Samaya

Designing the interface and the first agentic experience in finance

Company

Samaya AI

My role

Senior Product Designer

Clients

Morgan Stanley, a top-5 hedge fund

Scope

Product UI, agent UX, brand motion

Samaya builds expert AI agents for investment professionals. I designed the product end to end, from the first query box to agents that plan, research, and report on their own.

The problem

“Can you get me ATOSS revenue in local currency from 2019 to 2024?”

  • Who is asking?An analyst or an MD
  • When?Before earnings, or after
  • Why?A model, a pitch, a client note
  • Where do they work?Sell side or buy side
  • What's the bigger task?One number or a full initiation
  • How do they like data?Local currency, fiscal or calendar

The same question means different things depending on who asks, when, and why. The product had to understand all of it.

What great design does

  1. 01Knows the user
  2. 02Anticipates intent
  3. 03Adapts to context
  4. 04Removes friction
  5. 05Creates delight
  6. 06Builds trust
01

Agent Studio

Describe the report you need in plain language. An agent drafts the outline, takes your edits, runs the research, and puts itself on a schedule.

02

Plan before it runs

Before any research starts, the agent shows how it understood you: the companies it detected, the time range, and the sources it will read. Every assumption can be edited.

03

Reasoning you can watch

Explore, validate, synthesize, summarize: each stage of the agent's work is visible as it happens. Analysts always know where it is and why.

04

Evidence over confidence

Answers fill in cell by cell, each one cited. When the data isn't there, the table says no evidence instead of guessing.

05

Your sources, your rules

Filings, transcripts, broker research, the web, a team's own documents: analysts decide exactly what the agent reads, down to a single folder.

06

One place to start

Quick answers and deep agent work begin in the same input. Q&A and agents share one entry point. Stepping up from a question to a recurring report is easy.

07

Always on

Agents keep watching once the report ships. Triggers like a guidance change or a spread widening turn a one-time answer into ongoing coverage.

01 / 07

Brand motion

Making the invisible legible

Motion for explaining how Samaya's agents are built and where they read from

Agent architecture, from orchestration to summary
Every source, one agent

Impact

10,000+

Seats deployed at bulge-bracket banks and top asset managers

~40

Vendors in the competitive pilots a top-5 hedge fund ran before choosing Samaya

3

Morgan Stanley divisions using it: Research, Sales & Trading, and Banking

Company milestones, as published on samaya.ai

Next project

Amity →