Our goal is to ensure your voice, tone, and personality is reflected in the auto responses. We do this by taking in previous conversations you provide to get a sense of how to best respond in your voice. If you provide conversations with actual customers that have a different tone (e.g., a favorite or loyal customer), then we will take that into account with that specific customer. If you choose to provide your voice, it will help better personalize the generated assistant persona.

Yes (eventually), with the provided conversations, you can create separate variations of your persona, each with unique ids. If you want a customer to receive that variation then in the Customer Context data set, you would change the field “personality_variation” from “default” to the id of the persona variation.

At the moment, we only support text responses. We plan on providing voice and video responses in the future.

When you create your persona you will be automatically creating conversation workflows. Each workflow has a set of unique key trigger phrases (so we know when to trigger it) and a goal statement that lays out the core objective. You can modify the trigger phrases and goal statement to best suit your customer success. For advanced use cases, you can also provide optional context params and conditional statements that can be tied to your context data sets, see the Workflow API for more information.

If a customer contacts you through a Scout9 authenticated channel, we try to decipher what type of conversation it is being initiated based on the conversation context types provided. If a conversation type with 90% accuracy is determined, an auto response will be generated. If we can’t determine the conversation type then no response will be generated and you will have to reach out to the customer manually.

If your virtual self cannot progress through the given conversation workflow within 1 customer response iteration, then it will disengage and notify you to manually address the customer. You can optionally have the your virtual self provide a disengagement response to buy you some time (e.g. “Let me check and get back to you”)

Once you manually reply to that customer, your virtual agent will stop responding to that conversation. You can also use the dashboard to stop auto responses without replying to the customer manually.

We use a set of proprietary dynamic persona context generators that are applied to the language models to guarantee a response that resembles you. We use a variety of language models from Google, OpenAI, and Microsoft. Each language model might have different ways of responding and achieving the desired goal based on the context we provide, for more complex cases you might need to configure different AI language models

We use a prosperity dynamic persona context generator to supply the proper context to guide the LLM to generate a response as close in your voice as possible. We then run post processing on those responses and make modifications to further translate the text in your words as best as we can.

We currently only support gmail and it requires you to authenticate your gmail to allow for read and write privileges. We will be providing support for Outlook as well as Scout9 provided emails your customers can email.

We currently only support Scout9 provided phone numbers that your customers can text. But we are planning to provide Android and iOS applications that would allow you to have your personal phone become an authenticated Scout9 channel for auto responses.

We currently do not support any OTT applications at this time, but there’s no reason we couldn’t support them in the future.

Yes, we let you select from a variety of popular large language models from our support list such as Google, Meta, Microsoft, and OpenAI model lines. You can also integrate your own model for more domain expert use cases by following the API documentation and meeting input requirements.

Yes (eventually), with the given context and existing conversations on our platform you can compile that into your own domain expert LLM.

Yes (eventually), we do charge hosting fees but we can host your LLM and provide API endpoints for other applications.

Yes, you can configure your account to a variety of webhooks, view our API for more information.

Yes, we have a Node.js, Python, and REST api that require a business API key you can get on your account dashboard.

No, if a customer reaches out to you in any channel, they will be treated as a create customer and their fields will be entered automatically. However, your virtual form will be asking for data such as their name which could be awkward if you already have an existing relationship with that customer. To avoid that situation, it would be best to upload your customer context beforehand.