By Chip VonBurg, Field CTO, Hyperscience
Intro
I have recently joined the Hyperscience team as the Field CTO and as I write my first blog for the company, I wanted to share my perspective on why I joined and began my journey with the company.
Although I am new to Hyperscience, I am not new to the Intelligent Document Processing (IDP) space. I have been in this industry for many years, working for many vendors, and I’ve seen it grow and evolve.
In reflecting on my journey in the IDP space, it’s clear how much has evolved since I first entered the market.
Initially, the challenge was straightforward – the question was, can we achieve full-page OCR (Optical Character Recognition) to accurately read documents from machines? Overcoming this hurdle unlocked a multitude of possibilities for automation and processing.
The next step in the evolution involved how to get meaningful data from those documents. This began with structured forms using zonal recognition (aka fixed forms). This was a significant advancement, but it had its limitations. These systems often struggled with slight variations in form layout and although they worked well for machine print, they were ok at best at handprint (but could not even touch real life handwriting).
As we progressed, semi-structured documents like invoices and bills of lading (BOLs) came into play. By applying rule based models, we gained the flexibility to process a wider array of documents than ever before. However, this required a dedicated team of well-trained specialists to create and maintain the rule sets, which added to the complexity of our operations.
Then came the promise of machine learning. What if we could teach machines to read documents? This revolutionary shift offered the potential to significantly reduce both the cost and time associated with building and maintaining document models. It simplified the process while simultaneously enhancing accuracy and increasing automation rates.
The need for good User Interfaces (UI)
However, even as machine learning advanced, a critical need arose for good user interfaces (UI). While machine learning is powerful, it can be daunting without a team of data scientists. Effective interfaces are essential for loading, labeling, training, and testing models. They enable users to visualize document groupings and identify where additional samples are needed to improve the training set.
Most mature IDP platforms have some sort of user interface for their training data, but the reality is most lack the depth and detail that is really needed to give users the insights needed to not only create models, but to keep them performing well over time.
But how do you measure accuracy?
Measuring the accuracy and correctness of these models became another challenge. Achieving good results during the training process is only part of the equation; continuous quality assurance and reporting are essential to ensure that systems perform well in real-world scenarios.
As a result, model management emerged as a vital area of focus. As models are deployed, they can degrade over time, necessitating a system for tracking performance and managing new variants that require additional samples.
How do you integrate into the business?
Behind every document is a process and because of that just reading the document often isn’t enough. There exists the needs for integrations, decisioning, and the ability to inject some process intelligence into your document processes.
Although some vendors allow for customization of processing steps, often a third party vendor is needed for true end to end automation.
Hyperscience: Empowering Success Through Innovation and Accuracy
I’m proud to say that Hyperscience has anticipated and addressed all of these challenges through their core platform, the Hyperscience Hypercell. Built with AI at the core and based on proprietary machine learning models, we’ve developed a system that empowers humans to teach the technology effectively. The platform’s workflows facilitate seamless process orchestration and integration with downstream enterprise systems while allowing you to utilize the power of GenAI and LLMs. Hyperscience has developed a platform that consistently delivers the accuracy and automation rates our customers require.
This isn’t just theoretical; our customers are experiencing tangible benefits. At a recent event, one customer from a Fortune 500 financial services firm remarked, “When I first started with Hyperscience, I thought the 99.5% accuracy and 98% automation rates were just marketing claims, but I’m here to tell you they are real.” Companies are more than ever striving for automation to drive profitability goals, and Hyperscience is making these ROI’s a reality.
Hypercell is designed so that you can securely deploy your way. No matter if that is SaaS, OnPrem or PrivateCloud on all the major platforms. The architecture allows you to deploy as you see fit.
The automation landscape has transformed significantly over the years, and Hyperscience stands at the forefront of this evolution. We are committed to helping our customers navigate these advancements and realize the full potential of their document processing efforts, surpassing the struggles faced by many competitors in the industry.
I feel a profound sense of excitement about being part of Hyperscience. Our commitment to innovation and our ability to meet the evolving needs of our customers truly sets us apart.
Looking ahead, the future for Hyperscience is bright. With the rapid pace of technological advancements, we are well-positioned to lead the way in intelligent automation. As we continue to refine our platform and push the boundaries of what’s possible, I am eager to see how we can empower our customers to achieve even greater efficiency and success. Together, we are not just adapting to change; we are shaping the future of document processing and creating lasting impacts across industries. The journey is just beginning, and I can’t wait to see where it takes us!
Join us for an exclusive webinar on ‘Seasons of Change: Future-Proof Your Business with Hyperscience’ on December 17, 2024, where we will discuss how AI-native hyperautomation is revolutionizing document processing, outperforming traditional legacy systems, and driving efficiency and innovation in today’s fast-paced business landscape. Click here to register: https://hyperscience.mc-staging2.net/webinars-events/future-proof-your-business/.
Chip VonBurg brings a 25-year career in intelligent automation to Hyperscience, including tenures at Autonomy, Filetek, and Digital Documents. VonBurg spent the last 12 years at ABBYY, most recently as Chief Customer Officer. Throughout his career, Chip has been focused on driving innovation that delivers tangible results and ROI, always acting as an advocate for customers.
