The News: Automation Anywhere held a briefing with industry analysts to discuss its ongoing initiatives around AI and automation, and provided additional insight into the way it is deploying generative AI guardrails. Steve Shah, SVP Product at Automation Anywhere, laid out the specific steps the company is taking to safely deploy generative AI within its platform.
You can read the original Press Releases focused on Automation Anywhere’s recent generative AI announcement with Google Cloud and Amazon Bedrock, and the company’s announcement of Automation Co-Pilot + Generative AI for Business Users and Automators, and Document Automation + Generative AI.
Automation Anywhere Outlines Generative AI Guardrail Strategy
Analyst Take: Automation Anywhere has recently made several generative AI announcements, including those announcing partnerships with Google Cloud and Amazon Bedrock, as well as the company’s announcement of its core AI offerings, including Automation Co-Pilot + Generative AI for Business Users, Automation Co-Pilot + Generative AI for Automators, and Document Automation + Generative AI. In addition to discussing the enhanced functionality and efficiency afforded by incorporating generative AI into the platform, Steve Shah, Automation Anywhere’s SVP of Product took time to discuss the specific steps the company is taking to enact generative AI guardrails. He notes that data governance and the proper application of controls are at the forefront of Automation Anywhere’s AI strategy.
Securing Partnerships With LLMs That Respect Data Governance and Customers
At the heart of Automation Anywhere’s strategy for implementing generative AI guardrails is the decision to only partner with LLMs that respect data governance and privacy of Automation Anywhere’s customers. Further, Shah says that the company puts a premium on the use of secure models, and its selection of partners such as OpenAI, Google, Amazon, and Vertex AI underscore the commitment to only using models that handle data properly and securely.
Implementation of Prompt Guardrails to Limit Abuse or Misuse
Shah described several steps the company is taking to eliminate and lessen the likelihood of prompt abuse. Automation Anywhere is doing a significant amount of prompt engineering and controls development to ensure that users cannot submit prompts that are designed to elicit inappropriate or inaccurate results. Shah says that the company uses a lot of checks and balances to ensure that prompts are not able to go off the rails, and has established best practices for both prompt engineering and secure development with AI.
This is a sound and proper approach to implementing generative AI guardrails. A key best practice revolves around ensuring that inadvertent or intentional instances of prompt abuse or manipulation have been considered and addressed, prior to the tool being rolled out. However, it is a process that needs to be revised and updated, particularly as the models change, and the number and types of users and use cases increase.
Restricting the Use of Generative AI
Generative AI tools are new and shiny, and everyone wants to experiment and play with them. However, Shah says another key strategy for enacting generative AI guardrails is to provide enterprises with the ability to restrict its use. Shah adds that not everyone needs nor should have access to generative AI tools, and Automation Anywhere has implemented a control level that can be administered by the organization’s Center of Excellence (CoE) Manager, which has the ability to turn features and users on and off.
As the old sports adage goes, not everyone gets to touch the ball. While there will be some grumbling, it is far more prudent to limit the use of newly-minted generative AI tools to the employees and roles that truly will see and drive benefits from its use.
Deploying a Human in the Loop
Like other vendors deploying generative AI, Automation Anywhere also promotes the use of a human in the loop (HITL) to review generated answers or content. This ensures that no matter what the AI does, there is a path for human to double check the result before the content is executed or acted upon. Shah says that due to efficiency and productivity benefits, the slight slowdown in workflow due to a HITL worth it, in terms of implementing generative AI guardrails that reduce the risk of blindly relying on generative AI output.
Continuous Performance Monitoring
The final step deployed by Automation Anywhere is the use of continuous monitoring of generative AI tools to ensure they are delivering value and are being used responsibly. Shah says this message has been resonating with enterprises, which are rightly focused on making sure that investments in generative AI are providing ROI, which will continue to rise in importance as the costs of rolling out generative AI across dozens or hundreds of use cases continues.
Disclosure: The Futurum Group is a research and advisory firm that engages or has engaged in research, analysis, and advisory services with many technology companies, including those mentioned in this article. The author does not hold any equity positions with any company mentioned in this article.
Analysis and opinions expressed herein are specific to the analyst individually and data and other information that might have been provided for validation, not those of The Futurum Group as a whole.
Other insights from The Futurum Group:
UK AI Regulations Criticized: A Cautionary Tale for AI Safety
Generative AI Investment Accelerating: $1.3 Billion for LLM Inflection
Google Cloud Engineering Exec: Welcome to Generative Engineering
Author Information
Keith Kirkpatrick is VP & Research Director, Enterprise Software & Digital Workflows for The Futurum Group. Keith has over 25 years of experience in research, marketing, and consulting-based fields.
He has authored in-depth reports and market forecast studies covering artificial intelligence, biometrics, data analytics, robotics, high performance computing, and quantum computing, with a specific focus on the use of these technologies within large enterprise organizations and SMBs. He has also established strong working relationships with the international technology vendor community and is a frequent speaker at industry conferences and events.
In his career as a financial and technology journalist he has written for national and trade publications, including BusinessWeek, CNBC.com, Investment Dealers’ Digest, The Red Herring, The Communications of the ACM, and Mobile Computing & Communications, among others.
He is a member of the Association of Independent Information Professionals (AIIP).
Keith holds dual Bachelor of Arts degrees in Magazine Journalism and Sociology from Syracuse University.
