Automation Anywhere AI Agent is quite good overall, and I believe the main area for improvement is working on different OCRs to extract data properly. The OCR and document understanding of Automation Anywhere AI Agent still struggle with complex and unstructured formats such as invoices with multiple layouts, poor scan quality, or handwritten fields. Even after training IQ Bot, I sometimes see accuracy issues and a need for manual validation. Another improvement area is error handling and debugging since when a bot fails in production, the logs are not always detailed enough, increasing troubleshooting time.
Account Director at a tech vendor with 10,001+ employees
Real User
Top 10
Jan 17, 2026
Automation Anywhere AI Agent could be improved in a few ways. Since I am currently working with Salesforce and I understand that to compete Automation Anywhere as a middleware product with Mulesoft or Informatica, the implementation phase needs to be quicker. The POC is always successful, but the implementation phase has to be quicker. The business needs to understand this better. Automation Anywhere should get into the business team and focus on explainability at the business level, not just a technical thing, but in the language or discussion where the business team can understand. There is clear business reasoning missing right now. The learning is strong within the process for sure, but there is limited reuse of learning across processes, which I think is very important. Simple model governance for business users is needed so that anyone should be able to use or create bots on Automation Anywhere. Right now, consultants and partners are needed in place. Additionally, Automation Anywhere should have native integrations with core platforms such as Oracle, SAPs, Salesforce, and ServiceNow. It should have out of the box integration rather than going for APIs because the deals which are not getting closed with Automation Anywhere or UIPath are purely because these organizations do not have out-of-the-box connectors available.
Co Founder at a media company with 10,001+ employees
Real User
Top 20
Dec 23, 2025
When considering weaknesses and improvements, the platform does not give us the liberty to use our own features where we can bring out creativity. We have to map the process to whatever is available, and the bidirectional integration is missing. If we have another agent coming from another tool and we are using it, integration is very difficult. Price is an issue because when we go with AI, we are getting more efficiency. Comparatively to other tools, it is on the higher side. However, we do not consider the price point when we choose this kind of tool for banking. We go for the compliance side, license feasibility, and other factors. Pricing is high in the market, but we are not in pricing mode. We see other advantages where we can use license solutions that can be used in the majority of cases. We give business-as-usual support to the customer, so we bring in the cost to our bandwidth and do not charge the customer. There are many combinations we employ. Additional features I would like to see in the future include faster learning capabilities. For single use cases, we require multiple agentic AIs. If we could combine them into one and do the decision-making with a single price, that would be better. Currently, for each specific condition, we use one agentic AI and then combine another agentic AI on top of it that controls all the conditions. This creates a complex situation, but if it were simplified, it would be good.
Learn what your peers think about Automation Anywhere AI Agent. Get advice and tips from experienced pros sharing their opinions. Updated: February 2026.
Automation Anywhere AI Agent is the #7 ranked solution in top AI Finance & Accounting solutions, #8 ranked solution in top AI Sales & Marketing solutions, #8 ranked solution in top AI Content Creation solutions, #77 ranked solution in top AI Customer Support solutions, and #206 ranked solution in top AI Data Analysis solutions. PeerSpot users give Automation Anywhere AI Agent an average rating of 6.8 out of 10.
Automation Anywhere AI Agent is quite good overall, and I believe the main area for improvement is working on different OCRs to extract data properly. The OCR and document understanding of Automation Anywhere AI Agent still struggle with complex and unstructured formats such as invoices with multiple layouts, poor scan quality, or handwritten fields. Even after training IQ Bot, I sometimes see accuracy issues and a need for manual validation. Another improvement area is error handling and debugging since when a bot fails in production, the logs are not always detailed enough, increasing troubleshooting time.
Automation Anywhere AI Agent could be improved in a few ways. Since I am currently working with Salesforce and I understand that to compete Automation Anywhere as a middleware product with Mulesoft or Informatica, the implementation phase needs to be quicker. The POC is always successful, but the implementation phase has to be quicker. The business needs to understand this better. Automation Anywhere should get into the business team and focus on explainability at the business level, not just a technical thing, but in the language or discussion where the business team can understand. There is clear business reasoning missing right now. The learning is strong within the process for sure, but there is limited reuse of learning across processes, which I think is very important. Simple model governance for business users is needed so that anyone should be able to use or create bots on Automation Anywhere. Right now, consultants and partners are needed in place. Additionally, Automation Anywhere should have native integrations with core platforms such as Oracle, SAPs, Salesforce, and ServiceNow. It should have out of the box integration rather than going for APIs because the deals which are not getting closed with Automation Anywhere or UIPath are purely because these organizations do not have out-of-the-box connectors available.
When considering weaknesses and improvements, the platform does not give us the liberty to use our own features where we can bring out creativity. We have to map the process to whatever is available, and the bidirectional integration is missing. If we have another agent coming from another tool and we are using it, integration is very difficult. Price is an issue because when we go with AI, we are getting more efficiency. Comparatively to other tools, it is on the higher side. However, we do not consider the price point when we choose this kind of tool for banking. We go for the compliance side, license feasibility, and other factors. Pricing is high in the market, but we are not in pricing mode. We see other advantages where we can use license solutions that can be used in the majority of cases. We give business-as-usual support to the customer, so we bring in the cost to our bandwidth and do not charge the customer. There are many combinations we employ. Additional features I would like to see in the future include faster learning capabilities. For single use cases, we require multiple agentic AIs. If we could combine them into one and do the decision-making with a single price, that would be better. Currently, for each specific condition, we use one agentic AI and then combine another agentic AI on top of it that controls all the conditions. This creates a complex situation, but if it were simplified, it would be good.