Ricardo Craft gets candid about how Conexiom is specifically architected to get data off a document with 100% accuracy.
Based in North Vancouver, British Columbia, Ricardo Craft is a true outdoorsman. When he’s not keeping up to date with machine learning (ML) and artificial intelligence (AI) technology, he spends his time hiking, skiing, kite surfing, and open-water swimming with his family.
When he is not outdoors, Ricardo is at the helm of the Conexiom platform. Ricardo serves as the executive vice president of product management at Conexiom, where he provides insights into the strategy and roadmap for Conexiom’s products.
I ultimately have responsibility for the product and the product direction.
Conexiom focuses on sales order automation, invoice automation, and other document types for manufacturers and distributors. So I think about the product today, but also what enhancements to make to the product to help us compete effectively, and more importantly, deliver more value for our customers.
First, we deliver massive value for our customers. I feel privileged in that you can look at our 5-star reviews on G2 and read multiple customers say that Conexiom made a substantial difference within their organization. Conexiom has tangible proof points. Close to 10 million orders per year flow through our platform, and the value of those orders being processed for our customers’ customers are worth hundreds of billions of dollars.
Second, it’s an interesting time to be running product at Conexiom. ML and AI technologies are evolving quickly, so I spend a fair amount of time thinking through how those technologies will deliver incredible performance for our customers.
“Conexiom has tangible proof points. Close to 10 million orders per year flow through our platform, and the value of those orders being processed for our customers’ customers are worth hundreds of billions of dollars.”
If you look at, say, 30 different purchase orders (POs) from different trading partners, every one of those documents is going to have a different header, footer, bill to, ship to, vendor type, and ship type that live in different places within a document. Also, there may be orders with 1 or 2 line items or upwards of 500, and within those line items, there may be 1 line of text or it may wrap into 2 or 3 lines of text.
The degree of variability is an extremely hard problem to solve. Trying to set up and teach an RPA tool to deal with that variability without a high degree of subject matter expertise is not only time-consuming, but the ongoing maintenance required wastes resources and creates a high degree of technical debt. At Conexiom, we learned how to solve these problems, and our technology is purpose-built for this use case.
Also with RPA, you typically use OCR to extract the data from a document. If that data extraction isn’t 100% accurate (which OCR is not 100% data accurate, as it tries to recognize characters from an image), you compound data errors on top of structural document errors. That’s why RPA and OCR implementations come with high maintenance costs and high levels of frustration for its users.
Lastly, even if you are successful in extracting the data from a document, you must ask, “Now, what do I do with it?”. The information on a PO from a trading partner may not align with how your ERP needs to create an order.
“Part of Conexiom’s secret sauce is the ability to answer, ‘Now that I have this information, what do I do with it?’. We have the ability to transform the data from the document and apply very complex business rules so that we can create touchless transactions. This is what differentiates what we do and, more importantly, delivers value to our customers.”
Conexiom deals with purchase-order processing in a completely different paradigm. We don’t take broad horizontal tools to solve a specific problem. We understand the manual sales order- and invoice-processing problem, and therefore, our automation platform is specifically architected to get data off a document with 100% accuracy, understand the relative position of that information, how to deal with it, and how to apply specific business rules.
First, we deliver on the promise of 100% data accuracy and touchless orders. I had multiple conversations with customers where they shared that they had employees spending 70% of their time key punching in orders and only 30% of their time doing value-added services. But once they implemented Conexiom, that ratio flipped. They stopped doing unnecessary and time-consuming manual work.
And during the initial scramble of the pandemic, many customers in healthcare or cleaning-supply companies, for example, had order volumes spike to unprecedented levels. But because Conexiom was in place, they could deal with the influx of orders without asking team members to work overtime or finding and training new hires.
Many clients told us that they wouldn’t have been able to help so many of their customers deal with the pandemic if they didn’t have Conexiom in place to handle increased order volume. It speaks to the value Conexiom can deliver to customers.
My wife would ask me to automate what she asks me to do, so I actually remember to do it with out her having to remind me. Honestly, that’s why we bought an Amazon Alexa device. For that… and Alexa is a great DJ.
Hear More Conexiom Voices:
Senior Manager, IT Operations Peter Pacuch Streamlines Processes to Maintain Customer Trust
Lead Software Architect Chris Pirvu Guides Conexiom’s “Purpose-Built” Technology
Senior Manager, SaaS Operations Andreas Freyvogel Ensures Conexiom Maintains High Quality of Service
Product Manager Paloma Scheck Leads Conexiom’s Apps Within Salesforce
Product Manager Conor Lees Gives Us Something to Look Forward To—The Partner Portal Release
VP of Software Development Will Anderson Explains ML and AI in a Way Non-Techies Will Understand
EVP of Product Management Ricardo Craft Ensures Conexiom Delivers 100% Data Accuracy
