Guest post: Introducing our Partner ScaleHub and their Collective Intelligence Solutions

‍Artificial intelligence is taking the world by storm, and that’s no surprise. It’s changing the game in multiple industries and our partner Hypatos showed just how intelligent a task like document processing can be. With a mix of OCR and machine learning, Hypatos’s customers have been able to digitize documents with speed and accuracy that blows an OCR-only solution away. But while, for example, a success rate of 98% is an amazing leap forward, it still isn’t totally automated.

There are any number of reasons why AI can’t, as a standalone solution, reach 100% automation for IDP. But you’ll be surprised by the one factor that can get it over that hurdle and make it completely automated: humans-in-the-loop. It sounds counterintuitive, but it's the reason why our partnership with Hypatos makes so much sense.

The Power of Collective Intelligence for Document Processing

Artificial intelligence excels at racing through the kinds of repetitive, tedious tasks that the human brain quickly bores of. And, while machine learning means that AI can get better at these tasks as it learns, it still takes human intelligence to set up and control the training of the AI, as well as fill  the inevitable gaps and account for the small amount of exceptions to AI’s otherwise excellent capabilities.

Many are already doing this in a linear way: documents take a first pass through AI and any exceptions it can’t handle are forwarded to humans. ScaleHub streamlines this process, blending Hyaptos’s artificial intelligence efficiently and securely with the human intelligence of our >2 million global crowd contributors, in line with one another. In this way, the solution becomes one of collective intelligence, with machines handling the heavy lifting and people picking up the slack. From your perspective, this looks like 100% automation: put a document in, have correct data come back. Yes, there may have been a human intervention, but you didn’t see it.

Here’s how it works: Any exception the AI can’t handle will be handed over to the ScaleHub Portal. To ensure data privacy, documents are broken down into small pieces - which we call “snippets” - so that the data it contains is stripped of context (think: a barely readable handwritten address will be cut into snippets for validation that contain eg. only the family name, only the postal code, etc..) This is one way that Scalehub maintains data privacy. And in case you might be concerned about your data being processed by people unvetted, untrained and far removed from your business: Rest assured, we’re able to help you find crowds that fit your function, even in cases of specialized knowledge for specific industries or use cases.

Incidentally, that kind of crowd customization is perfect for those of you whose business goals also include ESG initiatives. ScaleHub works with various impact sourcing partners and can facilitate a consciously-chosen crowd that addresses the S (social targets) in ESG. For those struggling with talent shortages, this kind of solution is a win-win scenario.

Meet us in Berlin this November

For those of you who want to learn more about ScaleHub’s solutions and what makes Hypatos and us perfect partners: Come see us at the Intelligent Automation Summit in Berlin, November 22-24. Be sure to come by booth 4,where we’re making our first joint appearance, we’re looking forward to answering any and all questions you have about our solutions. If you can’t make it to the summit, but still want to learn more about us, our partnership with Hypatos, or crowdsourcing as a managed service, you can reach out to us here.

‍