How Lula cut 30 hours a week of data work by automating data
Fabi is the genie that knows all of your data and you can ask it anything and it'll give it to you.
Summary
Lula is a platform that helps convenience retailers and quick-service restaurants (QSRs) bring their business online. By implementing Fabi.ai, Lula was able to reduce manual data analysis work by 30 hours per week, and set themselves up to scale their operations efficiently without compromising service quality.
Context: Helping brick-and-mortar retailers and QSRs go online
Lula Commerce's mission is to help convenience retailers and quick-service restaurants (QSRs) bring their business online to meet their customers where they are. They offer an all-in-one platform that makes it possible for companies such as Valero, Shell and Jacksons Food Store to quickly go online by connecting them directly to third-party marketplaces: Uber Eats, DoorDash, and Grubhub.
With nearly 1,000 stores and over 433,000 rotating items in their system, Lula's ability to quickly analyze and act on data is crucial for maintaining service quality and managing risk for their customers.
Challenge: Finding the needle in the haystack
Lula serves convenience stores and QSRs that sell thousands of items a day. This comes with all the challenges of selling a large quantity of SKUs across hundreds of stores: transactions disputes and counter-disputes, monitoring for illicit items by geo and order cancellation processing, just to name a few.
All the data they needed to process and handled these requests or issues were stored in their PostgreSQL database, but they faced a few challenges with their data operations:
- Limited bandwidth: With only a few team members capable of deep data analysis, routine requests created bottlenecks. As Adit noted: "Unfortunately, while we do have technical team members, they're busy juggling multiple projects. This created a gap between being able to get the data and then communicate it to the right people."
- Time-consuming manual processes: Team members would export data to CSV files and share them manually, a process that became increasingly unsustainable as the company grew.
- Abandoned analysis: Many potentially valuable data investigations were abandoned due to time constraints because of the heavy lift of generating insights and getting them to stakeholders in a format they could use. As Adit shared: "If it takes me five minutes, it's worth it. But if it takes me like 45 minutes to an hour, it's like I don't have the time to do that right now. So it would just not happen."
For Lula this wasn’t just an issue of questions not getting answered or taking weeks to get answered, it was also a matter of serving their customer and even reducing liabilities. Every missed counter-dispute or illicit item sold represented lost revenue to Lula and their customers and even the risk of legal liabilities. Prior to rolling out Fabi.ai, the system was not scalable and often meant hiring overseas support to help them inspect records, flag items, and share as weekly reports.
Solution: Automating reports and alerts with Fabi.ai
To address these challenges, Adit Gupta (CEO & co-founder) and Phil Tribe (Head of product) decided to integrate Fabi.ai with their existing tech stack, which included AWS databases, PostgreSQL, and various internal tools. Three core components immediately made a difference to their data operations:
- A fully integrated SQL, Python and AI environment: With a direct connection to their database, the team used AI to generate queries in a fraction of the time. They could then process results using AI-generated Python for advanced analysis techniques, including fuzzy-matching, to identify items and transactions requiring attention
- The native connection to Slack: Once data was processed and flagged items were identified, they quickly built a Slack integration to proactively push alerts to team channels
- Support for scheduled jobs: Using Fabi.ai's scheduling functionality, they automated reports to refresh every few hours and push alerts to Slack multiple times a day, rather than waiting for daily or weekly reports
A key "aha moment" came when the team realized how quickly they could validate business intuitions with data. Or, as Adit puts it: "I look at our sales data every single day. Before, if there was a bug or a change in our sales pacing, I’d have to sink hours into tracking down the cause with the team. Now, I’m able to build a quick week-over-week chart of sales in 20 minutes and we can all quickly course correct as needed.”
Results: Hundreds of hours saved and stronger data culture
Implementing Fabi.ai has made both a positive financial and cultural impact for the Lula Commerce team. Among the most impactful results are:
- The elimination of 30 hours per week of manual data work
- Savings of approximately $10,000-11,000 annually in human costs
- Reduced analysis delivery time, meaning insights get in people’s hands in minutes instead of days
- Rapid scaling by automating key workflows
- More empowered team members thanks to self-service access to critical data
- Improved anomaly detection and service quality
Future: Scaling strategy and data operations
There are three main takeaways from Lula’s data story with Fabi.ai (so far):
- AI can help write SQL and Python scripts to both accelerate time to insights and automate manual or repetitive tasks
- Using Python to push alerts to Slack and meeting the team where they work can help unblock them, boosting productivity
- Reducing inefficiencies and costs is not just a matter of better operational efficiency but also a matter of scaling and growing an organization
As Lula prepares to onboard another 700 stores, Fabi.ai will play a crucial role in their scaling strategy. The platform will help them maintain service quality while dramatically expanding data operations without proportionally increasing costs or data debt.
Ready to improve your data workflows? You can get started with Fabi.ai for free in less than five minutes and make complex analysis a breeze.