Button Text
Home
arrow
Blog
arrow
Document Automation
Jun 30, 2022
Min Read

Next-Generation AI Document Automation Software Understands Any Document

Share on TwitterShare on Twitter
Share on TwitterShare on Twitter
Share on TwitterShare on Twitter
Share on TwitterShare on Twitter
super.AI
Chief AI for Everyone Officer
SUMMARY

Document automation software has been around for more than three decades. However, recent advances in AI are spawning new entrants in the document automation market that are rapidly transforming the industry. Companies spanning a wide variety of industries can now leverage AI document automation software to eliminate manual effort when processing even the most complex documents.

This blog will cover a brief history of document automation, including the evolution of the technology’s first generation over the past decade, and how second-generation document automation startups are beginning to disrupt the market with fresh approaches.

A brief history of document automation

To handle the growing volume of corporate mail, companies centralized their mailroom operations.  Over time, mailrooms started using scanners to digitize and distribute documents, and then optical character recognition (OCR) to convert them to text. Initial leaders in this space were printer companies like IBM (IBM DataCap), OpenText, Kofax, and ABBYY. These companies became market leaders by adding data extraction and workflow capabilities on top of OCR.  

Initial data extraction relied on templates. This approach worked well for structured documents, such as standardized forms (W2, etc.) where the position of each field is constant. Early extraction techniques worked partially for semi-structured documents such as invoices. However, each invoice vendor required its own unique template. Ultimately, this approach was not scalable and the solutions were hard to set up and use.

First-generation AI document automation software

To overcome the challenges of a template-driven approach for processing semi-structured and unstructured documents several companies entered the market with an AI-based approach. These new solutions offered a greatly simplified user interface, enabling business users to combine document automation with robotic process automation (RPA) to automate document-centric business processes. They also provided some flexibility, allowing users to select one of several OCRs to improve results for a given use case. They also provided out-of-the-box solutions for common use cases. However, first generation AI document automation software suffered from the following limitations:

  1. High touch: Solutions from companies like WorkFusion and Hyperscience perform well in varied application scenarios, but are difficult to set up and maintain. Others work well for a small set of use cases, but require extensive training and maintenance when applied to new use cases.
  2. Proprietary AI models: Solutions like IQ Bot from Automation Anywhere and the Rossum Platform have greatly simplified the user interface, but rely on a fixed set of OCRs and proprietary AI models for document processing. This limits both platform flexibility and output quality, as the best model for a given task can’t be easily tested and deployed.
  3. Limited human resource management: All major solutions provide a human-in-the-loop interface with limited (if any) capabilities to escalate tasks, satisfy service level agreements (SLAs), gamification to keep human workers engaged, or access to a curated set of crowd-sourced resources to streamline deployment.
  4. Not output guarantees: Most tools will provide confidence levels at the document or field level for extraction, leaving the job of determining output quality to customers. They also lack the ability to automatically allocate work between humans, AI, and software bots to ensure quality, cost, and speed goals are met. This means AI projects can easily fall short of expectations or become economically unviable.
  5. Documents only: First generation document AI automation software is limited to processing documents only, leaving enterprises to find other vendors for processing other unstructured data types, such as emails, images, and videos.

Second-generation AI document automation software

Second-generation AI document automation software from companies like [super.AI](http://super.AI) are taking advantage of advances in the availability and capability of machine learning models with a fresh approach to AI-based document automation. Compared with first-generation AI document automation software, second-generation solutions offer:

  1. Low touch setup: Second generation AI document automation solutions streamline setup, model selection, training, ongoing maintenance, as well as creating and deploying new AI workers to continuously increase automation rates. Users can select prebuilt solutions from a marketplace, or use a no-code user interface to build their own applications from any combination of AI, human, or software workers.
  2. AI model agnostic: AI models are constantly evolving and improving, quickly becoming commodified. Rather than investing in proprietary models in an attempt to compete with technology giants like Facebook and Amazon, newer document automation solutions allow users to leverage any AL/ML model, or combination of models, to ensure the highest quality results at all times.
  3. Comprehensive human resource management: Humans are critical for the success of AI-based document automation.  But human resource management is often an afterthought. Second generation solutions offer access to crowdsourced workers adept at training AI models for deployment and validating results in production. Features like gamification are used to keep workers engaged, and sophisticated escalation rules to make sure a given validation task is completed in time to meet SLAs.
  4. Outcome guarantee: Newer solutions have moved beyond offering just a confidence level for the output of their AI solutions. They allow users to define the trade-offs between quality, cost, and speed, then automatically allocate resources between AI, humans, and bots to guarantee the outcome.
  5. Any data type: Second generation AI document automation solutions are built on platforms that can process any unstructured data type, including documents, emails, images, video, audio, and more. This gives users a one-stop-shop for unstructured data processing, reducing the need to purchase point solutions as business needs evolve.

Additional AI document automation resources

Great strides have been made in AI-based document automation software in the last decade, and the pace of innovation and quality of automation is only accelerating. For more information on automating document processing with AI, check out the following resources:

Other Tags:
Document Automation
Share on TwitterShare on Twitter
Share on FacebookShare on Facebook
Share on GithubShare on Github
Share on LinkedinShare on Linkedin

You might also like

close