Introducing Keragon’s AI

Nucleus · AI Editor

Keragon's AI lets you create, update, and troubleshoot healthcare workflows using plain English — inside a HIPAA-compliant platform. Choose between Agent Mode (build immediately) or Plan Mode (the AI thinks, plans, and confirms with you before building anything). You stay in full control at every step.

Keragon AI Editor — chat panel and workflow canvas side by side

What can the AI do?

⚙️

Create workflows

Describe what you want in plain English. The AI wires up the trigger, actions, connectors, and logic automatically.

Example

"Create a patient in DrChrono when a Jotform intake is submitted."

✏️

Modify existing workflows

Ask the AI to add steps, change logic, or reconfigure a workflow you've already built — no manual rewiring needed.

Example

"Add a check to notify the care team if insurance is missing."

🧩

Plan complex workflows

Switch to Plan Mode for multi-step workflows. The AI thinks through the architecture, asks clarifying questions, and presents a full plan before building anything.

Example

"I need intake-to-EHR sync with insurance checks — plan this out first."

🔍

Fix run errors from chat

The AI reads your run logs directly and fixes failures on the spot — no copy-pasting, no jumping between screens.

Example

"What's the run error here? Fix it."

HIPAA-compliant by default. All AI interactions run inside Keragon's HIPAA-compliant environment. Zero data retention for model interactions. Nothing is published without your approval.


Agent Mode vs Plan Mode

The mode selector at the bottom of the chat input lets you choose how the AI works with you. Both modes are available to all users.

⚡ Agent Mode

Build immediately

  • The default mode — the AI starts building as soon as you describe your workflow
  • Best for straightforward automations where you know what you want
  • Fastest path from description to published workflow

"Sync new Healthie patients to GoHighLevel as contacts."

🧩 Plan Mode

Think first, build second

  • AI analyses the full workflow architecture before writing a single step
  • Asks clarifying questions to get your intent exactly right
  • Presents a complete implementation plan for your review, then builds on approval
  • Best for complex, multi-step workflows — or any time getting the details right matters more than speed

"Build a patient intake flow with insurance verification, EHR sync, and care team routing."

Agent-Plan-modes.gif
💡

Rule of thumb. Use Agent Mode when you can describe the workflow in one sentence. Use Plan Mode when you'd need a paragraph — or when a wrong assumption mid-build would mean starting over.


How Plan Mode works

1

Select Plan Mode

Click the mode selector at the bottom of the chat input and choose Plan.

2

Describe your workflow

No need to get every detail right upfront — describe the goal, the systems involved, and any key rules. The AI will fill in the gaps.

3

AI asks clarifying questions

Before touching anything, the AI asks what it needs to know — connector details, conditional logic, edge cases — to eliminate unknowns upfront.

4

Review the implementation plan

You receive a complete plan: triggers, steps, logic, mappings. Review it, request changes, or approve it to proceed.

5

AI switches to Agent Mode and builds

Once approved, the AI builds the full workflow automatically — on rails, start to finish. No mid-build corrections, no lost context.

plan-mode-questions.gif

Debug run failures from chat

Keragon AI can now read your workflow run logs directly from within the chat. When a workflow fails, you no longer need to copy-paste error messages or navigate between the Runs page and the workflow editor.

Try saying

  • "What's the run error here?"
  • "My workflow has been failing — why?"
  • "Fix the error from the last run."

The AI will pull your recent run logs, identify the cause, and apply the fix directly to the workflow. Works for data type mismatches, missing fields, connector errors, and more.

running-logs.gif
💡

Run log access works in Agent Mode. Open the workflow in chat and ask about any failure — the AI handles the rest. Run log metadata (error type, step, field mapping) is read to diagnose the issue. PHI in your patient records is not accessed.


Why use Keragon's AI?

Get value in minutes

First-time users build a production-ready workflow in under 10 minutes. No automation experience required.

🧠

Expert-level accuracy

Built on thousands of real healthcare workflows, with deep knowledge of EHRs, intake tools, and care communication platforms.

🛡️

Healthcare compliance

Runs on Keragon's HIPAA-compliant platform. Encryption, BAAs, and zero data retention for model interactions.

🔐 Full control — always. Review every step, edit configurations, switch to the visual builder, or revert changes at any time. Nothing is published without your approval.


How to get started 🚀

  1. Sign up or log in to Keragon
  2. Open a workflow — new or existing
  3. Choose your mode — click the mode selector at the bottom of the chat. Select Agent to build immediately, or Plan to think through the architecture first
  4. Describe what you want in plain English
  5. Let the magic begin ✨ — you'll have a working automation in minutes

FAQ

Is the AI available in my plan?

Yes. All users can access Keragon's AI-powered editor — Agent Mode, Plan Mode, and run log debugging — including during free trials.

Is the AI safe for PHI?

Yes. All workflows and AI interactions run inside Keragon's HIPAA-compliant environment with BAAs and zero data retention for model interactions. When the AI reads run logs to debug a failure, it accesses error metadata only — not the PHI in your patient records.

When should I use Plan Mode vs Agent Mode?

Use Agent Mode when you have a clear workflow in mind and want to move fast — it builds immediately. Use Plan Mode when you're dealing with something complex or multi-step, and you want the AI to ask the right questions before touching anything. If a wrong assumption mid-build would cost you time to fix, Plan Mode is the right choice.

Can I edit the workflow after the AI creates it?

Absolutely. You have full control via the visual workflow builder at all times. Every step is editable, and you can always revert to a previous version without losing anything.

Workflow version history panel
What if the AI makes a mistake?

You can ask it to fix errors automatically — it can now read your run logs directly and apply the fix from chat. Or manually adjust any step in the visual builder. Nothing is published without your approval, and you can always go back to a previous version.

Does the AI replace the workflow builder?

No — it enhances it. Many users start with AI to scaffold the workflow, then fine-tune using the visual builder. Both work together seamlessly.

What's next Build your first workflow Put the AI to work — create a real healthcare automation in minutes. Open Keragon
Was this article helpful?
5 out of 6 found this helpful

Articles in this section