- What Launched: Helix Assistant for PingOne DaVinci adds opt-in AI help inside the flow canvas.
- Who It Helps: It helps DaVinci admins and identity teams explain, troubleshoot, and build orchestration flows with an embedded AI assistant.
- Why Now: Some teams lack approved AI clients but still want AI help in DaVinci.
Meet the Helix Assistant for PingOne DaVinci
Ping Identity has launched a new Helix Assistant for PingOne DaVinci, which brings AI help into the canvas to explain, troubleshoot, and build orchestration flows that deliver seamless user journeys.
Figure 1. The embedded assistant offers prompt suggestions to help users ask questions about their DaVinci work.
The new AI assistant is a chat box built directly into the DaVinci canvas. You ask questions in plain language, and it helps you explain flows, troubleshoot problems, and build supported flows.
DaVinci is Ping's SaaS identity orchestration service, which connects the steps of an identity journey, like sign-on, multi-factor authentication (MFA), and risk checks, into one visual flow. The assistant runs on PingHelix, the shared, governed foundation Ping built to embed AI securely and safely across its products and platform.
Figure 2. The assistant appears alongside a visual sign-on flow, keeping AI guidance next to the orchestration canvas.
Why Identity Orchestration Flows Are Hard to Decipher
An orchestration flow holds many decisions in a single canvas. Over time, those decisions pile up, and the people who made them often move on to other work.
When you inherit a flow, you trace nodes, subflows, embedded scripts, and variables one by one to learn what it does. That takes time before you can change anything with confidence.
Failed executions add more work, since you have to connect log data back to the configuration. Building a new experience from a blank canvas also takes effort, even for skilled teams.
Why Embedded AI Matters for Identity Teams Now
Ping sees embedded AI and programmable identity features becoming an expected part of how buyers evaluate identity platforms. It's a practical way to bring AI help to the people who run identity every day.
AI-first headless identity is Ping's approach to delivering programmable identity through agent-ready markdown docs, skills, MCP servers, and a command-line interface (CLI) on top of existing APIs, software development kits (SDKs), and infrastructure-as-code patterns. Teams with approved AI clients can work with DaVinci through headless features like the PingOne MCP Server overall and specific DaVinci MCP Server.
Some organizations haven't approved those AI coding agents and clients yet. One large enterprise customer told Ping its organization hasn't approved any AI coding agents yet. For now, that team prefers assistants embedded in Ping products.
If your organization won't give you AI agents, you can still take an intent-driven approach to orchestration. Intent-driven means you describe the outcome you want, and the assistant helps with the steps.
Early adoption of this feature already reflects that interest in embedded AI. As of September 2, 2026, seventy-eight paid DaVinci environments across forty-eight customer organizations had already enabled the assistant and none had opted out.
Where DaVinci Teams Put the Assistant to Work First
Different roles reach for the assistant for different reasons. Here are three places it can help right away.
Admins inheriting flows built by someone else
A new admin often takes over flows that someone else built long ago. The assistant gives that admin a clear summary to start from, instead of a long trace through the canvas.
Follow-up questions fill in the gaps, like how many subflows a flow uses or which variables it sets. That makes it easier to plan a first change with confidence.
Helpdesk and support teams triaging failed sign-ons
When a user can't sign on, the helpdesk needs a likely cause quickly. With an execution ID, the assistant can analyze the failed run and suggest where the problem sits.
Support can then hand the identity team a clearer starting point. The team still confirms the cause in analytics and logs before changing the flow.
Architects reviewing flows before launch
Architects want to know a flow is ready before it reaches users. Asking the assistant to check for configuration problems adds a quick review pass alongside normal testing.
It also helps architects explain flow logic to stakeholders who don't work in the canvas every day. A simple summary makes design reviews easier to follow.
Explain, Troubleshoot, and Build DaVinci Flows with AI
The assistant supports three kinds of work in the canvas. Each one starts with a plain-language prompt and ends with you deciding what happens next. In each interaction, the assistant can also look up information from PingOne DaVinci’s documentation and surface it in the chat window. This helps admins by keeping context on the canvas and not having to navigate for information in another browser window.
Explain flows, subflows, and embedded scripts
Explaining a flow means getting a plain-language description of what it does. You can ask about a whole flow, selected nodes, a subflow, an embedded script, or a variable.
Example prompts include:
- Flow Overview: "What does this flow do?"
- Flow Structure: "How many subflows are associated with this flow?"
- Embedded Scripts: "Help me understand the script embedded in this flow."
- Variables: "List the variables configured in this flow."
When you want more detail, the assistant can point you to relevant Ping documentation. Follow-up prompts let you go deeper on any node or step.
Troubleshoot configuration issues and failed executions
Troubleshooting means finding out why a flow doesn't behave as expected. The assistant can validate a flow's configuration, flag likely problems, and recommend fixes.
For a failed run, share the execution ID or simply ask about a specific user, and the assistant analyzes what went wrong. It can also summarize flow executions from the past fifteen minutes.
- Configuration Check: "Are there any configuration issues in this flow?"
- Failed Sign-On: "Why was the user <email> not able to log in?"
The assistant recommends fixes, and you decide which ones to apply. Your analytics and log data stay the reference point for confirming the cause.
Build supported flows from a natural-language prompt
Build goes beyond advice: the assistant creates the flow, not just the instructions. Describe the experience you want, review the proposed flow, and select Add to Canvas. When you Add to Canvas, it increments the version so that you can revert back to an earlier state at any point if you are unhappy with a change. Each time you Add to Canvas, you get a configuration to test and refine—not a chat response you still have to implement by hand.
The process follows three steps:
- Describe the experience: Try a prompt like "Create a registration flow where the user supplies username, email, and password followed by email verification."
- Answer clarifying questions: The assistant asks for any details it needs to shape the flow.
- Review and add the flow: The Helix Assistant presents a flow summary. If the summary sounds good, click Add to Canvas.
Build works for these supported use cases:
Each proposed flow is a starting point to review, test, and refine. You stay responsible for confirming it meets your requirements before it goes live.
How PingHelix Keeps People in Control of AI-Assisted Changes
PingHelix is the foundation Ping built so its product organization can embed AI with a consistent, governance-oriented approach. The goal is one shared design across products, instead of separate, unrelated assistants.
PingHelix is designed around three principles:
- Auditability: PingHelix treats auditability as a core design principle for AI in Ping products.
- Guardrails: PingHelix applies guardrails intended to keep each assistant focused on its approved tasks.
- Product-Focused Agents: Each assistant is scoped to its product and designed to decline out-of-scope requests.
These are design and control principles. They don't guarantee that every response is accurate or that every misuse is blocked.
That's how people stay in control at every step. The Helix Assistant for PingOne DaVinci is opt-in, and nothing reaches the canvas until an admin reviews it and clicks Add to Canvas. Even then, you decide whether and when to promote the flow to production.
Best practices for reviewing assistant output
Outputs can be inaccurate or incomplete, so build these habits into every session:
- Validate Before Applying: Check each recommendation against your flow and requirements before you make a change.
- Use Follow-Up Prompts: Ask a narrower question when an answer is too broad or misses context.
- Compare with Analytics and Logs: Confirm troubleshooting guidance against your analytics and log data.
- Treat Answers as a Guide: Let responses guide your work, and rely on your own testing as the final word.
What DaVinci Admins Gain from AI Assistance
The assistant gives identity teams practical help across daily flow work.
- Clarity on Inherited Flows: You can get a plain-language explanation of a flow without tracing every node yourself.
- A Starting Point for Diagnosis: Execution ID or Username/UserID analysis and configuration checks point you toward likely causes and fixes.
- A Head Start on Common Experiences: Supported flows like password reset begin from a proposal you review, instead of a blank canvas.
- AI Help without Having to Procure a New Tool: Teams that haven't approved external AI clients can still use AI inside DaVinci.
How to turn on Helix Assistant for PingOne DaVinci in the PingOne console
The assistant is opt-in, and there's no additional fee for DaVinci customers. An admin turns it on from the PingOne console in each environment where the team wants it:
- Go to Settings > Environment Properties.
- Select Manage Opt-Ins.
- Select Helix Assistant for PingOne DaVinci.
- Click Save.
Figure 3. Administrators enable Helix Assistant for PingOne DaVinci from the environment’s Feature Opt-In settings.
After you opt in, open the assistant from the button in the upper right header bar next to your user profile. You can also select part of a flow, right-click, and choose Ask AI to focus on just that selection.
Because the assistant works where you build, you don't need to switch tools or copy flow details into a separate chat. Your questions and its answers stay next to the flow you're working on.
AI-Assisted Orchestration Puts Intent First
Identity teams spend real time translating between what they want and what the canvas shows. With AI in the canvas, more of that work starts with intent. You describe the outcome, then review and refine.
That shift can make orchestration easier to share across a team, including people who didn't build the original flows. It also gives organizations a measured way to adopt AI, with people reviewing each step.
Identity stays a team effort. AI takes on more of the interpretation, and people keep making the decisions.