We thought it about time that we took a good look at Google Document AI; the household names foray into the world of Document Capture. Call it Cognitive Capture, Capture 2.0, or Intelligent Document Processing (IDP), this once sedate and glacially moving sector has taken off in the past few years at a rate that few, if any, could have predicted. Microsoft and Amazon have been active in the space, and so has Google. You can read our full report (Vendor Vignette) for free here, but in summary, it’s pretty good stuff.

Let’s be honest, when you have the resources, finances, computing power, and access to vast swathes of data that a firm like Google has, you would expect it to be pretty good. But if I am being honest, we approached our research with a bit of skepticism, mainly as some grand claims regarding the accuracy of Google OCR (Optical Character Recognition) have been made. Google’s OCR is very accurate, easy to access, and doesn’t cost much to use. So too, are its forms capture tools good? Again you can read more in the report, but the question this all raises is, is this an industry game-changer? A Goliath-like Google entering into any tech market usually represents a sea change of fortunes. For those old enough to remember, Google’s entry into the enterprise search market back in the day was disruptive. Several enterprise search vendors of the time described it as destructive. This time around, it’s fair to say that Google’s entry will be less destructive and more constructive. The reason being Google Document AI is a building block, not a solution. It’s meant to drop into one’s program as a service, so it is attractive to all the programmers and business analyst level coders developing a document capture app. Compared to, for example, a product from the likes of Kofax or Abbyy where one gets a boatload of functionality before and after the actual data capture piece. Furthermore, the cost per page is competitive to the Capture specialists. Google’s advantage here is that it’s so darn easy to budget the monthly fee because it’s entirely usage-based; there are no seat licenses or “volume licenses.” 

This being said, Google isn’t directly competing with the specialists, so comparisons to ABBYY, Kofax are problematic. What Google has done here is to commoditize the underlying data capture technology. And in time, it will force the capture and document processing vendors to add value upstream in the process. Whereas in the past, a good bit of the value of any document processing solution was extracting the data from an image. Google has made it possible for a new generation of cloud-native, highly automated IDP solutions to emerge, built on the Google AI platform instead of by the old pieces and parts approach (OCR engine, image viewers, converters, filters, templates, RegEx models, etc.). New players don’t have to develop the core data capture layer and instead place 100% focus on solving specific document process problems by industry and use case. This is a sea change in IDP product development. To put it another way, it’s not apples to apples. Instead, Google is now growing all the apples one could need, you don’t need to grow your own anymore, rather, now is the time to upskill and start producing ciders and apple crumbles!