Amazon Web Services (AWS) this week revealed it is making it possible for end users who have installed its Amazon Quick artificial intelligence (AI) assistant to build agents that, as they run in the background, will continuously automate a wide range of tasks.
Announced at the AWS New York Summit, the AI agents will perform a range of autonomous tasks in the background, including processing orders or monitoring interactions occurring across emails, Slack channels or customer relationship management (CRM) applications.
Additionally, AWS has added an activity feed through which end users can consolidate email, messaging, calendar, and tasks into a single prioritized view that over time learns which messages are more important and which threads or topics are routinely skipped. Every interaction, correction, and outcome serves to make the AI agents better over time.
Finally, AWS also revealed 16 new integrations have been added to Amazon Quick, including connectors for Adobe, Cisco Webex Meetings and Video Messaging, Dun & Bradstreet, Figma, Google Chat, HG Insights, Microsoft OneNote, Moody’s, Shopify, Smartsheet, Snowflake, Visier, WhatsApp, Zapier, and ZoomInfo.
End users can discover, install, and share Skills, Agents, and Connectors via a built-in catalog that addresses use cases involving sales, finance, marketing and project management. Applications and AI agents built in Quick can also be published and shared with a wider team.
Collectively, these capabilities will change the way individuals work and collaborate in enterprise organizations as AI companion tools are increasingly put directly into the hands of end users, says Jose Kunnackal John, director of product for Amazon Quick.
The challenge and the opportunity is to not just automate existing workflows and tasks but to completely rethink how those tasks are actually completed, he adds. In effect, AI has now become a change management challenge, says Kunnackal John. “It’s really now about AI fluency,” he says.
In general, two distinct types of AI tools have emerged for end users. One is a personal assistant that learns end user preferences, while the second is various types of AI agents that have been trained to perform a specific task. Amazon Quick combines those capabilities within a single application that makes it simpler for the average individual to manage AI workflows across multiple tasks.
Each end user can determine for themselves what level of autonomy to assign to any agent, while organizations that adopt Amazon Quick are still able to centrally apply guardrails and manage governance policies, says Kunnackal John. For example, an organization may decide that while they are comfortable allowing an AI agent to read email, they might not yet want to allow an AI agent to send emails that might contain mistakes or, more troubling still, outright hallucinations.
It’s not clear to what degree AI assistants and agents have already changed the way individuals and teams work, but the implications are profound. In fact, the only thing that remains to be seen now is how AI agents will navigate conflicting goals that various teams within and outside an organization may have. Regardless of the current level of AI fluency within an organization, the one thing that is all but certain is that much of the cognitive load that makes work more tedious than it should be is about to be sharply reduced.
