Lume Launches: Automate Data Mappings Using AI

Lume recently launched!

‍

^‍

^{"Ingest client data, normalize data from different sources, and build data pipelines, automatically."}

^{TL;DR — Lume helps teams map data between systems with AI. If your team spends time mapping data for customer data onboarding, data normalization between systems, or mapping for data integrations, they can help! Sign up for a free trial here!}‍‍‍

_‍Founded by Nicolas Machado, Robert Ross, and Nebyou Zewde

‍

The founders at Lume are on a mission to automate the painstakingly manual process of data mapping, after experiencing this frustration as engineers themselves. They have onboarded several customers who are growing and are excited to help more teams use AI in their data mapping workflows.
‍

🧨 Problem

The usual mapping process involves a labor-intensive cycle: analyzing data to determine what's relevant, selecting the appropriate properties, developing the mapping logic, and constantly updating mappers to accommodate schema changes in source or target systems. This process, they learned, takes days, weeks, or even months for most teams, and automating it has traditionally been borderline impossible due to unique differences in data.

✨ Solution

Lume uses AI to automatically map data between any start and end schema, whether it is customer data, external sources, or anything else. Lume provides this via an API and a no-code platform where you can generate mapping logic, review and edit it, and manage multiple data pipelines. Whether you want to onboard customer data, normalize data from multiple sources, or create auto-mapping UIs over Lume, Lume delivers. With automated transformations and data delivery, error and type checking, auto-maintenance with schema inference, and execution via an API or a no-code App for different use cases, teams can spend more time delivering their core value to customers instead of wrangling and manually mapping data.

🚀 Key Features

• ✨ Generate mapping logic in seconds: via their API or the Lume app, simply pass in a sample of your source data and a target format, and Lume will generate the mappings.
• 🛠️ Edit mapping logic: review and edit the mapping logic and mapped data as needed to ensure you have the desired outcome. Leverage tools ranging from natural language all the way to full code edits to edit the mappers quickly and easily.
• ✅ Deploy mappers: Use the mappers deterministically in your code via their API or the app, allowing you to reliably move data between schemas as you scale your data pipelines.
• 🤖 Auto-maintain mappers: Lume detects changes in your schemas, notifying you and allowing you to use AI to update the mapping logic, effectively automating your maintenance.
• 📊 Manage and organize mappers: use Lume’s dashboard to have visibility over all of your mappers and live data pipelines.
• 📑 Handle any format: Lume supports JSON, CSV, XLS, PDF, and other formats! You can move data through the API or the app.
• 🔮 Upcoming: adding custom logic with lambda functions, custom validation rules, Python SDK, transferring your existing mappings when migrating to Lume, and more.

⚙️ How it works
‍

• To create mappers: specify your target structure and provide a sample of source data. This can be done via the Lume API or the Lume Platform.

‍

• ✨ Lume’s AI system creates the mappings between any two schemas, ranging from simple 1-1 mappings, time-series aggregations, complex calculations, and ontology classifications.

• Use the generated mappings and mapping logic: Use the Lume API or the Lume Platform to upload new incoming data and retrieve the corresponding output mapped data and mapping logic. Your data has just been automatically mapped with AI!
‍

‍

• For creating robust data pipelines, review and edit the generated mapping logic before running production data. Once saved, you can confidently use these mappers as deterministic pipelines. This is helpful for your data integrations between systems.
‍

‍

• Alternatively, build on top of the Lume API to deploy your own auto-mapper UI to allow customers to self-onboard.

💡 Use Cases

Lume handles three core use cases:

• Onboarding Client Data
• Normalizing Data from Multiple Sources
• One to Many Integrations

Here are three customer success stories:

• For a growth-stage battery analytics company, mapping each customer’s sensor data went from 2 weeks to 1.5 hours with Lume.
• For a large HR analytics platform, Lume mapping accuracy reached 95% for their most recent batch of integrations, saving them weeks of onboarding time for mapping and classifying data from incoming systems.
• For an early-stage financial product startup, Lume is handling ~500 live data pipelines that are externally facing, constantly mapping and adapting to new incoming data.

All of these have the common theme of having to map data between unique schemas, where even discrepancies as minor as column name variations make this process time-consuming and near-impossible to automate. This gets even worse at scale. Clients previously were allocating engineers, customer success teams, or offshore labor to analyze incoming data, map the data, and route it to their new systems. This process used to take up to multiple months for some teams, costing significant time and money.

They serve multiple industries ranging from ecommerce, insurance, manufacturing, and financial products.

‍

Learn More

_{🌐 Visit www.lume.ai to learn more
‍
👥 Share this with anyone who works on data mapping for customer onboarding, data normalization, or moving data between any two systems. They support a variety of use cases!
‍
📝 Sign up for a free trial to get started by visiting their website or booking a call here. They’ll get you on a Slack and onboarded quickly.
‍
⭐️ Upvote and share this post!
‍
🎉 Follow Lume on LinkedIn and X
‍
‍💥 The Deal: Mention you saw Lume via Launch YC, and we will give you a 50% discount for your first 6 months. We also offer a free pilot.}