Alpheous·
Platform Overview

How It Works

Alpheous runs as a team of AI agents that work the way your firm already works: they watch, they draft, and they wait for you to approve. This page explains that operating loop once, plainly, so the rest of the docs can refer back to it. Whether you spend your day in Slack reviewing drafts or you run the agents from the operator console, the same three steps describe everything the product does.

The Operating Loop: Watch, Draft, Approve

Every agent follows the same pattern. It is the spine of the whole product.

  1. Watch. The agents keep a continuous eye on the things you care about: your prospect universe, the public regulatory record, your inbox, and your calendar. They do not wait to be asked. When a filing lands, a meeting gets booked, or an advisor relationship goes quiet, the right agent notices.
  2. Draft. When something needs attention, the agent does the work end to end. That means the research, the ranking, the writing, and the compliance check, all before anything reaches you. What you receive is finished work with the reasoning attached, not a half-formed prompt.
  3. Approve. The agent posts the draft and stops. Nothing goes out to an advisor, a wholesaler, an LP, or a prospect until a person says so. You approve it as drafted, edit the wording, ask for a redraft, or skip it entirely.

:::info Why this matters The watch, draft, approve loop is also the human-approval guarantee. The agents do the slow, repetitive preparation; you keep judgment and every outbound decision. No email sends, no note hits the CRM, and no brief reaches the team without a person approving it first. :::

One Morning Queue

You do not have to chase the agents across a dozen tools. The work they prepare overnight and through the day collects in one place: a single morning queue of cards waiting for your decision. Each card carries the draft, the context the agent used, and the action buttons. You move through the queue, approve what is right, edit what needs a tweak, and skip what does not apply.

Because the agents have already pulled the context from your CRM, your calendar, and the public record, the drafts arrive with the right facts in them. Your time goes to the decision, not the lookup.

You Act Where You Already Work

There is no new console to learn before you can use the product. Most review happens in Slack, in the channel you already use. You mention the product to make a request, and the answer comes back as a card you can act on without leaving the conversation. For a fuller end-to-end walkthrough of one request, see Your First Real Action.

Operators and admins have a second surface. From the dashboard you configure which agents are active, connect the tools they read from, and monitor what the agents are doing across the firm. The daily reviewer never has to touch that surface; the operator uses it to keep the whole team healthy.

Configured to Your Firm

The agents are not generic. They are set up around your firm: the funds and strategies you raise for, the territories your wholesalers cover, the LP relationships your account leads own, and the regulatory record that applies to you. That configuration is what lets a draft sound like it came from someone on your team rather than a template. When you connect a tool or change who covers a territory, the agents pick up the change and work from it.

Meet the Roster

The agents are grouped by the job they do for your firm. Each one runs the same watch, draft, approve loop inside its own area:

Every one of these pages assumes the loop on this page, so you only have to learn it once.