💻Document Automation Swag Discussion Series Week 2: Share for a chance to win SWAG

The main “aha” moment was:

• If you have used IQBOT, you will realize how quick it is to create a LI with GenAI prompts and start using it in a few minutes. Having said this, I would recommend testing with a large set of documents to make tweaks in setting up Field/Document Rules along with GenAI prompts for better results.
• Watching the “Setting Up Control Room Prerequisites for Document Automation” tutorial. It looks a bit scary, but you will get the hang of it once you set it up. This is gold 🪙. If you have ever done this setup for the first time, you would really appreciate this video.

Hi, Before attending the camp I had a question:

How to set up a Document Automation process for Business users to Validate?

From the learnings, in short below are steps for setting up users.

1)Login as Control room administrator and create a new co-pilot users for business user account to validate Documents with automation co-pilot license.

2)Create a customer role with required permissions(view/run bots, Launch validator) and assign it to Copilot users(Business).

3)Next, login as co-pilot administrator, Create a team and add validator users.

4)Assign the team to the process and assign scheduler user(which allocates CR resources to process for BU).

Thanks,

Lahari

As someone whose never used document automation, I appreciated learning how to get the system set up, training a model, and creating aliases.

I had this question before joining the live stream🎥, and it got answered during the session.

Question: Is there a way we can reduce processing time for extraction in bot runners?

Answer: The processing time will be high depending on the number of fields you are extracting. If you are using GenAI for all the fields, it adds a lot of time. Overall, it depends on the model and the deployment time of Automation Anywhere.

Hi @Lu Hunnicutt,

I wanted to share some insights from my first two weeks of Document Automation Training Camp, where several key questions were addressed:

Q: How can I effectively use Document Pre-Processing?
A: I learned how to improve document quality using techniques like grayscale conversion and image enhancements, boosting OCR accuracy for challenging documents.

Q: How does Document Automation handle varying field labels?
A: By defining aliases for different document types, I can ensure consistent data extraction even when labels differ, using practical examples from the course.

Q: How is Generative AI impacting Document Automation?
A: Generative AI is transforming document automation by enabling accurate data extraction from unstructured documents. With tailored prompts, I can boost precision and tackle specific needs, like distinguishing between property and mailing addresses in lease documents.

Q: How can I efficiently set up a document automation process using Automation Copilot?
A: The "Control Room Prerequisites for Document Automation" course provided valuable guidance on creating and managing roles, workflows, and practical tasks such as data extraction, resulting in a fully automated and efficient process.

Hi @Lu Hunnicutt ,

for me the “aha” moment was the use of the aliases these work amazing! 🚀

Also seeing the benefit of “correct” prompting Open AI to get better results out of the model.

Looking forward to the next weeks game plans! 💪

Keep up the great work on the courses you provide for us! 👍

Hi @Lu Hunnicutt,

The question I had before starting the Document Automation Training Camp were:

Q1: How is Document Automation different than IQ Bot?
Ans: The Document Automation uses Gen AI along with the OCR engines to extract data from unstructured documents.

Q2: How user-friendly the Document Automation would be?
Ans: It is very easy to understand and user-friendly as it is a process and NOT just a task bot.

I still have a question that I would like to ask in this forum.
Q: Is there a plan to add more OCR engines (along with Abbey and Google Vision) in the future?

Hi @Lu Hunnicutt,

Does it possible use DA , Extracction action, without  processing in a runner, Like iqbot?

Does it possible reduce the time for processing?, because comparing with Iqbot, it’s too slow

Thanks!

Hi @Lu Hunnicutt,

As for @Vatsy , my “aha” moment was extracting data from unstructured documents like a legal or even a resume / CV using GenAI capabilities which earlier was very difficult to extract using IQ BOT or event with string manipulations.

Even reducing the training time as compared to IQ BOT in Document Automation.

Thanks for the mention @Zaid Chougle!

Hi @Lu Hunnicutt ,

I’d like to share a question and answer on Document Automation that was addressed during the second week of the Document Automation Training Camp.

Q: How does Generative AI actually improve Document Automation, and what new skills can I expect to gain as an RPA Developer from this course?

A: Generative AI is transforming Document Automation by enabling accurate data extraction from unstructured documents—a challenge I often face as an RPA developer. In this course, I’m learning how to use and customize prompts to fit specific needs, like distinguishing between a property address and mailing address in lease documents. Mastering these techniques has been a major “aha” moment, as it significantly reduces errors and saves both time and resources. These skills boost the accuracy and efficiency of automations!