Describing the Problem Statement / Business Need :-
•
IT
Production Support Monitoring and Fixing is a task which runs without any
interruption.
•
The
concept of Onshore/Offshore partnership is currently prevalent in our IT
industry to take care of this round-the-clock task to address any production
issue which can happen at real time.
•
Humans
are employed to ensure the production Jobs/Processes/Programs are running
without any issue. In case of issue they resolve
•
Humans
are expensive. They can make errors. We are not machines and naturally feel
tired, or can fall sick.
•
Software
Robots do not have the disadvantages mentioned in the previous line.
First we need
to think, “Goals” are defined by whom and why. It is people like us, who need to understand from the past statistics,
and accordingly generate a necessity of a Goal to be defined. Our objective
should not be limited to think within the already set Goals, but also to think
beyond (if possible). And that is exactly what we are trying to propose out
here.
As Production Support needs to be given 24/7, 365 days a year, so we can reduce
human intervention with software robots for Production Support Monitoring, and
also train the robots accordingly to address error fixing for those errors
which are of generic nature.
Additionally,
I propose to implement Machine Learning
Algorithm within our proposed software robot. As a result the robot will
itself predict the repetitive errors and suggest/execute/trigger the required
fixing steps.
Points to be
adhered to, for a prototype:-
Among the different types of Machine Learning tasks, a crucial distinction is drawn between
supervised and unsupervised learning:
•
Supervised
machine learning: The
program is “trained” on a pre-defined set of “training examples”, which then
facilitate its ability to reach an accurate conclusion when given new data.
•
Unsupervised
machine learning: The
program is given a bunch of data and must find patterns and relationships
therein.
With the implementation of Pattern-Recognition using Machine-Learning-Algorithm
within our proposed software robot, the robot can predict/anticipate probable
production issues.
§ The robot will itself predict the
not-so-frequent repetitive errors (through an Auto Learning Circuit, which will
be responsible for monitoring the Drives/Servers and the respective Websites
using the site maps) and execute/trigger the required fixing steps, when the
errors actually happen.
§ In such a situation, we can gradually
remove the dependency on Humans and Manual-Production-Fix-Documents.
Ø
We
will develop automated tool modules for respective error fixing steps,
which will be controlled by our primary software robot.
Ø
We
will refer to the previous year’s error logs, which will help us as statistics for the known Use-Cases (as
per the Manual-Production-Fix-Documents), which needs to be given for the
already known repetitive production errors.
BENEFITS (Tangible/Intangible):-
•
Software
robots can lower the costs by using automated tools which can work 24/7, 365
days a year, picking up an employee’s repetitive tasks.
•
Usually,
one software robot can replace between two to five humans.
•
A
software robot is at least three times faster than a human and less prone to
committing errors, and thereby more efficient.
•
An
example of one Robotic-Process-Automation process is where we take data from
one application and match the data against specific rules from another
application.
•
Another
area of impact is IT infrastructure outsourcing. Here, software robots can keep
tabs on critical thresholds -- such as storage capacity -- and take corrective
action.
•
Software
robots can monitor the thresholds and, under certain threshold breach
conditions, robots can even go and address the fix if [they are] configured to
do so.
•
Processes
ripe for Robotic-Process-Automation include those that are rules-based,
definable, repetitive, high volume or span multiple systems.
Comments
Post a Comment