Skip to content

About

We give teams the outcome, not the recording

Every meeting reaches real decisions and real commitments. Then it ends, and most of them evaporate before anyone writes them down. That gap, between what the team chose and what the team remembers, is the problem Sol exists to close.

01Our story

Built by people who watched good decisions disappear

Maya led product at a workplace collaboration company and watched the same pattern repeat: a real decision was reached in the meeting, the call ended, and nobody owned it. Arjun built document-understanding systems at scale and saw that language models were finally good enough to pull structure out of messy, natural speech. They started Sol to bring those two things together.

Meetings are not usually unproductive. They are messy. A real decision gets made somewhere in the conversation, then gets buried under discussion and follow-up chat. No one writes it down in plain language. The commitment quietly disappears because the record never existed.

Modern language models are good at exactly the work that fixes this. They can read a full conversation, separate what was decided from what was only discussed, and name who committed to what and when, in plain language. That reading is the engine of Sol, not a feature bolted onto a notes app.

Sol is not a transcription tool. It reads the transcript and gives the team what they actually need: the decisions that were made, the action items with a named owner and a due date, and a short summary. Then it tracks whether those actions get done across meetings.

What happens without Sol

  1. 01Meeting ends
  2. 02Decisions stay in one person's memory
  3. 03Action items have no owner or due date
  4. 04Nobody replays the recording
  5. 05The team relitigates the same things next week

Sol turns the meeting into decisions, owned actions, and a short summary before the next one starts.

02Values

How we build

Four principles that guide every decision, from the extraction logic to the product.

Clarity, not a recording

The point of Sol is what you leave the meeting with. Everything we build is measured against whether the team is clearer afterward.

Only what was said

Owners and dates come from the meeting, never from a guess. A commitment Sol cannot ground is one it will not invent.

Follow-through is the product

Extracting an action is half the job. Sol earns its keep by tracking whether it actually gets done, meeting after meeting.

Quiet with your data

Your transcripts produce your outcomes only. They never train a shared model, and you can delete them on request.

03Team

9 people building meeting clarity

Maya and Arjun started Sol and build it full time. The rest of the team brings engineering, product, and machine-learning expertise from environments where clear decisions and real follow-through actually matter.

ME

Maya Ellison

Co-founder & CEO

Maya led product for a workplace collaboration company through its first thousand teams, and watched good decisions evaporate the moment a call ended. She started Sol to give teams the outcome, not the recording. She builds Sol full time.

New York, NY

AR

Arjun Rao

Co-founder & CTO

Arjun built machine-learning systems for document understanding at scale. He owns Sol's extraction engine: how a transcript is read, how models are routed, and how owners and dates stay tied to what was actually said. He builds Sol full time.

New York, NY

GL

Grace Lin

Founding Engineer, Extraction

Grace turns messy conversation into clean decisions and owned actions, and keeps the model honest about what it can and cannot claim from a meeting.

New York, NY

MB

Marcus Bell

Founding Engineer, Integrations

Marcus connects Sol to Zoom, Google Meet, Teams, and the tools where work happens, so transcripts arrive and actions land without anyone copying and pasting.

Austin, TX

SM

Sofia Marin

Head of Product

Sofia shapes how outcomes read the minute a meeting ends: short, specific, and easy to trust or correct. Her rule is that a busy manager should agree with it at a glance.

New York, NY

NC

Nathan Cole

Applied ML, Evaluation

Nathan builds the evaluations Sol is measured against, using real meetings, so the engine keeps clearing the bar as models and meeting styles change.

Remote, US

04Company

Legal entity

Sol, Inc.

Incorporated

Delaware, United States

Founded

May 2023

Stage

Pre-Seed

Headquarters

180 Lafayette Street, Suite 5NNew York, NY 10013United States

Phone

+1 (212) 555-0147

Team size

9 people

Come build the future of meeting clarity

We are a small team working on a problem every growing company has. If helping teams leave meetings with clear decisions and owned actions sounds like a good use of your time, we want to hear from you.