidential 31

5/16/2017

definedcrowd confidential 32

definedcrowd confidential 33

Market size

$36Bn

AI

$84.7Bn

Customer Care

2025

2020

5/16/2017

definedcrowd confidential 34

definedcrowd confidential 35

Watch our enterprise SaaS video: https://www.youtube.com/watch?v=vAWeT-edroc&feature=youtu.be

5/16/2017

definedcrowd confidential 36