Right time, right place

Editorial Type: Interview Date: 2021-04-01 Views: 2,722 Tags: Document, Channel, AI, Machine Learning, Recognition, Cloud, Aluma PDF Version:
George Harpur, founder of Aluma, has been working with intelligent document processing since long before most of us had even heard the phrase. He spoke to DM editor David Tyler about combining in-depth expertise with leading-edge cloud models

David Tyler: I'm sure many of our readers will know your name from previous companies you've worked with, including Neurascript and Focal Point Software as well as Kofax. How does Aluma represent the next stage in your story?

George Harpur: I've been working in document automation and intelligent document processing - with a specific focus on Machine Learning, as that's my background - since around 1997 when we originally set up Neurascript. When that business was acquired by Kofax in the mid 2000s, I worked for some time in what was then its Transformation Group. At that time the technology worked well, but within what we might call a relatively limited universe of documents. Historically we had been applying these systems mostly to forms, and of course the Internet changed the world of forms processing very dramatically!

Over time systems evolved to work with other types of documents of course, but very much limited to high end projects, with very high volume - invoices was probably the prime example at that time. I was left wondering why the technologies available weren't being applied more widely, as I thought we could be doing so much more. That was the thinking when we set up Focal Point Software in 2009: we realised that the key issues were ease of use and ease of adoption, and that is still central to our thinking now.

Our first priority then was to almost 'reinvent' the technology; we'd built several document classification solutions for example by this time, and seen what was good and what was bad. The bad part was that it generally required a lot of expertise to set these systems up, plus time spent on tuning and training. This could add up to months - and we believed we could cut those times down to days, or in some cases even hours.

This would require a whole new approach to the technology, of course, and even in 2009 we had recognised the cloud angle as being the future - the market certainly wasn't ready for that at that point, but we knew we needed to start from a clean sheet, building everything ready for the cloud, even though we were initially still selling to traditional on-premise users. This meant that what we were developing, when the time was right, would be instantly much more widely accessible as a cloud service.

DT: So in a way it's almost like your technologies were waiting for an environmental shift like the growth of cloud, in order for you to really take advantage of the wider market?

GH: The Focal Point offering had significant success, with some projects handling hundreds of millions of documents for customers in the USA and elsewhere. That success allowed us to build and mature the technology itself - always with the long term plan in mind to be able to package the offering in such a way as to make our technology much more accessible to the wider market such as SMEs.


"The traditional vendors have highly functional, but also heavyweight - in a number of ways: complexity of deployment, configuration and in many cases, cost - offerings, and then there is this whole new cloud-first generation at the other end of the spectrum… We see ourselves at the mid-point of that spectrum, and we occupy a different space than those at either end. We believe we combine the best of both worlds. We have a mature and proven technology, wide and deep market expertise, and can also offer all the benefits of the cloud service approach."
There are levels of types of documents that previously hadn't been seen as ripe for automation: not accessible at the right time or the right point in the process, or because the document volumes for a particular user is so low that it would have been hard to justify the budget for automation.

Once we felt the cloud technology was ready we decided - about a year ago now - the time was right to adopt and announce the Aluma brand as a new identity for our new-to-market cloud service offering.

DT: What about your go-to-market strategy - presumably Aluma is very focused on building partner relationships?

GH: We're looking at partners, but it's potentially a very broad set of types of partner - at the end of the day, we provide a service that hooks into an information management system of some kind, improving the end user experience. We're all about enabling and enhancing those existing systems.

We have a technology that we feel is very broadly applicable, so it makes sense for us to scale via partners. Among those partners we would include Business Process Outsourcers, for instance - we're not selling to people who are looking to process their own documents, rather to businesses who process documents on behalf of other people. In that sense there is a very clear line in terms of who we are selling to.

DT: So to sum up the Aluma proposition, it's about combining the intelligent document processing and machine learning skills and expertise you've built up over the years, with the ability to offer via the cloud, while also focusing on making the solution as easy as possible to use and to deploy.

GH: That's exactly the combination that we think makes the Aluma offering unique. There are a number of barriers to entry here, and different people are coming at it from different angles, but noone quite like us. Some legacy vendors are offering a 'cloud front-end' on existing technologies, which is fine, but likely to be somewhat limiting. Then there are 'pure AI' or 'cloud first' players who are making a lot of noise, but our view is that those vendors won't have the maturity and real expertise in working with document sets in a variety of often challenging circumstances. Where we feel we occupy a unique space in the market is in bringing those two approaches together.

The traditional vendors have highly functional, but also heavyweight - in a number of ways: complexity of deployment, configuration and in many cases, cost - offerings, and then there is this whole new cloud-first generation at the other end of the spectrum. This group of course also includes the likes of Google, Microsoft and Amazon, who all have offerings in this space too.

We see ourselves at the mid-point of that spectrum, and we occupy a different space than those at either end. We believe we combine the best of both worlds. We have a mature and proven technology, wide and deep market expertise, and can also offer all the benefits of the cloud service approach in terms of widespread adoption, subscription licensing models, and all those things that make that aspect attractive to users.

DT: Obviously it's very early days for Aluma, but how is the message going with your target market so far?

GH: The Aluma message and approach are going down really well with the market already, even before we've really launched to the wider world. People are understanding what makes us different - or indeed unique.

Yes, we have expertise and a long history in intelligent document processing, as well as AI and Machine Learning, but we're not necessarily pushing those as features; it's far more about how users can trust us with their document needs to deliver a solution that will work for them. Historically to get the most benefit from those kinds of technologies, a user had to in effect become their own expert in those systems - we're saying you no longer have to do that. Put it in our hands and we will make the technology work for you - we've already done all the hard work upfront, and we've packaged it so that it is super fast, and super easy to use.

More info: www.aluma.io