Mysore Call Girls 8617370543 WhatsApp Number 24x7 Best Services
Sajit Joseph - The road to AI for the enterprise
1. The road to AI for
the enterprise
@sajit_joseph
2. Revenues from AI
will expand from
$8 billion in 2016
to more than $47
billion by 2020
– IDC
3.
4. Artificial Intelligence in day to day use
Facial recognition
on social media
sites. Hashtag
correction on
Twitter
Netflix’s
personalization
engine is based on
machine learning
Music providers
like Pandora and
Spotify are based
on artificial
intelligence
Banks use artificial
intelligence for
credit card fraud
detection
We use AI in our day to day lives. Leading firms have been enhancing their product
capabilities using artificial intelligence
5. What about Artificial Intelligence
for the enterprise?
What can we do now to harness the value of AI?
6. Areas for immediate focus
• Bots and virtual assistants
• Smart speakers
• Predictive analytics
• Robotic process automation
• Other areas
8. What are they?
• Bots and virtual assistants are programs which
performs an automated task or process. They have
been around for many years now.
• Uses a Natural Language Processing engine
• Combination of text/speech
• Deployed on web channels, mobile apps and
social media apps
• Goal is faster access to information or task
execution
Bots and Virtual Assistants
Voice
Chat
BOT
9. Market trends
• Chatbots will be responsible for cost savings of over $8 billion annually
by 2022, up from $20 million in 20171
• By 2019, 40% of enterprises will be actively using chatbots to facilitate
business processes using NLP2
• 27% of consumers weren’t sure if their last customer service interaction
was with a human or a chatbot 3
• By 2020, the average person will have more conversations with bots than
with their spouse2
Sources: 1 - Juniper Research 2 – Gartner 3 - PWC
Bots and Virtual Assistants
10. Where can it be used?
• Widely used for customer facing applications but huge opportunities for
employee facing use cases
• Opportunity to improve customer experience and reduce costs.
• Applications could be:
• Answering questions on a topic
• Performing transactions (scheduling etc.)
• Event based approvals
Bots and Virtual Assistants
11. Suggested approach & considerations
• Identify one use case to start with
• Clearly identify the expected business outcome
• Leverage one framework which can be used for current and future use
cases
• Focus on monitoring user behavior, unanswered questions etc. to improve
the overall performance of the capability
Bots and Virtual Assistants
13. What are they?
• Smart speakers and conversational systems
are devices which can perform interactive
functions via conversations
• Users will increasingly interact with brands
using Smart speakers like Alexa and Google
Home
• The rate of adoption of these devices is much
faster than the adoption of smart phones
Smart Speakers
14. Market trends
• By 2020, 50% of all searches will be voice searches1
• 55 percent of U.S. households by the year 2022 will use smart devices like
Amazon Echo and Google Home1
• One-in-six Americans (16%) owns a voice-activated smart speaker3
The
rate of growth will be faster than the smart phone adoption which is at
around 66% in US 2
.
Sources: 1 - Juniper Research 2 - ComScore 3 - NPR and Edison Research
Smart Speakers
15. Where can it be used?
• These devices represent a new channel like web or mobile devices.
• Its not a question of ‘if’ but ‘when’ users will be able to conduct most
interactions with a brand via these devices
• Create conversational capabilities to allow users to ask questions, conduct
transactions.
Smart Speakers
16. Suggested approach & considerations
• Designing a conversational experience is not the same as designing a
user experience for web or mobile channels
• Define an experience which limits the amount of conversations needed to
get a task done
• Select a set of use cases which have the most business value and is easy
to design
• Ensure that the experience is personalized and contextualized for the user
• Deploy – Test – Iterate
Smart Speakers
18. What is it?
• Predictive analytics is using historic data to forecast
future outcomes
• Advanced analytics and machine learning can be
used to provide insights
• Example – Automated lead scoring of sales
opportunities can help a sales team to focus on
opportunities which have the highest likelihood of
closing
• Software providers like Salesforce are beginning to
add these capabilities in their products
• Most use cases with high ROI would require custom
model development
Predictive Analytics
Predictive
Insights
19. Market trends
• Using a machine learning algorithm, Amazon automated picking and
packing items in a warehouse logistics setting. Amazon’s average time
from ‘click to ship’ now stands at 15 minutes
• Netflix saved $1 billion in a year as a result of its machine learning
algorithm which recommends personalized TV shows and movies to
subscribers (by avoiding cancelled subscriptions).
Source: McKinsey quarterly
Predictive Analytics
20. Where can it be used?
• The applications of this concept are huge.
• Examples - demand prediction, churn management, fraud detection, sales
forecasting etc.
• The accuracy of the predictive model is based on the availability of data for
analysis.
Predictive Analytics
21. Suggested approach & considerations
• Start with focusing on a use case which has high ROI
• Make sure you have access to all the data elements which are linked to
the outcome
• In many cases access to external data might be needed to enrich the data
with different data points (e.g. weather data)
• Leverage professionals who have statistical modeling and machine
learning expertise to identify patterns and create forecasting models
Predictive Analytics
23. What is it?
• Robotic process automation focusses on using
tools to automate a manual process.
• RPA tools provide a wide range of options
• Few advanced RPA tools have started integrating
artificial intelligence to provide cognitive automation
• Cognitive automation could range from reading
handwriting or understanding scanned documents
to improved results by unsupervised learning
Robotic Process Automation
24. Market trends
• RPA offers a potential ROI of 30–200 percent—in the first year1
• RPA is expected to see a compounded annual growth rate of about 60.5
percent worldwide through 20202
• By 2020, automation and artificial intelligence will reduce employee
requirements in business shared-service centers by 65%3
Sources: 1 - McKinsey 2 - Transparency Market Research 3 - Gartner
Robotic Process Automation
25. Where can it be used?
• Traditionally used for back office functions but it can be used in most
places with manual processes using structured data
• Most common use cases tend to be in procure to pay, order to cash,
accounts receivables and reconciliation
• RPA can execute a use case in weeks/months but in some cases it might
be best to use RPA as a stop gap till an IT system is put in place to
address the root cause of the manual process
Robotic Process Automation
26. Suggested approach & considerations
• Identify the top list of target use cases
• Identify an RPA product which works best for these target use cases
• Start piloting the product using one use case and then work on automating
other use cases as per their priority
• Keep change management in mind while rolling out these initiatives
Robotic Process Automation
27. Other areas
Computer Vision • We are in the early stages of mainstream application of this technology.
• There are significant implications for various industries specially Retail and
Consumer Goods organizations.
• Amazon Go is a great example of how this technology has been used
Prescriptive Analytics • Next stage of evolution after Predictive Analytics
• Focusses on ‘What should be done’? Using rule definition, decision trees
and optimization models
Robots • Robots have been around for many years.
• AI has given new life to the possibilities of how robots could be used
• Soon robots will be used to augment and automate business processes
(e.g. Amazon Kiva)
Other Areas