Hey everyone 👋 We’re Anna and Tyler from OmniAI, and we help clean up data using AI.

🪄 Omni in action: Turn YC demo day videos into data

Here we started with a database of all the public demo day videos and extracted some of the most interesting business metrics — without needing to watch hundreds of videos.

Problem

A company’s most valuable data set is often its least accessible. Only 20% of corporate data lives in SQL, with the remaining 80% locked away in unstructured formats (reviews, chat logs, transcripts, pdfs). It takes some serious engineering work to answer questions like:

• “Graph the most common user complaints over the last 4 years”
• “What percentage of users complain about pricing on sales calls”
• “How many users are getting their questions answered by our chatbot”

And this problem will only get worse as AI interfaces become more ubiquitous and more user data is locked up in chat logs. 🤖

.

How are people solving this today?

Even YC founders don’t have a unified strategy here. From the hundreds of companies we’ve interviewed, it usually boils down to:

🌎 Outsourced data entry - Slow and surprisingly inaccurate

🔎 Spot checking - Slow and very prone to bias (i.e. read 10 user reviews and make sweeping business decisions)

💸 Hire some ML engineers - Best results, but it’s going to be slow and expensive

.

Solution

Use OmniAI to turn that unstructured data into tabular data. OmniAI makes it simple to transform and enhance unstructured data, all within your existing data warehouse.

• No engineering time needed
• No calibrating models
• And certainly no fighting with AWS SageMaker.

How it works

Run ML models across your data. Categorize, extract, summarize, translate, and more.

• Connect to a data warehouse - We support Snowflake, Postgres, MySQL, and MongoDB.
• Choose models to run - Define type safe schemas to run against your unstructured data. Use any of the hosted models, or define your own.
• Transform your data - We’ll run those models against your data, and keep your warehouse in sync as new rows/fields are added/deleted.
• Query with SQL - All the transformed data stays in your warehouse. Surface this data in your product, or analyze with your existing BI tools.

Meet Tyler & Anna

Anna comes from building ETLs and data enrichment tools at Hightouch. While Tyler has a background in ML and healthcare applications (most recently at Fair Square). We’ve dealt with everything from syncing CRMs to fax machine APIs.

We’re building Omni because we hate working with messy data, and we can make this problem a little better for everyone.

Our ask

• Data problems! We’re constantly adding new tools and would love to hear about the unique data problems you face.
• Book a demo with us at getomni.ai
• Tell people about Omni! Know anyone drowning in a sea of PDFs, or trying to build image classification from scratch? Chances are we can help!

Email us at founders@getomni.ai