The email security vendor has developed a way to use GenAI efficiently and affordably enough to assess every email rather than just a portion of them, according to Inky CEO Dave Baggett.
Email security vendor Inky is rolling out new capabilities that stand apart by utilizing GenAI to analyze every email intended for a customer’s inbox, rather than just a portion of emails, Inky Co-founder and CEO Dave Baggett told CRN.
While using Large Language Models to scan all emails is cost- and resource-prohibitive for most security vendors, Inky has developed a way to use GenAI efficiently and affordably enough to do so, according to Baggett (pictured).
[Related: Inky CEO Dave Baggett On Catching QR Code Phishing — Before Other Email Security Tools]
“This means that it doesn’t matter how cleverly the attacker tries to word something — if a human would understand it, then Inky is going to understand it also,” he said in an interview.
“That’s useful because it lets us understand the intent of the mail — is it asking the person to do something? What kind of tone does it have? Is it about payment?” Baggett said. “This really does massively level-up the detection capabilities of the system.”
The capability on Inky’s MSP-focused behavioral email security platform is available now for users of its pro and advanced bundles without any additional charge, the company said.
The enhancement took more than a year to develop, as a result of needing to address the multiple challenges of using infrastructure-intensive LLMs to assess the massive volumes of email — a large portion of which is spam or phishing — that customers receive today.
“You have to find a way to distill the parts of the knowledge of a large model that you can run efficiently and affordably enough so that you can include it in a base [email security] system,” Baggett said. “It’s a really difficult engineering problem.”
The answer, Inky’s team found, was to implement a trimmed-down LLM along with making certain hardware adjustments to make it affordable, he said.
Now, unlike competing vendors that only use GenAI to scan a fraction of emails due to the expense, “we’ve done it efficiently enough so it can run on every single mail,” Baggett said.
Given how different this approach is from those of other vendors in the email security space, the capability should be extremely compelling for Inky’s MSP partners, he said, offering a “great way for them to demonstrate the value of Inky to their own end customers.”
Crucially, the enhanced system also aims to stand out when it comes to “explainability” — providing specific details to partners and customers about why certain emails were marked as phishing or spam, Baggett said.
Partner Perspective
At Rudick Innovation and Technology, a Dallas-based MSP and partner of Inky, the new GenAI-powered scanning capabilities should prove massively useful with customers, according to Co-founder and CTO Chris Wilson.
Over four years of partnership, Inky has always been highly proficient at preventing spam and phishing from reaching inboxes, Wilson said. This was underscored by a recent quarterly business review the MSP held with a customer, which is only receiving a quarter of the emails intended for them because Inky is preventing the other 75 percent that are spam, he said.
Now, with the introduction of the newest capabilities to Inky’s platform, “using GenAI is going to see that email at a different level,” Wilson said.
For instance, the MSP will be enabled to identify whether an email’s content includes signs of urgency or has a payment requirement — “and we can now stack those parameters together and use those to make better decisions on their email,” he said.
The bottom line is that “the less we ask our customers to do with their inbox, the more successful we’re going to be as an MSP,” Wilson said. “This is going to make it even easier for us to deliver that capability to them.”
Advancing AI For Security
For Inky, the debut of the new GenAI-powered email scanning capabilities ultimately aims to build on the company’s original pioneering work around implementing computer vision for scanning emails at near-human levels of accuracy, according to Baggett.
The launch seven years ago involved taking known AI techniques and adapting them for use in email security, he noted. At the time, “that was a huge step-function improvement in email detection,” Baggett said.
Now with the latest capability from Inky, “we’ve done that again with this newer generative AI technology,” he said.
