The CEO and founder of Klarna, a fintech company preparing for an IPO, recently took to X to explain the reasoning behind his decision to move away from Salesforce’s widely-used CRM system in favor of an in-house AI solution developed by his team about a year ago.
In his latest comments, Sebastian Siemiatkowski made it clear that he does not expect, nor does he believe other companies should, mimic his choice. “This isn’t the end for Salesforce; it could very well be the opposite,” he remarked.
The revelation that Klarna had created its bespoke AI system, based on OpenAI’s ChatGPT, which enabled the company to terminate its Salesforce contract, gained significant attention in September. Siemiatkowski discussed this during an investor day, noting that the initiative resulted in the reduction of approximately 700 full-time contract positions and an annual saving of nearly $40 million.
In response, Salesforce’s founder and CEO Mark Benioff expressed doubts about Klarna’s methods for managing customer data and ensuring compliance. “Suddenly, @Benioff was queried on stage about Klarna’s departure from Salesforce. I felt a wave of embarrassment,” Siemiatkowski shared.
With speculation growing that the company might go public next month—looking forward to the release of Klarna’s confidential financial records—Siemiatkowski is seeking to clarify the situation.
As a fintech operating in a stringently regulated environment, he aims to dispel any notions that Klarna is simply uploading all customer data to OpenAI. He indicated on Monday that the initiative involved consolidating data from various SaaS platforms, including Salesforce, into a proprietary tech stack developed in-house.
While Siemiatkowski didn’t specify where all this data was relocated, he did mention that Klarna is leveraging the Swedish company Neo4j and its graph database technology.
“So no, we haven’t swapped out SaaS for a large language model; storing CRM data in such a model would come with its own challenges. Instead, we have engineered an internal tech stack that incorporates Neo4j and other solutions to amalgamate data into knowledge,” he elaborated.
“We enabled our internal AI to utilize this knowledge, and with the assistance of @cursor_ai, we found we could swiftly implement new interfaces and interactions,” he continued.
This debate underscores a longstanding issue in enterprise software: the choice between building in-house or buying pre-existing solutions.
Siemiatkowski believes that most organizations are unlikely to embark on building their own advanced AI-driven software.
Nevertheless, he anticipates significant consolidation within the SaaS industry. “Will we see all companies following Klarna’s path? I don’t think so. In fact, it’s far more likely that we will witness a trend towards fewer SaaS providers consolidating the market and emulating our approach to offer similar solutions to others,” he concluded.
Compiled by Techarena.au.
Fanpage: TechArena.au
Watch more about AI – Artificial Intelligence
