1. The document discusses how robotic process automation (RPA) is being used by Vueling Airlines to automate repetitive tasks in their Continuing Airworthiness Management Organization (CAMO) department.
2. RPA software robots mimic human actions to perform tasks like downloading and moving files without significantly changing existing IT systems.
3. Vueling worked with EXSYN Aviation Solutions to identify rule-based processes for automation, develop a scoring model, and program an RPA bot called gAVInBOT MK1 to automate the creation and printing of aircraft maintenance work packages from their AMOS system.
4. Automating this process with RPA reduced the workload for 3-4 employees by 2-
WSO2's API Vision: Unifying Control, Empowering Developers
Automating Humans back into Aviation
1. Automating Humans Back into
Aviation –
How Robotic Process Automation
is used on a daily basis within the
CAMO department of Vueling
Albert Almendro: Aircraft Structures
Engineer & AMOS
Administrator
Sander de Bree: CEO
gAVInBOT MK1: Software Robot v1
2. TABLE OF CONTENT
I. Meet gAVInBOT
II. RPA
III. Approach
IV. Scoring Model
V. Results
VI. Case Study
VII. To Summarize
VIII. Future Skills
IX. Questions
4. gAVInBOT MK1
EXSYN Aviation Solutions & Vueling Airlines
1st of May 2017
Amsterdam, the Netherlands
0 0
0
I. Meet gAVInBOT
II. RPA
III. Approach
IV. Scoring Model
V. Results
VI. Case Study
VII. To Summarize
VIII. Future Skills
IX. Questions
6. Why is it interesting to M&E / CAMO?
Uses Robotic Process Automation techniques to:
Copy human actions and perform them in the
same way.
Download files, move files, enter data in a system
or file.
In order to
• Facilitate cooperation between humans and robots.
I. Meet gAVInBOT
II. RPA
III. Approach
IV. Scoring Model
V. Results
VI. Case Study
VII. To Summarize
VIII. Future Skills
IX. Questions
7. Why is it interesting to M&E / CAMO?
Low intrusion: RPA performs the same actions a human
employee would, and therefore does not require
high impact changes in the IT infrastructure, both
hard- and software
Rule-based: Applicability is centered around rule based
process steps
Key to preventing skilled workforce to be mainly involved in
low-skilled repetitive tasks
I. Meet gAVInBOT
II. RPA
III. Approach
IV. Scoring Model
V. Results
VI. Case Study
VII. To Summarize
VIII. Future Skills
IX. Questions
8. How does it look like?
I. Meet gAVInBOT
II. RPA
III. Approach
IV. Scoring Model
V. Results
VI. Case Study
VII. To Summarize
VIII. Future Skills
IX. Questions
9. Benefits vs. Challenges
BENEFITS CHALLENGES
• Low intrusion
• 24/7 workforce
• Workload control
• Insight in process performance
• Removing the boring stuff
• Error reduction or prevention
• Finding the correct automation
cases
• Managing changes
• Exceptions
• Automating for automating
• Automating poorly designed
processes
• Human employee resistance
I. Meet AviBOT
II. RPA
III. Approach
IV. Scoring Model
V. Results
VI. Case Study
VII. To Summarize
VIII. Future Skills
IX. Questions
11. How to overcome the challenges in 5 steps:
Gather Analyze Visualize Score
Skill
learning
Map the different area’s in the organization and their processes
Output: Process landscape
I. Meet gAVInBOT
II. RPA
III. Approach
IV. Scoring Model
V. Results
VI. Case Study
VII. To Summarize
VIII. Future Skills
IX. Questions
12. How to overcome the challenges in 5 steps:
Gather Analyze Visualize Score
Skill
learning
Gather process information on:
• Volume
• Time investment
• Rule-based
• Exceptions
• Human errors
• Process changes
• Input/output
• System & application
• Risks
I. Meet gAVInBOT
II. RPA
III. Approach
IV. Scoring Model
V. Results
VI. Case Study
VII. To Summarize
VIII. Future Skills
IX. Questions
13. How to overcome the challenges in 5 steps:
Gather Analyze Visualize Score
Skill
learning
Establish flowchart documents to visualize the process
Output: flowcharts
I. Meet gAVInBOT
II. RPA
III. Approach
IV. Scoring Model
V. Results
VI. Case Study
VII. To Summarize
VIII. Future Skills
IX. Questions
14. How to overcome the challenges in 5 steps:
Gather Analyze Visualize Score
Skill
learning
Rate the processes based on pre-defined factors for RPA.
Using the scoring model of EXSYN
Output: Listing of process candidates for RPA
I. Meet gAVInBOT
II. RPA
III. Approach
IV. Scoring Model
V. Results
VI. Case Study
VII. To Summarize
VIII. Future Skills
IX. Questions
15. How to overcome the challenges in 5 steps:
Gather Analyze Visualize Score
Skill
learning
Educate (code) the RPA bot to perform these processes
Output: Robotized Process
I. Meet gAVInBOT
II. RPA
III. Approach
IV. Scoring Model
V. Results
VI. Case Study
VII. To Summarize
VIII. Future Skills
IX. Questions
17. The technical rating will indicate the technical feasibility of a process. Performed by
the RPA expert.
* Weights depend on RPA technological changes
TECHNICAL ASSESSMENT WEIGHT SCORE
Clearly defined rules 5
Non complex process 3
Predictable process 5
Structured input data 4
Application accessibility 2
WEIGHTED SCORE 0
Technical Score
I. Meet gAVInBOT
II. RPA
III. Approach
IV. Scoring Model
V. Results
VI. Case Study
VII. To Summarize
VIII. Future Skills
IX. Questions
18. The value rating will indicate a ‘ROI’ of the process, or how much value is expected to
be gained from automation. Done by the process expert under guidance of
automation expert.
* Weights are client dependent
VALUE ASSESSMENT WEIGHT SCORE
Time spend 5
Efficiency 4
Value of the process 4
Process is not going to change 3
Workload distribution & predictability 1
Prone to human errors 3
Regulated process 2
WEIGHTED SCORE 0
Value Score
I. Meet gAVInBOT
II. RPA
III. Approach
IV. Scoring Model
V. Results
VI. Case Study
VII. To Summarize
VIII. Future Skills
IX. Questions
20. PROCESS NAME Tech.
Score
Value
Score
Combined
Score
AD preparation 70 93 62,3
SB reporting 79 78 59,0
Damage mapping for fan blades 62 90 53,4
Workpackage creation 86 83 68,3
TSDF import 75 86 61,7
Scoring Example
I. Meet gAVInBOT
II. RPA
III. Approach
IV. Scoring Model
V. Results
VI. Case Study
VII. To Summarize
VIII. Future Skills
IX. Questions
22. About Vueling
• Low cost Airline founded in 2004
• Part of IAG group together with British Airways, Iberia,
Aer Lingus and Level.
• Over 150 destinations
• Hubs in Barcelona and Rome
• 115 aircrafts fleet consisting of A319, A320, A321 and
A320 NEO
• Largest airline in Spain by fleet size and number of
destinations
23. Business Problem
Printing of workpackages from AMOS frontend. These WPs will
go to the MRO at the end of the day. Every single item in the
maintenance forecast is one execution of the process. Daily
creation of workpackages weighs heavy on the daily workload
Current situation
• 3-4 employees.
• Spending each 2-3 hours printing WPs.
• 120 WP on average per day.
• Planning 7 days ahead.
• Skipping non-mandatory forecast items.
I. Meet gAVInBOT
II. RPA
III. Approach
IV. Scoring Model
V. Results
VI. Case Study
VII. To Summarize
VIII. Future Skills
IX. Questions
24. Current Business Process
I. Meet gAVInBOT
II. RPA
III. Approach
IV. Scoring Model
V. Results
VI. Case Study
VII. To Summarize
VIII. Future Skills
IX. Questions
25. Approach to Automation Solutions
1. Initial process analysis with planning department process expert
2. Scoring the process
3. Development of automation skeleton
4. Iterating improvements for business rules and reliability changes
5. Testing the automation in a test environment
6. Creating documentation for the end-product
Challenges
• Vueling upgraded from AMOS v10.90 to v11.30
• AMOS is not controllable through UI objects
I. Meet gAVInBOT
II. RPA
III. Approach
IV. Scoring Model
V. Results
VI. Case Study
VII. To Summarize
VIII. Future Skills
IX. Questions
26. I. Meet gAVInBOT
II. RPA
III. Approach
IV. Scoring Model
V. Results
VI. Case Study
VII. To Summarize
VIII. Future Skills
IX. Questions
27. New Business Process
I. Meet gAVInBOT
II. RPA
III. Approach
IV. Scoring Model
V. Results
VI. Case Study
VII. To Summarize
VIII. Future Skills
IX. Questions
28. • 4500,- euros worth of aviation engineer man hours saved within 6 weeks.
• 1 employee for monitoring
• Automation has enabled 15 days of buffer
• No more human error causing items to be missed
BEFORE AFTER
Employees involved 3 1
Man-hours spend per day 8 0,5
Workpackages per day 120 150
Days planes ahead 7 15
Duration of one cycle (min) 2 1
Experience of stress High Low
Daily overtime Yes No
Overall motivation employees Low High
After Automation
I. Meet gAVInBOT
II. RPA
III. Approach
IV. Scoring Model
V. Results
VI. Case Study
VII. To Summarize
VIII. Future Skills
IX. Questions
30. To Summarize
1. RPA as technique provides CAMO’s and maintenance organization the ability to
robotize processes
2. Low intrusion as RPA mimics user interactions with current systems
3. Creates a 24x7 workforce
4. Allows human intelligence to focus on the creative tasks rather then the
repetitive tasks
5. Significantly reduces human error
6. Simply reduces costs
I. Meet gAVInBOT
II. RPA
III. Approach
IV. Scoring Model
V. Results
VI. Case Study
VII. To Summarize
VIII. Future Skills
IX. Questions
32. Future Skills
What can gAVInBOT MK1 already do:
• IPC/AMM downloads from MyBoeingFleet
• SAP document extraction
• EASA publication extraction and downloading
• AMOS Workpackage creation and printing
• File archiver
What is it currently learning:
• ADS-B data retrieval
• Repair & Purchase Order Management
• Further processing of EASA publications (OCR)
• Data Transfer automation
• Aircraft Lease documents verification
I. Meet gAVInBOT
II. RPA
III. Approach
IV. Scoring Model
V. Results
VI. Case Study
VII. To Summarize
VIII. Future Skills
IX. Questions
33. Future Skills
Future of EXSYN’s gAVInBOT MK1:
“Robots struggle when there is deviation from rules. They don’t make decisions.”
John Kilbride (Hunt, 2016)
AI technologies like machine learning and cognitive computing are being used to
train gAVInbot MK1 and be able to perform decision making.
I. Meet gAVInBOT
II. RPA
III. Approach
IV. Scoring Model
V. Results
VI. Case Study
VII. To Summarize
VIII. Future Skills
IX. Questions