
September 2023 Releases
Copilot started as the place you asked questions. Over the last few months it has become the place where agents are improved, tested, and monitored. Here is what's new:

Start your agent build with existing files: upload an existing script, screenshots or CSVs from your business, and Copilot uses them to ground the build so your agent reflects how your team actually works.
Try prompts like:
Once Copilot gets to work, Canvas opens in a tab beside the chat. You can make manual edits, test the agent's logic and test audio without ever leaving the conversation.
Every change Copilot makes is saved as a new draft version, and it suggests what to do next: test it, publish it, or keep making adjustments.
.png)
Giving an AI the ability to edit a live agent is only useful if you trust it. So when you ask Copilot to make an edit, it now shows an in-line diff of the proposed change and waits for your approval before saving a new draft.
Try prompts like:
That means Copilot can make large, sweeping changes (rewrite the objection handling, restructure a prompt) while you still review every line that moved. You get the speed of delegation with the control of a code review.

Every deployment has a weird edge case: an eligibility rule that lives in a spreadsheet, a lookup against an internal system, a calculation nobody wants to put in a prompt. With Regal Functions, you describe what you need and Copilot confirms the requirements, writes the code, explains what it does, and gives you example inputs and outputs so you can check the behavior before anything goes live.
Try prompts like:
The code is hosted on Regal, and you can use it across AI Agents, IVRs and Journeys. No developer required, and no separate infrastructure to maintain.

Most voice AI agent tools stop at prompting. At Regal, orchestration is built into the platform. Journeys are where calls, texts and handoffs get sequenced and triggered, and the agent, the journey and the data they produce all live in one place.
Try prompts like:
Because Journeys are native, Copilot can build them the same way it builds agents. Describe the outcome you want, such as calling new leads, texting anyone who doesn't pick up, and handing interested leads to a person. Review what Copilot proposes, then refine it in Canvas.

When you add Custom AI analysis points to your agent, you can pick up on fuzzy signals like intent, objection reasons and, competitor mentions on every call.
Try prompts like:
Now Copilot can surface them in plain language and with interactive graphics. For example, ask how many leads were interested in a Medicare plan last week and get an answer, a data visualization, plus the links to the corresponding transcripts.
Copilot can build agents from your existing files, show an in-line diff before saving any edit, write and host custom code with Regal Functions, build Journeys, and answer questions about your call data with charts and transcript links.
No. When you ask for an edit, Copilot shows an in-line diff of the proposed change and waits for your approval before saving a new draft version. Change Proposal is live for Single-State Agents today.
Regal Functions lets you describe custom logic, such as an eligibility rule or a lookup against an internal system, and Copilot confirms the requirements, writes the code, explains it, and gives example inputs and outputs to check. The code is hosted on Regal and works across AI Agents, IVRs, and Journeys.
Yes. Describe the outcome you want, such as calling new leads, texting anyone who does not answer, and handing interested leads to a person. Copilot builds the Journey, you review it, and you refine it in Canvas.
With Custom AI analysis points on your agent, Copilot can surface intent, objection reasons, and competitor mentions in plain language with interactive graphics. Ask how many leads were interested in a Medicare plan last week and you get an answer, a data visualization, and links to the matching transcripts.
Ready to see Regal in action?
Book a personalized demo.



