Here’s Every Product Innovation Announced at Imagine 2024 (and Even a Few That Weren’t!)

This week at Imagine, we unveiled the AI + Automation Enterprise System to give every customer expanded capabilities, including AI Agent creation, multi-agent orchestration, AI Agent and model management and governance, and the ability to go from discovery to automation in record time.

New AI Agents are a game-changing, powerful combination of AI and automation and represent the future of enterprise productivity, helping companies automate what was once impossible. AI Agents are able to complete complex cognitive tasks, like deciding on the best replacement product for a stock outage or understanding and routing incoming customer service tickets to the right place. They can also learn from and adapt to data and actions. But what makes our AI Agents more powerful than what you would see anywhere else is their ability to collaborate together across mission-critical enterprise processes spanning any function.

With the new AI Agent Studio, we have also delivered a platform that makes it fast, easy, and secure to develop, deploy, and manage specialized AI Agents for any use case across any function.

These are just part of the massive wave of product announcements from Imagine 2024. For complete details on every new AI capability, feature, and productivity-igniting solution, keep reading.

Many of these new capabilities are generally available with our latest v.33 release (noted with “GA” for generally available), with other innovations that will follow over the next few months (also noted). Reach out to your Automation Anywhere contact for more information on availability and pricing.

Let’s go!

AI Agent Studio to easily build, manage, and govern AI Agents

The new AI Agent Studio (GA) is the first low-code AI-powered solution to easily build, manage, and govern custom AI Agents for any cognitive task and with no data scientist required. Developers of all skill levels can create AI Agents, ground them in enterprise data, tune and templatize the agent prompts, and create guardrails for what agents can and can't do to protect company and customer data.

Here’s what else AI Agent Studio brings to the AI + Automation Enterprise System:

• AI Skills:
  • To make AI Agents even more capable, new AI Skills (GA) enable the packaging of multiple generative AI capabilities to support any task across the enterprise. Each skill includes a foundational model of choice, the option to connect to enterprise data via RAG, and an optimized prompt for consistently accurate and relevant outputs.
  • New prompt templates to experiment with prompt and model combinations to determine the best for each use case.
    • New custom prompt templates (GA) allow teams to create and share a prompt for use within an AI Skill so others can quickly infuse generative AI capabilities, even without previous experience.
    • New native prompt templates (January) provide out-of-the-box prompt templates for tasks like classification, summarization, and composition of emails.
  • Automation Anywhere’s native RAG service (GA) for connecting to enterprise knowledge.
  • New hyperscaler RAG services to embed enterprise knowledge into AI Skills. Expect RAG capabilities for Amazon Bedrock (October) and additional hyperscalers, including Azure (January).
  • New AWS and Google Cloud RAG service connectors (GA) provide new bot store packages for connecting to hyperscaler RAG services.
• Connect to any LLM:
  • Our open platform enables connections to any LLM, including private instances of third-party LLM models (GA) for the highest level of governance and control, including support for Google Gemini, OpenAI Assistants, and Anthropic Claude (GA); OpenAI GPT-4o and GPT-4V (July-October); and Hugging Face and Meta Llama (January).
  • New custom model support enables connections to any custom model built on AWS (January).
• AI security and governance controls:
  • New AI analytics with dashboards (October), including explicit and implicit quality metrics to monitor the use of AI and models, plus deep, drill-down analytics for automations using AI.
  • New monitoring and auditability (GA) to track data inputs and outputs with audit logging and analytics for prompts and model responses.
  • New automated model evaluation (later this year) to create custom evaluation criteria and rules that align with company policies to control model outputs and take action to block or flag responses.
  • New user feedback on model outputs in Automation Co-Pilot (later this year) allows users to rate model responses, improving model accuracy and performance.
  • New policy enforcement for model usage (October) when building automation using LLMs.
  • New toxicity scanning of prompt outputs (January) to ensure compliant and on-tone responses.
  • New prompt data security (GA) to manage prompt data storage or forward to an integrated security and event management (SIEM) platform.
  • New data masking AI guardrails (January) to scan input prompts and model responses for abusive content, threats, and sensitive data by intercepting queries to and responses from the LLMs to protect sensitive data like customer PII using smart obfuscation techniques.
  • New prompt data security (GA) to manage prompt data storage or forward prompts to an integrated security and event management (SIEM) platform. Data at rest and in transit is secured using industry-standard encryption algorithms, and access is controlled via permissions.

Automation Co-Pilot, your embed-anywhere enterprise assistant

• Automation Co-Pilot is now conversational thanks to an integration with Amazon Q (January), currently in preview. Now, business users can ask questions to request automations and AI Agents or quickly access knowledge bases via chat.
• New multi-session support (GA) across various business and messaging applications for the same user makes it easier to automate across the entire application landscape.
• New dynamic license mapping (GA) for automated user account provisioning, role mapping, and license allocation based on role so functional groups can be onboarded faster.

Document Automation for any document type

• New on-premises support for generative AI (GA) so customers can connect to private cloud generative AI models on all major hyperscalers, unlocking any document type, including unstructured documents.
• New document table extraction with generative AI (GA) so users can extract more complex table information.
• New Anthropic LLM support on GCP and AWS including on-premises, private cloud scenarios (GA). Anthropic also offers an increased context window for more complex queries and data extractions.
• Enhanced support for 32 additional languages, including Korean and Japanese (GA).
• New Document Workspace (October) to quickly execute search query testing, versioning, and batch testing for faster model setup, faster document extraction setup and testing updates in real-time, tracking versioned changes, and benchmarking performance.
• Scale and performance enhancements that improve extraction speeds (October) and offer a new option to run high-scale processing in the cloud (January and beyond).

Automator AI to accelerate the automation lifecycle

Enhanced GenAI Process Models (GA) are now tuned using over 300 million automations for discovery, documents, automation, and resiliency. And, we’ve announced even more new innovations to speed automation and AI deployments.

• For Autopilot:
  • New Autopilot (GA) uses generative AI to quickly transform process maps into automation drafts.
  • New Autopilot support for BPMN input (October) expands Autopilot beyond process discovery output (PDD) to accept inputs in BPMN format from any mining tool to build automations.
  • New text- and table-based process description (coming soon) supports free text or spreadsheet-based process descriptions as a starting point to build automations with Autopilot.
• For Automation Co-Pilot for Automators:
  • Text-to-automation for natural language inputs (GA) to generate process flows, including RPA bots, APIs, and forms.
  • Suggested next action (GA) to auto-complete automation development with in-app next-best actions.
  • Recorder support for natural language inputs (October) to generate UI automation flows with Recorder.
• For Generative Recorder:
  • New generative AI (GA) to identify changes and update in real-time, saving days of potential automation downtime and resources to search for the change.
  • New fallback visibility (October) for improved reliability for strategic apps and enhanced Recorder action to display supplementary information used in fallback.
  • Vision LLM-powered fallback (October preview) to use Vision models as another fallback approach to increase resiliency and reduce execution failures.

For Process Discovery:

• Updated Privacy Enhanced Gateway (PEG), now with Process Discovery redaction customization (GA), to enable the protection of sensitive data when capturing screenshots via Process Discovery including:
  • Define a customized value to replace sensitive data so users will still understand the screens, even if the data is masked.
  • Enable real-time redaction so power users can visually see the impacts of redaction in captured screenshots.

Solution Accelerator to speed the time to business value

The new Service Operations Solution Accelerator (GA) is the first of many designed to help teams get started with AI Agents quickly with AI Agent blueprints, out-of-the-box process templates, and how-to guides. Service Operations teams can deliver faster, always-on service for internal teams and external customers across use cases including order management, returns processing, ticket management, and Q&A.

Coming soon are Solution Accelerator packages for finance, HR, IT, healthcare, banking, and manufacturing.

Even more AI-powered innovations from Imagine 2024

Continuing the innovation momentum, here are more new features and enhancements that transcend the sections above!

• New Automation-as-a-Service:
  • New 100% cloud-based services (GA) for API-based tasks and processes to fully run automations on the cloud and alleviate customers of the cost and management burden of in-house infrastructure.
  • New Connector Builder (GA) to create packages to connect with any custom business application and expand automation across any complex enterprise process.
• Enhanced Process Orchestration:
  • New Process Templates (GA) enable developers to swiftly build automations with predefined templates for various business use cases, accelerating creation of process automation, RPA, and API tasks.
  • New real-time process intelligence (January) to make every process more efficient with better outcomes by tracking and monitoring the progress of complex process automations inside Automation Command Center.
  • Enhanced deployments of process automation from anywhere (January) via any third-party events like API, webhooks, or regular schedules.
  • CoE Manager expansion to manage an entire automation program and customize tracking of process ROI (October) across every process automation path in conjunction with new integrations to ServiceNow, UiPath, Jira (GA), and Blue Prism (January), plus visualizations in Power BI and Tableau (GA).
  • New child task debugging (GA) to simplify troubleshooting of complex and long-running processes that span many layered tasks.
• New Automation Command Center:
  • New Document Automation widget (GA) to track detailed usage, including page count, by various methods.
  • New export to PDF and Tableau (GA) to put outcomes in the right place for integrating with existing reports. Export to PDF makes it easy to integrate status into existing roll-up documents.
  • New process and AI analytics (GA) for detailed insights about long-running, complex processes and how they are using AI tools like LLMs.
  • New dashboards for Generative Recorder (October) to track how often Generative Recorder steps in to ensure automations successfully complete.
  • New monitoring of models and tokens used in bring-your-own generative AI subscriptions (October) to maintain real-time dashboards for AI consumption by automations.

Remember, these and other innovations are GA now or will be delivered over the coming months. Reach out to your Automation Anywhere contact for more information on availability and pricing.

Or, to get a personalized overview of the power of our new AI + Automation Enterprise System, request a live demo.