Chatbots have evolved significantly since the early programs like ELIZA in 1966. Notable developments include IBM's Watson in 2011, Siri in 2010, Google Now in 2012, Alexa in 2015, and Cortana in 2015. These chatbots used natural language processing to have conversations, answer questions by accessing online data, and perform tasks for users. In 2016, Facebook launched bots for Messenger and Microsoft created Tay, although Tay had to be shut down after it rapidly developed harmful speech.
5. What Robotic process automation (RPA)
5
● Robotic automation refers to process automation where software drives existing enterprise application
software in the same way that a users does
● Robotic automation is a tool or platform that operates other application software though the existing
application’s users interface
● Robotic process automation (RPA) is defined by the Institute for Robotic Process Automation (IRPA) as ‘the
application of technology allowing employees in a company to configure computer software or a ‘robot’ to
capture and interpret existing applications for processing a transaction, manipulating data, triggering
responses and communicating with other digital systems.’
6. What Robotic process automation (RPA)
6
● What is automation ? Sequence of steps performed without human interaction
● What is process ? sequence of activities (meaningful)
● What is robotic : mimcking a human
=> RPA : Automate the process by mimcking a human
7. Types of RPA
7
Back
office
robot
❖ Run unattended
❖ In a batch mode
❖ In virtual environment
❖ Send heartbeats to the server, so that
knows instantly when a robot is down
Front
office
robot
❖ Share the same workstation with an
employee who has control over when it
is used
8. Choosing the RPA Types
8
Back
office
robot
❖ Work in an unattended manner
independent of any human intervention
❖ Use in manual, repetitive high rule
based back office activities not
requiring any human intervention
❖ Can triggered both from local machine
or directly from server
Front
office
robot
❖ Work side by side with human agents
and assists them to automatically
complete the process
❖ General use in manual, repetitive,
highly rule-based activities containing
decision points that requires human
intervention
❖ Communication with server is
bidirectional :
➢ Robot to server : execution logs,
automated process upload
➢ Server to robot : automated process
version deployment only
❖ Can triggered only from the local
machine (either from triggers)
generated by humans in other
applications via a scheduled tasks in
windows task scheduler or manually
9. Choosing the RPA Types : requirement
9
Back
office
robot
❖ Can be either based on structured data
or unstructured data. Structured data
operations typically include data
operations on various databases.
Unstructured data operations typically
include processing of documents,
images and multi-modal data.
Front
office
robot
❖ Involves chatbots or conversational
agents.
❖ usually require natural language
processing or natural language
understanding
10. 10
The Benefits of RPA
• Like chatbots, RPA can be a powerful tool for digital transformation by providing the following
benefits:
• Reduce labor & operational costs
• Increase processing speed
• Improve customer experiences
• Extend service hours.
• Reduce human errors
• Provide comprehensive operational visibility
• Increase security & compliance
11. 11
The Limitations of RPA
• Despite the benefits of RPA and similar solutions for the enterprise, automation is not always the
right choice. Common disadvantages or weakness of RPA include:
• Employee support/morale: Fear of layoffs and redundancy can hinder operations and increase employee and institutional
knowledge attrition.
• Complex nonlinear processes: Automating processes that are too variable, unpredictable, or infrequent can raise the
complexity and cost of associated robot algorithms and subsequent deployments. Human touch points may still be required,
but not all products offer assisted or semi-automated processes.
• Error creep: Since most RPAs automatically execute the instructions found within scripts or business objects, errors can
replicate hundreds or thousands of times before the problem is spotted and corrected.
• Management challenges: While some RPA solutions provide dashboards for supervising robot queues and workflows,
managing thousands of bots executing transactions across an enterprise presents a challenge. Especially since it is common to
have many robot instances executing the same business processes in different locations.
• Hidden costs: Some RPA solutions require additional Virtual Machines (VM) or Virtual Desktop Infrastructures (VDI), which
raises the virtualization/OS cost for each new bot or process after the initial go-live. Scaling may cost more than originally
anticipated.
• Additional governance: Most RPA software bots, or at least those that take a script-based approach to automation, are
designed to never deviate from established algorithms. Any changes to an internal or external aspect of an existing business
process are thus likely to cause the software bot to fail to perform the intended function. This often necessitates additional
centralization, standardization, and optimization of processes prior to deployment or shortly after
13. Which Processes for RPA?
13
The general characteristics of a process which is ready for robotic process automation would be:
• repetitive and rules based
• accesses structured data sets
• utilises applications on a Windows or Web based platform
• the process is documented and has been standardised in practice
• three or more staff are hired to complete the process
• data input is prone to human error
14. Use cases for RPA?
14
Finance
&
Banking
● New account
verification
● Data validation
● Customer account
management
● Financial claims
processing
● Report creation
Form filling
● Change of address
● Loan application
processing
Insur
ance
● Claims processing
● New account
creation
Telec
omm
unica
tions
● Collection &
consolidate
customer data from
client phone
systems
● Backing up
information from
client systems
● Uploading data
● Transferring
customer data
between
applications
● Extracting data
about competitor
pricing
Healt
hcare
● Patient data
migration and
processing
● Reporting for
doctors
● Medical bill
processing
● Patient record
storage
15. Use cases for RPA?
15
Retail
● Automatically
updating online
inventory &
product
information
● Importing website
orders and email
sales into back-
end systems
IT
● Creating new
accounts
● Software
installations and
updates
● Batch processing
● Printer set-ups
Finan
ce/A
ccou
nting
● Invoice processing
● Accounts payable
and accounts
receivables
● Reporting
● Bank reconciliation
● Fixed assets
analysis
● Master data
management
● Vendor and
customer account
creation
● ERP logging from
another system
HR
● Employee on-
boarding
● Leave of absence
management
● Populating
employee data
into multiple
systems
● Performance
appraisal
management
17. Understanding the Spectrum of RPA technology
17
● Robotic Process Automation is a class of software
that replicates the actions of humans operating
computer systems in order to run business
processes. Because the software ‘robots’ mimic
exactly what the human operators do
○ Example : by logging into a system, entering
data, clicking on ‘OK’, copying and pasting
data between systems, etc.
the robots will do exactly what you have trained them to do
each time, every time. This is both their greatest strength
and their greatest weakness.
➔ strength because you need to be sure that the robot
will carry out the process compliantly and accurately
➔ weakness because it precludes any self-learning
capability.
● This inability to self-learn leads to two distinct constraints for RPA,
○ The first is that the robots require structured data as their input, whether this be from a
spreadsheet, a database, a webform or an API. => AI, NLP, NLU
○ The second constraint for RPA is that it can’t make complex decisions, i.e. it can’t use
judgement in a process. => cognitive reasoning is useful
18. Understanding the Spectrum of RPA technology
18
The key dimensions defining the software vendors are :
1. Data: The level of sophistication in dealing with business data (structured or unstructured);
2. Type of Tasks predominately performed: tasks are either rules based or require
knowledge from multiple sources to complete the process.
3. Interoperability: working across multiple applications (single application or multiple
applications and platforms)
4. AI: The level of artificial intelligence provided by the application (none, machine learning
based on pattern recognition and statistics, or emerging true AI).
21. RPA tools selection
21
Tech
nolo
gy
• First and foremost is the
technology in which RPA
tool has been built, like
Microsift .NET, IBM
Mainframe, Java, web etc.
The tool has to be platform
independent, considering
many of the organizations
perform their day to day
tasks outside the local
desktop using Citrix or
Virtual machines etc. So
the automation solution
should be able to support
any application and
platform.
Inte
rface
• Next most important
feature is the RPA tool’s
user interface. A complex
user interface will delay
the process of
implementation and
increase the learning
curve and adaptability. A
more user friendly
interface like drag and
drop, auto capture, image
recognition etc. would be
better.
Man
age
men
t
• It is important to know
how effectively and easily
the robots can be
managed considering
these robots are going to
be replacing your human
workforce. There should
be a high level of visibility
and control in terms of
process monitoring,
process change,
development, re-use etc.
Secu
rity
• Security is one of the most
important factor to be
considered. How safe are the
robots when compared to the
humans? After all robots/RPA
tools are also a piece of
software, so how much more
security controls can be
implemented, test the code
rigidly and lock it up, the better.
Automation solutions that
supports the compliance
process (like HIPPA – For
Healthcare industry, SOX –
Financial Sector, PCI DSS –
Credit card related organization
etc.) are some of the factors to
be considered in terms of
security.
23. 23
Robotic Process Automation - How it Works
• RPA software products are typically composed of three fundamental elements:
• Developer Tools Most RPA products provide various tools that are used to model and define bot assignments or jobs at either
the individual user level (via work desktop/ laptop) or enterprise level (via centralized dashboard). These assignments or jobs,
which are recorded manually or automatically, consist of detailed step-by-step instructions that the bot must follow in order to
successfully replicate the business process being modelled. The detailed instructions are composed of rules or conditional
logic such as if/then or if/then/else statements and are often encapsulated within scripts, business objects, or other containers.
Most developer tools for RPA provide visual decision tree editors, configuration wizards, and drag-and-drop controls.
• Robot Controller The robot controller acts as the eyes and ears of your software robot by playing three roles. The controller
serves as a repository for all bot assignments by storing performance data, history, and bot credentials. It provides built-in
controls and workflows to help govern the creation, testing, approval, and deployment of new assignments and provides a
versioning system to track ongoing configuration changes. The robot controller can also monitor and report on the status of bot
activities and transactions to the operations manager or other designated employees.
• Software Robots Software robots interact directly with enterprise back-end systems and applications to process transactions
and other tasks. Most RPA products provide at least some pre-built bot functionality with the ability to custom-code additional
capabilities. Software robots are also typically designed to automatically identify the important elements of a business systems
user interface, such as buttons and input fields, by examining the application code in order to simplify the deployment process.
For virtualized desktops, however, software robots must sometimes rely on less-effective techniques such as searching for
pixel-based object locations.
26. Best practices for implementing RPA
26
That consideration leads to the obvious question, “So just exactly how do you implement RPA correctly?” :
1. Access applications and data using the most efficient, accurate, scalable way available. The
desktop user interface isn’t the best, but may be the only option.
2. Make sure the IT department is involved and that you’re on the change control list. — Caution: If
you’re using RPA with an application that sees a lot of UI changes, this could hamper IT and delay
application improvements requested by other departments; and that will not make you very popular
with the folks in IT. Remember: RPA tools rely on a consistent user interface. If your IT department
changes the user interface to improve usability, RPA robots may stop working.
3. Centralize the logic and runtime execution so that you have better control. — A free RPA tool could
proliferate rogue departments’ automating applications without the knowledge of their IT
department. This could cause major issues and cost the organization a lot of money. It gets back to
having a good automation strategy and, possibly, a center of excellence.
4. Provide full auditing and reporting capabilities.
5. To avoid issues with changing UIs, use APIs, Web services, and database calls as much as
possible.
6. Use RPA when it’s the only option available, and avoid the desktop as much as possible.
29. 29
What are chatbots
• Unlike RPA, chatbots are intermediaries between systems that you can talk to. They communicate
with applications, things, and people to take action on revenue generating activities immediately.
• A chatbot is basically a program that’s designed to talk to you and collect information from your
conversation. Depending on how it’s developed or how intelligently it’s been built, a bot can use that
information to do things for you, such as :
• book you a flight,
• suggest personalized offers or promotions,
• pass you along to a human in times of frustration, based on how you’d like the bot to help.
• and much more.
31. 31
History
• 1966, Eliza: Imitated human conversation by matching user prompts to script responses - he was able, at least for a time, to pass the Turing AI test
• 1972, PARRY: Inexplicably simulated a person with paranoid schizophrenia. PARRY was more serious and advanced than ELIZA - he was described as
"ELIZA with attitude"
• 1988, ABBERWACKY: One of the first attempts to create AI through human interaction. Mainly a form of entertainment, its ultimate goal was to move
from a text-based system to an entirely vocal system
• 1992, DR. SBAITSO: An AI speech synthesis program created for MS DOS computers. Designed to highlight a digitized voice, the role of a psychologist
in interacting with users
• 1995, A.L.I.C.E: "Artificial Linguistic Internet Computer Entity". A.L.I.C.E was a natural language processing robot. It could apply rules of matching
heuristic models to human input - in other words, having a conversation
• 2001, SMARTERCHILD: An intelligent robot widely distributed on SMS networks. With features such as fast data access and a personalized fun
conversation, it was considered a forerunner of Siri Apple and Samsung S Voice.
• 2006, IBM's Watson: Watson was specifically designed to complement in Jeopardy !, which he won in 2011 against two of the former champions of the
show. Now, watson uses natural language processing and automatic learning to reveal knowledge from large amounts of data.
• 2010, SIRI: An intelligent (and cheeky) personal assistant, part of the Apple IOS, which has a natural language user interface to answer questions and
perform Web service requests. Siri would pave the way for all subsequent AI Bots and APs.
• 2012 GOOGLE NOW: Developed by Google for the Google Search mobile app, it uses a natural language user interface to answer questions, make
recommendations, and perform actions by delegating requests to a set of Web services
• 2015, ALEXA: A voice that inhabits the amazon echo device. Alexa is capable of voice interaction - she uses natural language processing algorithms to
receive, recognize and respond to voice
• 2015, CORTANA: A smart personal assistant created by microsoft, CORTANA sets reminders, recognizes natural voice commands and answers
questions using the Bing search engine
• 2016, BOTS FOR MESSENGER: In April, Facebook launched a Messenger platform allows developers to create robots can interact with FB users. In
June, 11,000 bots were available
• 2016, TAY: Designed to mimic the speech and habits of an American teenager, Microsoft Tay rapidlt's chatbot developed a vicious paranoia that made
the HAL 9000 softer. It was closed after only 16 hours of existence, and has not been heard from since
32. 32
The Benefits of Chatbots
• AI-rich chatbots facilitate intelligent dialogue between people and systems and provide the following
benefits:
• Enhance digital: Create higher-value purchases and service interactions for your customers. Chatbots can access more
customers and open new sales outlets.
• Reduce app fatigue: Sync brands and systems with the communication channels customers already use. Help customers by
cutting down on the number of apps and interfaces that must be mastered.
• Unlock efficiencies: Minimize manual, resource-intensive business and supply chain processes.
• Simplify complex interfaces: Form-intensive interfaces can be broken down into easily digestible, natural language driven
conversations via popular communication channels such as websites, email, SMS, messaging platforms, and more.
• Streamline work: Let employees get routine tasks done in less time, with communication tools they know and love - enhancing
productivity and efficiency. Allows for better resource allocation and decreases calls and emails to support services.
• Reduce communication overload: Decrease firstline support calls and emails. Allows for better resource allocation, faster
support, and better user experiences. Chatbots provide 24/7 customer support and immediate answers.
• Provide comprehensive visibility: Analytics generated from chatbots can provide unmatched visibility into customer and
employee behavior.
• Retain human feel: Chatbots deliver personalized, memorable conversations and experiences through natural language while
helping to connect people and systems. They can sound and act like humans, which leads to higher customer and employee
engagement when compared to static forms and information sources. Plus, human agents are usually only a click away,
ensuring customers always get the help they need.
33. 33
The Limitations of Chatbots
• While the benefits of chatbots are vast, there are some disadvantages associated and common
mistakes made. Common weaknesses of chatbots include:
• Use cases are endless: Identifying valuable use cases that solve a real problem can take a lot of effort, and there’s
typically more than one way to get there. Partnerships may be important for developing high-value use cases.
• Not all bots are created equal: Developers can quickly build bots, with or without experience, effective tools and
frameworks are key to creating successful chatbots.
• Narrowness: Many chatbots, like software robots, are designed to perform a few similar tasks or a single exact
tasks. This can quickly lead to either complex bots or a plethora of bots. Universal bots that can coordinate the
work of multiple other bots can solve this issue.
• Security & compliance: Personally identifiable information and business data may be at risk unless the chatbot is
designed and implemented with the right level of encryption, authentication and auditability, which is only in true
enterprise-grade secure development platforms.
35. Chatbots - Key vendors
35
From the inside out, the concentric circles represent:
● Platforms: The messaging platforms that enable
the existence bots through robust send-and-
receive APIs, frameworks and ecosystems.
● Brands: Companies which have launched and
experimented with bots, in that particular
quadrant (for example, Managed x Support).
● Providers: Companies who have the capabilities
to deliver exceptional work in that quadrant.
● Tools: The supporting tools used by providers,
brands, or developers of bots in delivering bot
experiences.
36. 36
Vision du marché
Api.ai Wit.ai Luis Lex Watson Rasa Chatscript Do you
dream up
Editeur Google Facebook Microsoft Amazon IBM Rasa Chatscript Do You
Dream up
Open
Source
French
langage
Integration
Pricing
On-premise
38. 38
How it Works
• Chatbots are typically composed of (4) fundamental elements that serve to provide rich
conversational interfaces for customers and employees:
• Tasks
• Channels
• NLP & Speech
• Intelligence
• Advanced chatbot platforms also provide the following fundamental elements:
• Bot & Dialog Development Tools
• Platform Middleware
• Enterprise Capabilities (advanced encryption, administration, and compliance functionality)
40. 40
How it works
New Process With Chatbots- High-Level View
API / DataBase
Horus, Gédéon, Céphée, ...
User sent request
1
The chatbot sends the
request to the
conversational engine
2
Conversational engine
The engine identifies the intent
and the context3
The chatbot will search for the
appropriate service or data4Data or service is sent to
the chatbot
5
The chatbot
package the data
in an appropriate
response
6
Cephee
0
Natural language engine
Chatbot
SI
42. 42
Process Automation
RPA
● Used to fully automate workflows. Can
partially automate processes (humans
initiate or approve actions)
● Best suited for processes that are
perfectly understood and/or
documented. Little-to-no room for
deviation
Chatb
ot
● Used to simplify manual processes
● Used to more effectively handle
processes that have some automation
and some human elements
● Can be used to automate workflows
● Can drive smarter human decisions and
proactively change human behavior
● Can connect, automate, or enable a
more effective way to respond for
processes that are ill-defined or
documented, that are uncommon, or
that involves deviation and/or human
intervention
43. 43
Human Involvement
RPA
● Primarily limited to initial
configuration/ scripting/
implementation and error/ exception
handling
Chatb
ot
● Humans engaging directly with systems
44. 44
System Interaction..
RPA
● Primarily limited to initial
configuration/ scripting/
implementation and error/ exception
handling
Chatb
ot
● Replaces existing (often complex) GUIs
with conversational interfaces across
channels
45. 45
Intelligence
RPA
● Primarily limited to initial
configuration/ scripting/
implementation and error/ exception
handling
Chatb
ot
● Can add intelligence to a process via
natural language. Has the ability to
understand, remember and learn from
the information gathered during each
interaction