To learn more about Docugami, please check out this article in Geekwire.

© 2019 Docugami Inc.

The Team







It is time for all of us building software to take a hard look in the mirror.  

For years, we promised we would solve the world’s information management problems, but the reality is we have just created a whole new set of problems.

Today, 85% of business information is “dark data,” lost in a sea of disconnected documents, emails, Slack conversations, voice-to-text messages, and other digital flotsam and jetsam. 

As the digital transformation accelerates, it is harder to ensure quality, consistency, accountability, and regulatory compliance.  

It is not enough to shrug our shoulders and say “hey, we just make the tools… we are not responsible for how people use them.”  

This “document dysfunction” affects nearly every type of organization, from finance to health care to real estate to government and more, impacting millions of citizens, customers and companies. 

One badly-engineered document ended up costing a Maine dairy company $10 million in driver overtime costs. Another poorly written contract resulted created lengthy litigation and a nearly $1 million loss for a major Canadian telecom company. The <healthcare story***>. The <real estate story***>.  

The tech industry helped create these problems, so we need to help solve them.  

Right now, there are lots of smart people putting lots of time and energy into using artificial intelligence to solve problems that boggle the imagination, like mining asteroids or AI-enhanced humans.  

We think that is great.  

But we hope people across the industry will join us in figuring out how to use AI to solve the mundane information management problems that plague businesses large and small.  

If that sounds boring compared to human settlements on Mars, that is okay. We think “Boring AI” could be a pretty big deal. 

We are a document engineering company. Our entire reason for being is to build “Boring AI” that will help stop document dysfunction. 

And we know we can’t do it alone. We are issuing a call for others across the software industry to recognize this issue and join us in working to solve these problems, helping unlock billions of dollars in increased efficiency, improved compliance, and business insights for companies around the world. 

OK, how can the software industry solve these problems now, when smart people have been working on them for 50 years and they’ve only gotten worse?


We see four principles that can lead us to more effective solutions: 

First, we need to bring together multiple scientific domains in new and powerful ways – that means artificial intelligence in natural language processing but also in image recognition,  semi-structured information, even approaches from natural sciences like biology. Breaking down the walls and combining these disciplines will give our industry new insights and new tools to solve these very hard problems.  

Second, instead of “Big Data,” we need AI that understands “Small Data”– the unique sets of business documents distinctive to individual companies. There is a lot of “Small Data,” and it is all different. A lot of what people call Big Data artificial intelligence these days is really highly supervised, labor-intensive, prohibitively-expensive machine training that cannot meet the internal needs of an individual company. We need technology that can scale across any enterprise and any commercial segment, with only minimal learning and guidance. We need algorithms that are smart enough to figure out your specific documents without massive supervision.  

Third, past attempts to use AI to try to solve business data and document problems have failed because they focused on the wrong altitude – helping to complete words or sentences. That only helps people create even more disconnected content even more quickly. We need a completely different approach. We need to apply AI to the document as a whole. Algorithms needs to understand the structure and strategy behind a company’s business documents, not just the meaning of individual words and phrases. If we can create tools that can understand the unique priorities and approaches of an individual company, COOs will have powerful new ways to accelerate performance, monitor accountability, and ensure legal and regulatory compliance.  

And to be truly effective, we need solutions that do not disrupt existing workflows or require massive investments in staff training or armies of consultants reinventing processes. From the start, AI should enrich the tools and routines that frontline workers already use to get their work done. The past 50 years have proven that you cannot force employees to adapt to straitjacket templates, you have to provide solutions that fit into how they already work and make them more productive. The more users accept the AI’s help, the smarter and more helpful the AI should become. It’s a virtuous cycle.  

It is fashionable to say that “AI is going to take our jobs,” but we think that’s a superficial take. Companies that focus on AI to cut costs may do okay in the short run, but companies that use AI to empower their frontline workers and drive their strategic advantage will be the real winners.  

The future isn’t about AI making human beings obsolete. The future is about AI making human beings and companies more productive and effective.  

We don’t think that’s boring at all.  

And we look forward to working with others across the industry to make that future a reality.