Flatpay uses Kernel agents to research small merchants at massive scale

Flatpay uses Kernel agents to research small merchants at massive scale

100
 outbound reps
Series B
Point of Sale
HQ: 

Denmark

Founded: 

2022

Website: 
flatpay.com

Problem

Flatpay sells to small European brick-and-mortar merchants. Below a certain threshold, the merchants are too small for Flatpay’s solution.

Flatpay did not have an automated way to remove the merchants it couldn’t sell to from account lists, which was a major source of inefficiency.

Solution

Kernel allows Flatpay to combine structured data (address, years of operation, etc.) and unstructured data (website, reviews, pricing, etc.) of 100,000s of merchants to estimate total payment volume and disqualify merchants before reps waste time on them.

Impact

Flatpay's annual recurring revenue has grown to over $30M in two years, and it is set to hire 250+ outbound sales reps across Europe in the next 12 months.

The estimated impact on rep productivity is 15% with an estimated revenue impact of $1M.

Flatpay uses Kernel agents to research small merchants at massive scale

Our 100+ reps used to speak with merchants, only to realise late in the process they are too small. Kernel agents now do research and sizing estimation across 100,000s of accounts.

Jacob Poulsen

Head of Marketing Expansion, Flatpay

Problem

Data coverage and accuracy for small merchants is notoriously low. So the only way Flatpay could filter out merchants that were too small was by having reps call them.

With hundreds of reps and many ramping at all times, calling merchants that should have been disqualified has been a major source of rep inefficiency.

Solution

Flatpay finds a high volume of merchants from Google Maps, but this information alone provides limited insights about each merchant. Kernel solves this challenge by enabling Flatpay to research every merchant in depth, as though a dedicated SDR had unlimited time to gather information.

By feeding a fine-tuned reasoning model with structured and unstructured data sourced by Kernel agents, Flatpay is able to estimate total payments volume and pre-qualify merchants prior to reps calling them

Let’s take an example account: ACME Japanese Bar and Sushi

The relevant publicly available data is:

  • Full-service restaurant in both Sushi & Japanese Cuisine, Vietnamese Cuisine
  • Facilitates payments both physical and online
  • High foot traffic in an urban location
  • 278 Google reviews
  • Average price point range is €15-25

Feeding this data to the model allows Flatpay to estimate TPV and qualify or disqualify the account.

The data is synced via API to Flatpay’s dialer-system, allowing the Flatpay team to review, provide feedback, and help Kernel further train the model and research agents.

Impact

In a test for Germany, Kernel agents disqualified 22 % of accounts that would have otherwise been prospected by reps.

The error rate where Kernel agents incorrectly disqualified accounts was 1.2 %.

The estimated impact on rep productivity is 15 % with an estimated revenue impact of $1M.

Flatpay uses Kernel agents to research small merchants at massive scale

Flatpay uses Kernel agents to research small merchants at massive scale