🎓 Teach Michael: What shouldn't you automate?

[Michael_S]

Hello automation gurus,

As part of my ongoing efforts to learn more about intelligent automation (I'm still super new, but I'm learning fast!) - I've been asking you for help on questions that have lurked in my mind.

You, being the incredibly kind folks you are, have already taught me so much across topics like:

- How to find the best processes to automate
- What's the role of Senior Leadership when scaling intelligent automation across an org?

Today though, I'm in the mood to play Devil's Advocate. So the question I'd love to hear from you all on is...

---

What are your rules / philosophies on things that should never be automated?

---

I'm hoping for a heated debate on this one! You might like to contribute:

• Your view on processes that should always be handled entirely by a human
• Automation projects you've tried in the past, which didn't succeed (and what you learned)
• Advice for beginners on automations that might seem simple, but are actually incredibly complex

Please do dive in and share your thoughts, every contribution is truly welcome. You'll not only be helping me, you'll also be teaching other newcomers with your experience, and that's what community is all about.

Thank you!

------------------------------
💙 Michael
(I'm part of the SS&C Blue Prism Community Team)
------------------------------

[lookman]

Hi Michael.

I love how you are going about learning the RPA world.

One thing that I think you would never automate is a process that is rarely used. For example, there may be a yearly or even a quarterly process that is pretty simple for a person to do. There are costs associated with automation so unless the process has significant value, it would not be worth the effort/costs.

------------------------------
Jack Look
Sr Product Consultant
Blue Prism
------------------------------

[harish.m]

Hi Michael


One of the project I have worked required hand off  by human and implemented on desktop attended automation need to be careful about automating and running the bots on user machine since there are lot of parameters involved and lead to inconsistent result.

------------------------------
-----------------------
If I answered your query. Please mark it as the "Best Answer"

Harish M
Lead developer
America/New_York TX
------------------------------

[Neel1]

I got stuck up in one use case which require bot to work upon multiple heavy excel file and large business rules . One of file is of 450000 rows and process involve strict SLA as well with real time progress to be made available to human. Bot processing time also increased due to all business rules and validation. 

------------------------------
Neeraj Kumar
------------------------------

[Michael_S]

First thing to say is, you are all so helpful and it's warming my heart 💙

@Jack Look - thanks for kicking us off! I love your reminder on ROI, and to consider that automation always has a cost. Great to keep in mind.

@Harish M - humans make mistakes, take caution whenever they're involved!  Message received 🫡

@Neeraj Kumar - that sounds like a NIGHTMARE. How do you even begin to break down all that complexity?!

Thanks for your horror stories and lessons so far everyone, I'm learning a lot!

------------------------------
💙 Michael
(I'm part of the SS&C Blue Prism Community Team)
------------------------------

[Babjee24]

Hi Michael,

Having worked with Banking domain quite a sometime now I've noticed not to automate the process that has multiple 3rd party web applications  involved(these applications are not maintained by operations team that we interact with) that has regular patches or changes in the application design, as these are 3rd party applications no notifications will be sent on prior notice, this leads to unexpected screens or change in UI elements resulting bots failure which ultimately results in breaching the SLA of the BAU since the change is not expected the operations team gets into tighter SLAs in the last moments.

Having the changes happening quite a few times bots fail multiple times and the RPA solution instead of cutting shot the manual work would infact doubles it up for the operations team.

------------------------------
Babjee Vangipurapu
Senior RPA Developer
Wonderbotz
India
------------------------------

[Denis__Dennehy]

Great question,

I think a clear prioritisation process to score your pipeline and make decisions on what to automate would usually include questions around 'system change',  'process stability',  'process complexity'  and you will come to a decision of what to automate and not automate based upon those questions.  Blue Prism used to have a prioritisation scorecard we gave to customers (originally from Co-op bank),  these days there is the PDT tool.

So,  is a system going to be totally changed in an upcoming update? - if so you probably want to postpone automation
Is the process a new one that is constantly changing?  - if so you probably want to let it stabilise first
Is the process extremely complex and will be costly to automate but only has a very small benefit?  If so you probably will have better candidate processes to automate first.

Like most things I say,  these rules are there to be broken.  We had a customer once that automated a complex process even though the main system was due to be replaced in 6 months.  They made that decision because the ROI was very high,  even for 6 months because 30+ FTE worked the process.  As it turned out IT over ran in their system replacement project (isn't that always the way) and the automate was running for 2 years.

[Mukeshh_k]

Totally agree on all of the above points @Michael Shillingford  ! Very well put together by the community members/automation experts in the above thread !

Few of the other things I would take a note of before adding anything to automation pipeline:

• Systems which are not likely supported by IT Team (to Sort Accesses before committing (Don't Run Before you can walk))

• Does not have a Knowledgeable team or SME or Process Owners with decision making capabilities(to avoid endless loop of lack of transparency at times, customer complaints, bottlenecks)

• Involved employees not having a vested interest in the outcome of the project

• Strategic relevance, Impact on Business KPIs and benefits calculations (anything less than 1 FTE should be avoided, if possible) 

In the initial days - One of the processes I have tried and could not make it through had to do with the GDPR policy & Outside EU access - Discovery sessions with SMEs involved failed to spot and consider few restricted security policies - time when we had to start doing IAM - we were notified by IT that we would require a Group Security Officer approval, as systems identified were accessing to highly secured and confidential Network drives  (Worked on SMEs configs but never got approval for RPA environment) - which indeed later was denied by GSO but took week's of our time and effort which zero result outcome but left us a learning to to ensure adding this Important check to get a GSO approval or make a query before committing or adding anything in the automation Pipeline.

My advice for new beginners would be : not to just focus only on the development but consider knowing the strategic relevance(purpose/problem statement you are solving), financial Numbers, selection criteria, ROM, ask questions around it and know your automation candidate well.

------------------------------
Kindly up vote this as "Best Answer" if it adds value or resolves your query in anyway possible, happy to help.

Regards,

Mukesh Kumar - Senior Automation Developer

NHS, England, United Kingdom, GB
------------------------------

[EVIPUTI]

I've had past experiences where automation projects didn't succeed because they were too complex or required too much customization, which made the cost of implementation outweigh the benefits of automation. From those experiences, I've learned that it's crucial to carefully evaluate the feasibility and ROI of an automation project before committing to it.

For beginners, I'd recommend starting with simple automations that have a clear ROI and can be easily implemented. It's important to take the time to understand the process thoroughly and identify any potential roadblocks before beginning the automation process.

------------------------------
------------------------------
Vipul Tiwari
Senior Process Simplification Developer
Amazon
------------------------------
------------------------------