Ramp uses Hex to quickly analyze financial and product data to make faster, better decisions
Ramp embeds Hex dashboards where teams already work — Salesforce, Notion, and beyond — enabling analytics engineers to build ML-level data products like sentiment analysis and shrinking feedback loops from days to hours.
About Ramp
Ramp is a finance automation platform that helps businesses manage corporate cards, expenses, bill payments, and accounting in one place, saving companies time and money.
Outcome
- Feedback loops shrunk from 2 days to hours
- Junior analytics engineers building sentiment analysis previously requiring ML specialists
- Data products embedded where teams already work, scaling decisions org-wide
Scale achieved
- Embedded Hex dashboards in Salesforce, Notion, and other tools
- Analytics engineers building ML-level work (sentiment analysis) in SQL
- Customer Feedback 360 data app unifying Zendesk, surveys, Datadog, and more
Challenge: Influencing strategic decisions across the org
When Ian Macomber, Ramp’s Head of Data and Analytics Engineering, first arrived at Ramp, the company lacked a single source of truth. Even basic questions like, “who are our customers? and “which accounts are delinquent?” were hard to answer. As a first step, Ian guided the team to build data foundations — starting with customer definitions, multi-touch attribution models, and analysis that explained sales fluctuations.
But Ian’s ambitions were bigger: use data to explain why changes happen and show stakeholders what to do next. He coached his team of analytics engineers and data scientists who were embedded in product, engineering, and design pods: “Your success is tied to your pod’s goals — and it’s up to us to deliver the ‘so what’ and ensure action follows.”
To influence strategic decisions across the org, the data team brought on Hex.
Tracking AI performance with a little help from... AI
When Ramp launched a new AI auto-routing feature, the data team turned to Hex’s Notebook Agent to keep tabs on how it was performing — and to make sure those insights actually landed with the folks who needed them.
They started by using the Agent to brainstorm success metrics: what should “good” look like? From there, they ran zero-shot chain-of-thought prompting to test hypotheses, without ever leaving Hex. Finally, they pulled it all together into insights built specifically for product managers — clear, actionable, and easy to share. Watch the full story here.
Prototyping and iterating fast on data products
In traditional analytics engineering, a simple data question can bounce between teams for days. At Ramp, it could at times take two full days to deliver an answer; with Hex, those feedback loops shrink to hours. Product managers and data scientists now co-create inside the same Hex notebook — sketching ideas, refining metrics, and seeing results in real time.
"The number of iterations you can get in per day is extremely high," Ian says. That pace speeds learning and keeps teams from prematurely locking pipelines into production. "Hex lets us prove what works before we fully commit."
Ian's experience reinforces a key lesson: not everything has to be production-grade from the start. Getting to the right decision requires iteration and Hex enables that rapid cycle of refinement and feedback.
Unifying data across multiple sources: Sentiment analysis isn’t just for ML specialists anymore
Hex is helping the data team push new boundaries. “Our machine learning engineers are building the coolest stuff imaginable. What I am equally excited about, however, is what we can enable an junior analytics engineer to do,“ Ian said.
Work that once demanded a specialist ML engineer now lands with an analytics engineer operating right in Snowflake. As Ian puts it, “The person who built this wasn’t our most crafted machine-learning engineer. It was an analytics engineer expanding the surface area of what they could build.”
In Hex, an analytics engineer built a sentiment analysis that people can explore in a Customer Feedback 360 data app. This is possible by bringing together data from Zendesk, customer surveys, Datadog, and more into a sentiment-analysis pipeline written entirely in SQL inside Hex.
With Hex, the team unifies structured and unstructured data — emails, call transcripts, logs, even images — and then transforms those insights into data products, all without sacrificing governance or scale.
Embedding data where teams work
Ramp places Hex dashboards inside the tools where colleagues already live: Salesforce for sales reps, Notion for product teams, and more. Meeting people in their workflow scales decisions far better than screenshots in slide decks.
As Ian puts it, "Being able to scale decisions is not just about having a great data product, but putting it into places where people are actually looking and working every single day." This approach helped the data team scale decision-making across the organization.
Impact: Driving Ramp 1% better every day
Ramp's data team is no longer just reporting on the past, they're shaping what comes next. Instead of being a service-based org, the ROI of Ramp’s data team comes from the actual internal and customer-facing product they build. With Hex, they've built a culture where all data professionals move fast, work alongside decision-makers, and ship data products that live where business happens.
Read Ian’s tips to drive influence across the org.