How Sunlight Financial knows anything in seconds.
What used to take hours or days now takes seconds, without a Slack message to the data team.
The analyst bottleneck.
When the COO needed a quick answer, like yesterday's average FICO for top-tier borrowers, it still required a request to the analytics team and a wait. The analysts spent most of their time fielding ad-hoc requests instead of doing actual analysis.
Every data question went through the same loop: stakeholder asks, analyst pulls. The team spent more time answering requests than doing analysis.
SQL Server ran ETL. Tableau handled visualization. Salesforce held custom reports. SSRS emailed reports. Answering a question that touched two systems meant manual exports and spreadsheet stitching.
20+ Tableau workbooks weren’t always current. Business logic was locked inside stored procedures that were difficult to maintain.
Open email, scan a table, close it, open the next one. Six times. No drill-down, no historical context, no interactivity.
“I used to start every day with six tabs of reports. Now I start with answers.”
An AI agent everyone can use.
Matt McCoy doesn't ping his analytics team for routine questions anymore. He asks Fi, Definite's AI data agent, and gets an answer in seconds.
Fi knows how Sunlight’s segments are defined. It knows which date field to use for Home Improvement vs Solar conversions. The analytics team went from fielding ad-hoc requests all day to doing higher-value work.
- >“What was the average FICO by month for Home Improvement?”
- >“What was our conversion rate from approvals to fundings by quarter?”
- >“How did promo volume compare quarter over quarter?”
- >“Break down borrower demographics for last month.”
- >“Which installers had the highest conversion rate last quarter?”
“Instead of pinging our analytics team every time I need a simple answer, I just ask the question and get it. I can self-serve.”
“People can now have a direct relationship with their data. My team gets to focus on higher-value work.”
Non-technical people picked it up in a week.
Definite ran workshops on-site at Sunlight's Charlotte office: two company-wide sessions, then 1:1s with credit, operations, finance, and product teams.
The team requested recordings to re-watch. Within a week, multiple people across the org were building their own reports and analyses.
Curtis Lynch, Head of Operations, started debugging his own call center data with Fi within a week.
“Even as a non-technical person, I can follow the AI agent's reasoning and understand the data it pulls.”
“After one session, I felt confident enough to teach myself from there. A week later I was building things I would have asked the analytics team for.”
“Every time we hit an obstacle, it feels like a team environment. If something doesn't exist yet, it gets built.”
Four systems became one.
Definite replaced SQL Server, Tableau, Salesforce reporting, and SSRS. Ingestion, data lake, transformation, modeling, and analytics run in a single platform.
The segment logic that used to live in opaque stored procedures now sits in readable semantic models the whole team can query.
- ×SQL Server stored procedures for ETL
- ×Data scattered across Salesforce, SQL Server, and flat files
- ×Business logic difficult to maintain
- ×20+ Tableau workbooks, tedious to update
- ×6 morning email reports, no interactivity
- →Automated connectors with built-in scheduling
- →One scaleable, unified lakehouse
- →Semantic models anyone can read
- →Live views with drill-down
- →One source of truth, always current
“We sunset Tableau entirely. The transition was seamless.”
What changed, in numbers.
“Definite isn't just a reporting tool. It gives everyone across the company access to data they never had before.”
“When in doubt, just ask. Fi will figure it out.”