Our first Adrima update: what we're building

Imagine an advisor preparing for a client review. Since the last conversation, the client has sent a document and raised a new question. There are notes to revisit and portfolio information to check. Before the meeting can be useful, the advisor needs to work out how those pieces relate.

That is the kind of work we're building Adrima around. Our recent Resources articles have explored the advisor-client relationship. For our first company update, we want to show how our work on meeting intelligence, CRM features, and the client app connects to that relationship, starting with the preparation for this meeting.

Before the meeting, bring the client into focus

The advisor begins with the client's recent activity. What prompted the new question? Was something left open after the last conversation? A note in the client record may explain more than the portfolio figures alone.

Our CRM plugin work focuses on making that relationship information easier to use alongside portfolio context. A customer relationship management system holds the firm's client records. We're developing features that help advisors rank, filter, and enrich their view of the client book using connected CRM and portfolio information.¹

The aim is to connect the broader view of who needs attention with the detail required to prepare for an individual conversation. Adrima is designed to work alongside existing CRM, portfolio, planning, and document systems, with connection scope depending on the systems involved and their availability.²

For the meeting itself, we're working on preparation briefs that bring together client goals, recent activity, unanswered questions, and supporting evidence. This is the preparation side of our meeting intelligence work.¹

We want the advisor to have a focused starting point for the discussion, with enough supporting context to examine what matters. The brief should help them decide what to ask and what needs a closer look.

A new document adds something to the conversation

As the advisor prepares, they turn to the document the client sent. It may help answer the new question. It may also introduce a detail that changes what they need to discuss. The task is to understand what the document contributes to this particular review.

We're developing document upload so clients can submit files and requests through the app, bringing those inputs into the advisor's workflow.² Our focus extends to what happens after the file arrives: making relevant information easier to find when the advisor needs it.

This connects to our work on retrieval-augmented generation, or RAG. Retrieval finds relevant information for an AI system to use when composing an answer.³ We're working on that capability to support responses grounded in firm information and approved sources.

For an advisor reviewing the document, the source matters as much as the explanation. They need to be able to assess whether the material applies to the question and whether something important is missing.

A retrieved passage may be outdated or need context from another record. We're developing these workflows to support that review. Finding information gives the advisor something to examine; deciding how it applies remains part of their work.

The meeting ends, and the next steps need a record

During the review, the advisor and client work through the question. One point becomes clearer. Another needs more investigation. The client raises a possibility for the future, but they have not decided to act on it.

Afterward, those distinctions need to survive in the follow-up.

This is the next part of our meeting intelligence work. We're bringing transcript review, notes, tasks, and a draft client follow-up into a connected workflow.¹ The advisor can review what was discussed and determine what belongs in the response.

We're building around the difference between a conversation and an agreed action. A useful follow-up needs to distinguish a decision from an idea, and a completed discussion from a question that remains open.

Our goal is to put the conversation and proposed next steps together so the advisor can check the draft before sharing it. That reviewed record should also help with later work, including preparation for the next meeting.

Back at home, the client returns to the explanation

Later, the client wants to revisit a point from the conversation. They remember the broad explanation, but a detail is less clear. They also need to send the additional information the advisor requested.

We're building the client app for this part of the relationship. It is designed to give clients access to firm-approved information, including portfolio context, goals, documents, meeting summaries, and explanations. Clients can ask questions, make requests, and upload documents through the same experience.²

Our site also demonstrates portfolio questions, investment reports, and a dashboard for reviewing portfolio information.¹ These are ways we're approaching a broader goal: helping clients understand the information their firm shares with them between conversations.

Advisor control is part of the design. Advisors determine what is appropriate to share, and questions requiring professional judgment go to them.² We want the client to have a useful way to revisit the work and bring unresolved questions back to the advisor.

The next conversation should build on this one

When another review comes around, the starting point should include what happened here: the question the client raised, the information they supplied, and the follow-up the advisor reviewed.

That continuity is what connects our current work at Adrima. Meeting intelligence, CRM plugin features, document upload, retrieval, and the client app each support a different step. We're developing them together so information gathered at one point can help with the next.

This first update describes work in progress. Our focus is on the connection between advisor preparation, reviewed follow-through, and client understanding. The meeting is one way to see it. The ongoing relationship is what we're building for.

Footnotes

1. Adrima feature descriptions and demonstrations, accessed September 12, 2026. ↩ ↩² ↩³ ↩⁴

2. Adrima product FAQ, accessed September 12, 2026. ↩ ↩² ↩³ ↩⁴

3. Patrick Lewis and coauthors, Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks, 2020. General technical background only. ↩