Conversational AI€¦ · …delivering clear business value… 6 1. Increase sales across digital commerce channels with personalized, 24x7 human-like “bots” 2. Lower costs of

© 2018 Cognizant, confidential and not for distribution © 2018 Cognizant

Conversational AI Where are we now, where are we heading and what should you do tomorrow?

Bart van der Mark, AI Europe

24 April 2018

© 2018 Cognizant, confidential and not for distribution 2

Bart van der Mark

Head of AI Venture Europe

[email protected]

Runs the Cognizant Conversational AI Venture in Europe.

This venture identifies, nurtures and grows promising

concepts, technologies and business models around

Robotic Process Automation, Cognitive Agents, Artificial

Intelligence and Machine Learning. He is a frequent

publisher and key-note speaker on these topics.


Where are we?

© 2018 Cognizant, confidential and not for distribution

Chatbots have evolved from dumb to fairly intelligent…

4

• 100% scripted

• Stand-alone

• System-driven

• Narrow scope

• Dumb

• Scripted + AI

• Integrated

• User-driven

• Narrow scope

• More intelligent


…quite a few serious examples exist…

5

Created a ‘bot designed to

be a ‘Personal Stylist in

your pocket’.

Integrated Amazon Alexa in its

newer cars, allowing drivers to do

nifty things like check tire

pressure, maintenance

requirements, gas, etc. right from

their homes.

Unveiled ‘My

Starbucks Barista’—

AI-based app for

mobile orders.

Launched an AI based ‘bot called

‘Luvo’, to help customers with

responses to financial services

questions.

Launched a Facebook

Messenger chatbot for ordering

pizzas.

Use a successful

chatbot for fuel- and

oil-related customer

queries

Enable clients to find

and book tickets

through natural

conversation


…delivering clear business value…

6

1. Increase sales across digital

commerce channels with personalized,

24x7 human-like “bots”

2. Lower costs of customer service, and

improve customer satisfaction and

loyalty

3. Increase employee productivity and

satisfaction by automating high

frequency and routine service desk

interactions


…research show a promising future (and some areas of attention)…

7

Source: LivePerson



8

Source: LivePerson



9

Source: LivePerson


… and this is what we see reflected in our day-to-day business:

10

1. Organisations see / expect value but are uncertain what to do and where to start

2. Unsurpassed demand from organisations to learn more

3. High demand for PoC’s and pilots

4. Moderate demand for production-ready bots

5. Still somewhat limited but growing demand for usage at scale

6. Initiatives are still too often stand-alone

7. Fast evolving technology landscape, but not perfect yet


Where are we

going?


All AI, including Conversational, follows a exponential curve

12


Exponential growth of chatbots is fuelled by 3 key elements

13


Characteristics for future bots at scale

14

• Primarily AI driven

• Integrated with front- and back-end

and other AI

• User-driven

• Broader scope

• Intelligent but by far not at human

level


Key recommendations!


Put the customer and business outcomes in the centre,

not tools & technology…

16


…position chatbots in the context of the overall engagement strategy…

17


…and ensue that your end-to-end landscape supports

18


And remember: it’s all about bots and humans working together

19


Ensure you have a chatbot plan and roadmap beyond pilots / PoC’s

1. Discover

&

Experiment

2. Plan &

Lay

Foundation

4. Maintain & Expand

3a. Build CoE & In-

house Skills

External In-house

• Learn

• Quick scan assessments

• PoC’s

• …

• Technology choice / reconfirmation

• Roadmap / business case

• C-level commitment

• Governance & Compliance

• Methodologies

• Partners

• Infrastructure

• …

3. Implement &

Scale

Typically our clients would be enabled to:

• Manage demand and priorities

• Execute low and medium complexity

changes

• L1 and potentially L2 support

• Build low complexity chatbots

• …

Build and implement the chatbot

based on the agreed:

• Roadmap

• Technology & Infrastructure

• Methodologies

• …

20


Start NOW and start with NARROW use cases

21


Don’t get lost in the tool-jungle; get help

22


For this we built a global Conversational AI capability ahead of the curve

23


In

Summary

1. Chatbots deliver value today and

will develop exponentially

2. It’s not about replacing humans

but working together

3. Focus on business outcomes, not

tools

4. Think end-to-end

5. Start small but think big =>

roadmap

6. Do a great job on narrow use

cases and not vice versa

7. Do not re-invent the wheel; get

help

Conversational AI€¦ · …delivering clear business value… 6 1. Increase sales across digital commerce channels with personalized, 24x7 human-like “bots” 2. Lower costs of

Documents