By Layne Alfred, Marketing Associate

Automation Anywhere has been fine-tuning their RPA solution for 14 years, and other companies are starting to take notice and adopt the growing RPA technology. It’s becoming exponentially popular due to the fact that it automates monotonous computer tasks and frees employees up to work on more value-added work (isn’t it about time?).

But before you jump right in to an Automation Anywhere RPA solution, it’s important to understand that Automation Anywhere’s bots have very different capabilities and are not “one-size-fits-all.”

For example, some bots love automating complex tasks while others are only able to complete tasks that are very structured, so it’s important to understand your business needs and current environment to find your perfect ‘bot’ match. The easiest way to start off on the right foot is to look at the most common type of data that your business encounters. Is it mostly:

Structured data? Any data source with clearly defined fields like:

  • Databases
  • Spreadsheets

Semi-structured data? Any document where the location of a field or line items may vary like:

  • Mortgage applications
  • Invoices
  • Contracts

Unstructured data? Any data that has no real rhyme or reason and is never organized in a specific manner like:

  • Instant messages
  • Voice recordings
  • Videos


If you answered “structured data,” TaskBot + MetaBot are the bots for you.

RPA taskbotTaskBots are simply a collection of action commands. They are the basic building blocks of Automation Anywhere’s RPA, meaning that you will start with a TaskBot no matter what other types of bots you plan to implement. TaskBots go hand in hand with MetaBots. The big difference between the two are that MetaBots are designed to be reusable and can quickly adapt to changes, whereas Task Bots simply record and mimic the actions of the human user on the screen (and they are required to launch a MetaBot). MetaBots start with a screen capture and map the data to a property within your system. If any changes or updates occur in an application, the user can just simply hit the “Calibrate” button. MetaBots will figure out what has changed and remap the field. They can also integrate with dynamic link libraries (DLLs) should you need to implement your own custom code. Since TaskBots automate rules-based tasks on the front end, they’re perfect for simple documents. Paired with MetaBots’ ability to scale, processing structured data is a breeze.


If you answered “semi-structured data,” IQ Bots are your guys.

RPA MetaBotThese are the most complex bots because they get smarter as they go. Using a combination of computer vision, machine learning and natural language processing, the IQ Bot processes semi-structured data with its own cognitive intelligence capabilities. This bot requires a bit more human interaction than the others because you have to teach it how to behave, especially at first. The bot will attempt to process the information on its own and show a green check mark next to data that it’s confident is correct, and will ask you, “is this correct?” for the items it’s not quite sure about. It will adjust itself based on the answers you provide and will try something new next time until it gets it right. IQ Bots will continue to learn from their mistakes in an effort to reach 100% STP. This makes the IQ Bot perfect for processing semi-structured data. When the format is generally the same across documents, the IQ Bot can very quickly pick up on the repetition, but also has the ability to recognize when a field is different.


If you answered “unstructured data,” your solution is: RPA + Artificial Intelligence.

RPA IQ BotWhile IQ Bots can learn from repeated exposure to patterns in semi-structured documents, RPA is not intelligent enough on its own to decipher completely unstructured data. Luckily, Artificial Intelligence (AI) applications capture and understand unpredictable data well enough to classify it, act on it and create automated workflows. Integrate with IQ bot and voilà– you’ve found a solution for processing unstructured data. AI services like IBM Watson, for example, utilize machine learning, predictive analytics and data visualization to make intelligent decisions. Like IQ Bots, AI technologies are always working to improve their own processes; they will continue to make better decisions the longer they’re exposed to the same type of documents. Plus, IQ Bots can easily integrate with open-source cognitive AI systems like Google TensorFlow via API if they need a little extra help deciphering and acting upon unstructured data.


RPA can be a powerful tool for automation if you match your process to the right bot. That’s why it’s important to establish exactly what type of data your business encounters most and which bot will do the job, even if that means your RPA solution will need a little help from AI capabilities. Just decide on your perfect match and get to automating!

Looking for a deeper dive into RPA? Click the image below to watch our free on-demand webinar on IBM’s use of Automation Anywhere RPA in their Automation Platform for Digital Business. 

IBM RPA Explained



Layne Alfred

Layne Alfred

Marketing Associate

About the Author: When I’m not writing or working on a content project, I love yoga, dogs and food.