How can mature IDP players compete with the cash-rich AI startups?

The intelligent document processing (IDP) market has been flooded with start-ups over the past five years. By our estimation, at least fifty new IDP software companies sprung into existence since 2017, surfing the crest of the Google AI deep learning wave and breathing new life into the very staid legacy Capture market.

Without doubt, this movement has been great for our industry, bringing document capture from an obscure backroom operation into a strategic enterprise initiative. Some of these start-ups have raised eye-watering investments and subsequently flooded the zone with advertising, all but drowning out the classic IDP vendors. 

Classic as in “successfully solving document capture problems for 25 years or so”. At Deep Analysis, we don’t age-discriminate. We cover vendors ranging from start-ups to the classics, regardless of time in the market. What matters most to our client community is how these vendors use innovation to solve unstructured data business problems, and do it better than ever before. 

Recently we caught up with two of the classic IDP vendors: Parascript and KnowledgeLake. Each has taken a different innovation approach to stay relevant in the new age of IDP.

Parascript was started in 1996 and is best known as the IDP supplier to the U.S. Postal Service as well as to several state and county election boards. The company has earned a solid reputation, first with outstanding handwriting analysis at speed and scale, and later with highly reliable check processing for the largest processors. As handwriting is by far the most difficult text for computers to read, Parascript was a pioneer at using machine learning (ML) and AI algorithms.

Over the past five years, the company has quietly expanded the product line and now offers a full-stack IDP solution for any document type; structured, semi-structured, and unstructured. The underlying AI/ML technology is as good as any of the start-ups we’ve seen. For example, version 2.0 of their SignatureXpert AI uses advanced trajectory as image methods that raise the bar for signature verification efficiency and accuracy. We’re hoping this tech is deployed in time for the 2024 USA election cycle.

KnowledgeLake began life in 2001 as a document capture product for SharePoint and quickly became the preferred Microsoft partner. When Microsoft signaled its intentions to enter IDP with what eventually became known as SharePoint Syntex, KnowledgeLake embarked on a bet-the-company strategy to expand its offering beyond the narrow capture focus and into a sprawling, end-to-end document management and automation stack. Our first impression was this looks a lot like an ECM-lite (or DM plus?) offering.

Earlier this year, KnowledgeLake released its new cloud-native platform called Tahoe. As one of us is British and the other lives in the UK, we love irony. Tahoe was the codename for the very first SharePoint release in 2001.

But Microsoft doesn’t have a trademark on great Western place names. (Although Apple might – be careful out there!) KnowledgeLake’s Tahoe is a solid platform with serious AI under the hood and their first no-code RPA module. The team said the initial adoption rate is promising, with existing and new customers choosing Tahoe as a reasonable alternative to clunky stacks that try to integrate IDP, RPA, and DM from different vendors.

Understandably, both Parascript and KnowledgeLake are touting their AI accomplishments in an effort to play in the new game. But after researching and talking with many vendors from both the new and the old guard, the “deep analysis” is this: having great AI is no longer a competitive differentiator. You cannot “out-deep-learning” the deep learning guys!

Then how can a mature, mid-size classic player compete with the shiny new AI-first stories? We had a frank discussion with both companies about the marketing challenge face by all classic IDP companies. Their AI pitches are threatened with oblivion, overwhelmed by the capital-infused buzz created by the flashy start-ups with those huge sales and marketing budgets.

We think they should stop playing the AI tech game and instead lead with their best asset: the vast experience and proof of success in solving many document problems for diverse industries and use cases.

Parascript, KnowledgeLake, and other “Classics” have deep and extensive knowledge of document processing. They employ technical people with years of tenure who have seen just about every weird and whacky document problem and will find a solution. This can differentiate them from the AI start-ups, who tend to focus on the AI tech and a specific document use case.

Honestly, how many more “invoice data extraction” products does the world really need? Anyone who has ever tried to automate data classification and extraction from invoices knows the solution requires far more than just a technically advanced IDP piece. This is just one of many use cases where the Classics may still have the advantage.

In summary, the AI tech playing field is levelling up fast, and that means every IDP vendor will need to up their game to close more deals. The Newbies must execute a plan to evolve beyond pure, cool product stuff and one-trick pony solutions. For the Classics, rather than trying to out-geek the Newbies, it may work best to focus on something they already possess in spades, something that is highly valued by the customer: experience.