Changelog, improvements, and roadmap milestones for our mission to revolutionize enterprise AI
Added OpenAI Decisions and Microsoft-Decision-1 alongside Jev in the shared Evaluate Decision activity. Connect directly to OpenAI, or use Microsoft-Decision-1 through OpenRouter or Microsoft Foundry. Evaluate conditions, choose between defined options, and score inputs against a rubric, with typed results in the Flow Designer, probabilities, confidence signals, and explicit refusal handling for review and fallback paths.
Follow workflow execution directly on the Flow Designer canvas. Track the latest run, see its route and active activity, and use visit-count highlights to understand repeated paths. Switch to a historical run when investigating earlier executions.
Reuse shared Randol workflows with your tenant's own service connections. Connection policies can keep a fixed Randol connection, allow an optional tenant override, or require your own connection, with the effective connection checked for the selected execution environment.
Expose published Randol workflows as authenticated MCP tools so external AI agents can invoke them. Validate inputs against the flow schema, look up execution status, and cancel runs through the same integration, backed by Randol's normal workflow lifecycle.
Use Amazon Bedrock models and embeddings in Randol agent workflows. Configure your AWS region, model IDs, and supported authentication through a dedicated service connection, with structured-output options and an embedded setup guide.
Connect your Ruddr workspace to Randol workflows to work with projects, clients, members, time entries, and other supported resources. List, retrieve, create, update, and delete records where supported, with cursor-based pagination, secure API-key authentication, and built-in connection validation.
Create email templates with subjects, To/Cc/Bcc recipients, HTML content, and a plain-text alternative. Organize templates in nested folders and insert schema-backed placeholders into headers and bodies using draggable chips or insertion menus.
Connect Telegram bots to Randol workflows with message triggers for private and group chats, sender and chat filters, and per-environment credentials. Send messages, receive attachment metadata and captions, download incoming files, and send stored files back to Telegram.
Build managed chat experiences with resumable conversation history, grouped agent progress, and file uploads and downloads. Preview supported file content inline or in a dialog, while keeping access scoped to the user and isolating scripted previews.
Introduced the shared Evaluate Decision activity under Agentic AI, bringing Jev decisioning to Randol through TypeSafe and OpenRouter. Flows can evaluate yes/no conditions, choose from named options, or score inputs against an ordered rubric, then use the returned probabilities and confidence signals to route work, automate decisions, or request human review.
Connect Randol agents to external agents through the Agent2Agent (A2A) protocol. Discover remote agents and delegate work, or expose Randol Flows as A2A agents with discoverable Agent Cards so external agents can find their capabilities and invoke them. Support immediate responses and long-running tasks that wait durably and resume as work progresses.
Connect Stripe, Adyen, PayPal, PayU, Przelewy24, and Tpay to Randol workflows. Create hosted payment requests, check payment status, and start flows from verified payment notifications. Access balances and transaction history where supported by the provider and your account permissions, with dedicated activities and embedded setup guides.
Bring your Databricks lakehouse into Randol workflows. Execute SQL queries and work with Unity Catalog, volume files, Vector Search, Genie, and Model Serving through dedicated activities, using OAuth service-principal credentials or a personal access token for your workspace.
Deploy runtime components into your own AWS account or Google Cloud project through Randol release plans. Use AWS ECS Express or Google Cloud Run, configure resource sizing and scaling, and manage environment-specific runtime deployments alongside Azure.
GA release with multi-cloud deployment blueprints and billing.
Added ability to define placeholder templates ready for embedded placeholder templates for email and SMS communications.
Added number of new integration connectors including Google Workspace, HubSpot, Confluence, SharePoint, and Slack.
Implemented UI designer, now able to build UI on top of services and see UI based on the data stores with instant UI feedback.
Implemented full agent vector store and memory storage capabilities for enhanced AI agent performance.
Introduced changesets allowing deployments based on changeset names for flows and UI bundles.
Chatbot connector introduced to allow flows to respond to chat requests directly.
First draft of the UI builder. Config driven. Now allows deployable UI based on services for data and functionality.
Onboarding our first enterprise cohort with expanded agent tools, environment templates, and audit exports.
Ability to configure agents and vector storage activities within a flow.
Flows triggerable by events. Supports event catalogs and external events from third parties (Example JotForm)
Ability to raise events from an operation handler. Based on pre-defined event contracts.
Ability to deploy generated services to dedicated environments with configurations per environment.
Ability to generate domain logic and application handler logic code via beta version of AI code generation service.
Ability to compile stored code via beta version of the AI code generation service.
Dynamic store for all generated code, initially for domain and application logic.
Listen for and publish events within a flow.
Select and execute Randol operations (microservices) within a flow.
Create new operation contracts and placeholder implementations.
Add new events to the catalog and make them available to flows.
Define and execute operations (dynamic microservices).
Interface to explore and manage dynamic data schemas and their relationships.
Flow surface and activity rendering/updating.
Foundation for UI patterns and service connectivity.
Core component foundation for dynamic workflow orchestration.
Activity/rules engine capabilities and core platform components: auth, DI, logging, data storage, messaging.
