Recalls are costly for — and damaging to — any company, no matter the size or market.

For instance, McKinsey estimates that, for businesses manufacturing medical devices, recalls have been as high as $600 million in recent decades. The reputational impact tends to be lasting; customers aren’t quick to forgive. A Harris Interactive poll found that 55% of purchasers would switch brands following a recall, and that 21% would avoid buying any brand made by the manufacturer of the recalled product.

So what’s a business to do? Well, perhaps turn to AI, suggests Daniel First.

First is the CEO of Axion Ray, a company creating an AI-powered platform to predict product failures by taking in signals — from field service reports to sensor readings — and correlating these signals along with geolocation and other data.

It’s big business.

Axion Ray, valued at $100 million, today announced that it raised $17.5 million in a Series A round led by Bessemer Venture Partners with participation from RTX Ventures, Amplo and Inspired Capital. The new tranche brings New Castle, Delaware-based Axion’s total raised to $25 million, which First says will be put toward expanding the platform’s capabilities, entering new industries and growing Axion’s workforce.

The idea for Axion came to First while he was working at McKinsey, he says, in their AI strategy division. There, he saw that AI-powered projects to prevent product issues would often fail because the AI wasn’t sufficiently fine-tuned.

“To be successful, AI solutions that proactively mitigate issues need to be layered within a product, with workflows that different groups can use to collaborate to solve problems, enabled by a scalable AI platform with high precision,” First said. “Without [the right solution], many different groups across the enterprise do siloed analyses about emerging quality issues. This creates duplication and lack of collaboration.”

First started Axion Ray in 2021 to not only provide a way to detect warning signs that a product might be failing, but to give the various teams at an organization — engineering, program, product, production, field quality and customer support — a unified view of the issues and any data associated with them.

“Product quality issues can have an impact on the end user if [the] issues aren’t addressed quickly and efficiently,” First told TechCrunch in an interview. “Manufacturers struggle to proactively manage emerging issues affecting their customers, because field quality teams spend countless hours manually analyzing messy data sources to understand potential emerging problems.”

That, First says, is where Axion Ray can help.

He gives the example of a particular car model’s anti-lock braking system malfunctioning. Axion Ray’s algorithms might initially detect the issue from mechanic field reports, then identify the same or similar issues across call center complaints, reports from car dealership visits and car telemetry readings.

“We use a specialized AI to scan messy, unstructured and disconnected data across various systems to flag emerging recurring product quality issues,” First explained. “We can help a manufacturer understand that updating the hardware and software on a camera, for example, resulted in a spike in certain error codes, telematics aberrations, calls to the call center and returned parts.”

Now that’s a lot of data Axion’s ingesting — and for good reason, First would argue. But how’s Axion handling this= from a privacy perspective?

Axion says that it normally retains data “for the duration of an active account” or as outlined in a customer’s contractual agreement. Product owners concerned about how long data’s being kept might find that nebulous policy worrisome. First asserted, however, that Axion will delete customer data within 30 days of receiving a request.

“We’re committed to responsibly handling customer data,” he added.

With a team of 70 employees and customers in healthcare, consumer electronics, aeronautics, automotive and industrial equipment, including Boeing and Denso, First said he’s feeling confident in Axion’s growth trajectory.

“There are multiple trends that have supported Axion Ray’s expansion,” First said. “Many industries are releasing new technologies — like electric vehicles or other software-rich products — that are introducing unforeseen issues. Manufacturers are also working with new suppliers they have never worked with before. This is resulting in more quality issues than ever. Finally, manufacturers want to upskill their workforce to benefit from AI in driving automation of more manual tasks.”

Added Bessemer Venture Partners’ Kent Bennett via email: “Axion Ray has emerged as a clear market leader in automating workflows for field engineers to identify quality problems faster. The excitement we’ve heard from customers about Axion tells us the company is delivering clear and massive impact. The ROI their AI command center delivers to improve uptime, customer satisfaction and reduce cost has been a catalyst for significant growth within the customer base.”

Recalls are costly for — and damaging to — any company, no matter the size or market.

For instance, McKinsey estimates that, for businesses manufacturing medical devices, recalls have been as high as $600 million in recent decades. The reputational impact tends to be lasting; customers aren’t quick to forgive. A Harris Interactive poll found that 55% of purchasers would switch brands following a recall, and that 21% would avoid buying any brand made by the manufacturer of the recalled product.

So what’s a business to do? Well, perhaps turn to AI, suggests Daniel First.

First is the CEO of Axion Ray, a company creating an AI-powered platform to predict product failures by taking in signals — from field service reports to sensor readings — and correlating these signals along with geolocation and other data.

It’s big business.

Axion Ray, valued at $100 million, today announced that it raised $17.5 million in a Series A round led by Bessemer Venture Partners with participation from RTX Ventures, Amplo and Inspired Capital. The new tranche brings New Castle, Delaware-based Axion’s total raised to $25 million, which First says will be put toward expanding the platform’s capabilities, entering new industries and growing Axion’s workforce.

The idea for Axion came to First while he was working at McKinsey, he says, in their AI strategy division. There, he saw that AI-powered projects to prevent product issues would often fail because the AI wasn’t sufficiently fine-tuned.

“To be successful, AI solutions that proactively mitigate issues need to be layered within a product, with workflows that different groups can use to collaborate to solve problems, enabled by a scalable AI platform with high precision,” First said. “Without [the right solution], many different groups across the enterprise do siloed analyses about emerging quality issues. This creates duplication and lack of collaboration.”

First started Axion Ray in 2021 to not only provide a way to detect warning signs that a product might be failing, but to give the various teams at an organization — engineering, program, product, production, field quality and customer support — a unified view of the issues and any data associated with them.

“Product quality issues can have an impact on the end user if [the] issues aren’t addressed quickly and efficiently,” First told TechCrunch in an interview. “Manufacturers struggle to proactively manage emerging issues affecting their customers, because field quality teams spend countless hours manually analyzing messy data sources to understand potential emerging problems.”

That, First says, is where Axion Ray can help.

He gives the example of a particular car model’s anti-lock braking system malfunctioning. Axion Ray’s algorithms might initially detect the issue from mechanic field reports, then identify the same or similar issues across call center complaints, reports from car dealership visits and car telemetry readings.

“We use a specialized AI to scan messy, unstructured and disconnected data across various systems to flag emerging recurring product quality issues,” First explained. “We can help a manufacturer understand that updating the hardware and software on a camera, for example, resulted in a spike in certain error codes, telematics aberrations, calls to the call center and returned parts.”

Now that’s a lot of data Axion’s ingesting — and for good reason, First would argue. But how’s Axion handling this= from a privacy perspective?

Axion says that it normally retains data “for the duration of an active account” or as outlined in a customer’s contractual agreement. Product owners concerned about how long data’s being kept might find that nebulous policy worrisome. First asserted, however, that Axion will delete customer data within 30 days of receiving a request.

“We’re committed to responsibly handling customer data,” he added.

With a team of 70 employees and customers in healthcare, consumer electronics, aeronautics, automotive and industrial equipment, including Boeing and Denso, First said he’s feeling confident in Axion’s growth trajectory.

“There are multiple trends that have supported Axion Ray’s expansion,” First said. “Many industries are releasing new technologies — like electric vehicles or other software-rich products — that are introducing unforeseen issues. Manufacturers are also working with new suppliers they have never worked with before. This is resulting in more quality issues than ever. Finally, manufacturers want to upskill their workforce to benefit from AI in driving automation of more manual tasks.”

Added Bessemer Venture Partners’ Kent Bennett via email: “Axion Ray has emerged as a clear market leader in automating workflows for field engineers to identify quality problems faster. The excitement we’ve heard from customers about Axion tells us the company is delivering clear and massive impact. The ROI their AI command center delivers to improve uptime, customer satisfaction and reduce cost has been a catalyst for significant growth within the customer base.”