Atheris for
Agentforce
Atheris is the platform of choice for implementing Agentforce, providing a clear and collaborative design approach.
Why use Atheris for Agentforce Implementation?
Quickly identify and articulate agent use cases
Collaborate with stakeholders on scope and guardrails
Design complete, working agents with clear logic and behavior using AI
Identify and capture requirements for actions and use AI to generate user stories
Accelerate Agent development by automating instructions
Faster evaluation, testing and debugging with AI generated test cases
Provide governance and stakeholder sign-off of diagrams and user stories
Control org complexity and tech debt with Agent metadata in the core metadata dictionary
Provide a change log of diagram, instruction and test cases versions
Smart features that accelerate Agentforce
Without a clear, process-led approach, designing and deploying autonomous AI agents can be complex. That’s why we developed Atheris for Agentforce —to provide you with the clarity, control, and confidence to design, build, and manage Agentforce agents effectively.
Whiteboard on ideas for agents that reinvent the business
Use AI to help ideate on use cases and provide the initial agent design.
Collaborating with cross-functional teams to fine tune the agent design and get everyone on the same page.
Keep track of your agents with their skills, topics (JTBD), instructions, and actions in an AIM (Agent Interaction Map).
Use AI to evaluate and suggest improvements to the agent design diagram before you waste time building and testing the agent.
Accelerate building topics with one-click automated instruction generation.
Use AI to generate user stories for metadata changes to support actions.
Ensure thorough test coverage. Each scenario – not just the “happy path” – is evaluated with the AI-generated test utterances.
Accelerate debugging by using the diagram to understand the changes required.
Iterate rapidly with confidence with an audit trail of diagram changes alongside the instructions and test utterances.
Apply light touch, yet rigorous governance. Design iterations are tracked in the audit trail. When you deploy the agent, version the design documentation with a formal sign-off.
Eliminate the risk of changes. Agent and Data Cloud metadata is automatically added to the metadata dictionary with impact and dependency analysis to the core metadata.
Get more achieved with your team by collaborating in live agent process mapping sessions with everyone on the same page.
Rapidly capture processes, then use diagrams to validate requirements, guide user acceptance testing, provide training and meet regulatory compliance.
Frequently asked questions
Agents use a reasoning engine to plan and take actions autonomously based on natural language but requires user input. Copilots can understand natural language but requires user guidance to perform a task.
Yes. Agentforce uses Prompt Templates, but they can be used independently using AI to deliver actionable outcomes such as account summaries or personalized emails.
Agentforce needs Data Cloud to be able to monitor and build event logs of agent activity. If Agentforce agents need data or unstructured data from 3rd party systems, then you should use Data Cloud as it is tightly integrated.
We are currently adding these new metadata types to the sync. Then they will automatically have all the notification, documentation, impact analysis, and dependency functionality.
We are prototyping a number of Agents that can accelerate the design, build, and monitoring of agents. Some are already available. We can draw process diagrams from text or images. We can write user stories for processes and suggest which metadata can be reused.
But we can see even more opportunities to build agents that build agents.