Impressions from ANTENNA 2020 AntWorks Annual Analyst Retreat

Enterprises don’t need to be educated about automation anymore. Most of them have, at the very least, run some proofs of concept or pilots and, barring a few, concluded that RPA is good, but it doesn’t deliver the bang that was promised during the sales cycle. When we really diagnose the issue, the biggest reason for this rather dampening conclusion lies in the unforeseen effort required to develop and maintain a steady stream of clean data. Consequently, several enterprises develop cold feet about extending the use cases beyond the initial projects.

At the just-concluded AntWorks analyst retreat, ANTENNA 2020, this data challenge was the focal theme and was introduced immediately after the business update. It is refreshing to see that the ‘ants colony’ leadership is fully aligned with the problem of unstructured data and are investing heavily in exciting technologies, such as Fractal Science based ML, and a top notch R&D team (headed by a true blue rocket scientist!) to bridge this gap. I have long held the view that AntWorks’ Cognitive Machine Reading (CMR) is a key differentiator for them as they have gone to the first principles to build it ground up rather than relying on existing engines and philosophies. And the upgrades showcased this time are exciting!

Most data input modes combine different types of data formats. We all know this – I was recently filling out an application form for an insurance policy, and it had printable (name, addresses, etc.) fields, check boxes, signature fields, handwritten text (medical condition details, etc.). From a data ingestion perspective this is a nightmarish quagmire. Most automation tools address this only partially, or at most rather inefficiently by running the same form through multiple specialized engines. However, AntWorks introduced the Hybrid Processing functionality which allows all these different formats to be parsed through at one go. Apart from just extracting the data, it allows for business rules to be applied and also does critical steps such as signature verification out of the box. With 75-80% levels of claimed accuracy while deciphering cursive writing (endorsed by their Insurance client in a case study during the event), this can become a great tool for enterprises looking to reduce exception handling scenarios. It also opens the field up for interesting per-use services such as check verification. Additionally, it already covers 17 languages with Arabic (in line with AntWorks’ Middle East focus) and Mandarin on the way. It would be great to follow the market response to these functional upgrades.

The other big theme is around complete automation rather than task automation. This has long been my gripe with the industry – if you keep the enterprise at the center of the universe, it makes absolute sense to bring together the various technologies to solve for full processes. However, there still seems to be an inordinate focus on identifying more and more tasks that can be automated than on enhancing process wide value. Thankfully, a number of tool providers have started exploring this. AntWorks’ ANTstein SQUARE, launched towards the early part of last year, is focused on just this. It brings together the newly introduced Process Discovery tools, hybrid data processing through CMR, and the QueenBOT RPA engine as a full stack Integrated Automation Platform (IAP) to solve for the end-to-end problem. Along the way, it relegates RPA to where it should be, just one of the tools used by enterprises to optimize their processes. This seems like a no brainer, but years of market noise and conditioning has made this approach difficult for enterprises to take. AntWorks showed demonstrations of a couple of their cloud-based SaaS offerings that marry the IAP engine with domain specific details (e.g. Accounts Payable as-a-Service, E-Mail agent as-a-Service, etc.) which might make it even easier for enterprises to adopt this approach.

On the business front, a new Chief Partnership Officer, with experience on both sides of the tool provider-implementation partner divide, is changing the ecosystem strategy to a more targeted one. As a philosophy AntWorks plans to work with a limited set of partners, classified into implementation, technology, niche and social impact partners, but with much more alignment on the right metrics and mutual success factors. What is even more interesting is some new types of partners they are looking at, for instance they plan to partner with hardware providers on the Business to Business side, including IoT focused players.