CIO Influence Interview with Eduard Frank, CTO of Scanbot SDK

Eduard Frank, CTO of Scanbot SDK, discusses the latest IT and tech trends, the impact of machine learning on enterprise software, and more about machine learning and computer vision in enterprise applications in this catch-up.

—————

Hello Eduard, welcome to the CIO Influence Interview Series. Please share the key learnings from your tech leadership journey.

Thank you for having me! One of the most valuable lessons I’ve learned over the years is how important alignment is across all functional teams. We have a core team, multiple mobile teams, a web team, and a QA team. In total, our technical team consists of over 30 people. That makes coordination a big challenge. How well these teams collaborate and work toward the same goals, though, is a vital indicator of success. When the alignment is strong, everything else falls into place.

Also Read: CIO Influence Interview with Bryan Litchford, Vice President of Private Cloud at Rackspace Technology

Scanbot SDK is known for its machine learning-based SDKs. Can you highlight how these solutions are impacting enterprise customers globally?

At Scanbot SDK, we build barcode scanning, document scanning, and data extraction solutions for enterprises to integrate directly into their apps or websites. Our mission is to “make capturing data as easy, accurate, and fast as possible.” That means we have three main goals: ease of use, reliability, and speed.

We achieve these aims by combining traditional computer vision techniques with custom-trained machine learning models. This ensures our scanning engine is both accurate and fast. Additionally, the SDK comes with ready-to-use UI screens designed by our UI/UX experts. They make the tools easily accessible to all users, even those with little tech experience.

For enterprises, this means they don’t need to spend time developing or customizing their own scanners. Our solutions can be integrated within hours, allowing companies to quickly enhance their workflows with minimal effort.

Talk to us about the key trends driving innovation in the B2B tech landscape.

Automation is a key trend in the B2B tech landscape. AI is everywhere, although it’s often implemented more for the sake of marketing rather than creating actual value. But it does have real potential when applied to automatic data processing.

The challenge, however, is the availability of quality data. Manual entry is just too slow and unreliable. Our scanning software helps bridge that gap by digitizing paper documents in flatbed scanner quality on any smartphone. These high-quality scans are what OCR engines need to extract data accurately. The same goes for parsing structured data formats like driver’s licenses or checks—we turn analog data into digital form for further automatic processing.

With the increasing focus on data privacy and security, how does Scanbot ensure the protection of sensitive digital data in its barcode scanning and document scanning solutions for enterprise customers?

One of our key differentiators is offline scanning. Although we use machine learning to train our scan engine, the final product operates 100% offline, entirely on-device, and without any server connection. In fact, there’s no network code in our SDK at all.

Because the scanning happens on-device, there’s no risk of server-side data leaks, eliminating a range of security concerns for both us and our customers.

We do additionally offer encryption features that help our customers securely store and transfer scanned data to their own servers. However, all this doesn’t concern the scanning process itself.

According to you, what are the biggest opportunities and challenges for companies looking to integrate machine learning and computer vision into their enterprise applications?

The biggest challenge for companies integrating machine learning and computer vision is quality input data. That means automating the digitization of analog data is a huge opportunity. Our solutions save our customers enormous amounts of time and minimize errors. For instance, our MRZ, driver’s license, and check scanners accurately extract complex data like names, dates, numbers, and even ID photos within seconds. They also return that data as simple key-value pairs, which are easy to process automatically.

Also Read: CIO Influence Interview with Mark Whitehead, CEO and co-founder, NDay Security

A major question companies face when it comes to a solution like a barcode or document scanner is whether to build or buy one. Developing a machine learning-based scanner requires a large team—ML engineers to develop the core, mobile engineers to ensure compatibility with the chosen app frameworks, and QA engineers to maintain reliability across different devices and operating systems.

Also, it’s not just a “build vs. buy” decision, but a “build and maintain vs. buy” one. Every new device or software update requires constant testing and updates to ensure compatibility, which adds to the ongoing maintenance cost. As a matter of fact, we’ve never seen a company over the years that decided to build their own solution and stuck with it.

Integrating our ML and CV solutions into enterprise apps, on the other hand, isn’t a challenge at all. Well, I might be a little biased here. At a recent Android conference, though, someone integrated our SDK in just 1 minute and 2 seconds. That’s how quick and easy it is.

Before we wrap up, in your experience, what are the misconceptions and myths around the role of a modern CTO?

A common misconception about the role of a modern CTO is that it’s all about coding. In reality, it’s more about setting a technical direction that aligns with business goals and ensuring teams are working together effectively. Another myth is that CTOs have all the answers. The truth is, it’s about asking the right questions, learning continuously, and adapting as the tech landscape evolves.

[To share your insights with us as part of editorial or sponsored content, please write to psen@itechseries.com]