Automatic for the people
13 October 2017 | 0
Automation in business is not, as many people believe, about robots coming to take all our jobs. Yet there are certain industries where it will be a key development in the near future. Financial services, the legal profession and even journalism have been named as industries where automation will impact heavily.
“There is a huge spectrum of activities described under the term automation, but we tend to talk about it in the context of intelligent automation. The concept of analytics and artificial intelligence (AI) are breathing new life completely into the whole automation agenda,” said Paul Pollock, managing director of Accenture Technology Ireland.
“Historically it was about workflow or business process management systems which allowed us to move paper electronically, and then present screens to people at the same time as they’re carrying out related tasks. That was probably the first generation of this activity.”
But now, according to Pollock, the industry has moved on to the point that many companies are investing in the concept of robotic process automation, the ‘soft’ equivalent of the robotics associated with heavy industry.
“It’s about replicating the actual activity that a human being would do in the process of completing a task. We’re seeing a lot of that across financial services, communications and so on, particularly in the back office functions as people seek to automate what are sometimes termed ‘swivel chair’ activities,” he said.
“This is where people take information out of one system and then key it into another in order to complete an activity. This is creating the idea of a virtual workforce.”
This virtual workforce is typically deployed to carry out clearly defined, repetitive and essentially simple activities that do not require a lot of decision making.
“Some organisations are at the proof of concept stage, some are dipping their toes in the water and some are adopting as it is proven. The reason this is happening is because it’s more efficient—it shaves the amount of time it takes to do repetitive tasks. It’s cheaper but it’s also allowing companies to develop tailor-made customer experiences for people,” said Pollock.
“What you’re now getting is the combination of data being used to drive an experience for a customer. When you ring into a contact centre, they have access to a lot of information on you before they answer the phone, and they use that to drive the customer experience with you.”
At first glance, a law firm is not where you would typically expect to find cutting edge artificial intelligence and machine learning being deployed. Historically, the law has been slow to adopt digital technology and paper has ruled supreme.
However legal firms are like every other type of company in that they’re under pressure to do more with less according to Josh Hogan, partner and head of McCann FitzGerald’s financial services regulatory group.
“I joined McCann FitzGerald back in 2002, and I worked in another firm in the 1990s, and at that stage even e-mail was distrusted. Things have come a long way, light years in fact,” he said.
“In more recent times though, we’ve become increasingly focused on data, basically because our clients are looking for more for less. They want us to deliver more service for less cost, and we’ve had to find ways to do that. And one way of doing that is through technology.”
The first major inroads of technology into law happened probably five or six years ago in the area of discovery in litigation. When a trainee joined a large firm twenty years ago, discovery involved poring over boxes of paper, manually marking up on paper and typing out things of interest.
Technology has come a long way since then but at the same time, there has been an explosion in the amount of documents that are typically relevant to a legal case.
“We’re all generating more data as we go through life in general. The only way to handle the volume is through automation. We regularly use different software tools around discovery for that reason,” said Hogan.
High Court approved
In 2015 the Irish High Court approved for use a process known as technology assisted review (TAR) in a precedent setting case, Quinn vs IBRC, and for the first time machine learning was applied to a legal case. A data set consisting of 10 million documents was processed using TAR methods and within six weeks was reduced to a relevant set of about 11,200 documents that could then be manually processed.
“How do you go through 10 million documents? We’d all be retired by the time it was finished. We got a precedent in front of the court to allow machine learning technology to be used and the judge agreed it was a reasonable use of technology,” said Hogan.
This was the first such use of TAR in Europe, but the technology actually began to gain traction in the US following a 2011 decision by Judge Andrew Peck in the da Silva Moore v Publicis case.
In that case, the judge held that TAR was an appropriate method for making discovery if it was properly done, with appropriate transparency as to how the system is trained. England followed suit in 2015 and Australia has now also approved it.
“TAR is very particular to litigation, and it’s a type of machine learning that is specific to dealing with large volumes of documents. It’s done using a training algorithm-based system to predict relevance in the data sets. In the Quinn case, it was about removing duplicate documents, documents in foreign languages and so on, getting down initially to 700,000 or so, and then down to 11,000 that were relevant,” said Hogan.
“The big challenge was to show that we weren’t just cutting out documents to suit ourselves, and that the methodology being employed was effective enough to accurately pick out the relevant documents.”
Machine learning platform
Discovery is not the only legal process where machine learning and AI are used by McCann FitzGerald—it also uses Kira Systems’ machine learning platform for contract analysis. Kira claims its software was used in $100 billion worth of transactions in 2015 alone.
The company also recently announced it was working with AI provider Neota Logic to design and market applications to assist McCann FitzGerald’s clients with reducing the time and cost of legal work. The agreement, which marks the first time that Neota Logic has been deployed in Ireland, will provide McCann FitzGerald with access to Neota Logic’s technology as well as its training and services.
“What we use Kira for is to extract clauses from documents and put them into a Word or Excel format and then carry out due diligence much faster than if humans were doing it,” said Tom Connor, data investigations technical specialist with McCann FitzGerald.
Connor works with different practice groups within McCann FitzGerald, primarily in the area of artificial intelligence with machine learning and AI expert systems.
“We develop AI solutions for our clients and also for internal process efficiencies, such as intelligent document automation, and smart legal advisory applications. I also run the firm’s rule-based machine learning programme,” he said.
“We’ve invested in AI software platforms to enable us to deliver these solutions to clients, and it’s only been by investing in machine learning that, together with our cloud platform, that we can really do that properly.”
According to Denis Kennelly, general manager of IBM Hybrid Cloud solutions, there are three things businesses should know currently about business process automation.
“We’re entering a new wave in process automation. A combination of three forces are pushing a new wave of process automation across multiple industries: technological advancements, increased pressures from businesses, and growing awareness of automation best practices,” he said.
“While the spotlight often shines on areas like machine learning, other technologies such as big data and predictive analytics also play an important role. Cognitive intelligence’s ability to adapt to changes and respond to unstructured data can enable nothing short of a revolution.”
IBM said its business process automation allows companies to focus directly on their mission.
“Even companies in the same space can have vastly different missions, so the real benefit of automation is that by improving processes, it frees your team to work on what’s important specifically to your business,” said Kennelly.
“Automation needs a strategy. The innovation in automation technology has brought numerous new tools to the market— many that have the potential to reshape many industries. However without a clear strategy for making the most of these tools, a strategy that identifies he best tasks to automate and managing changes within your organisation, you may not be able to capture the full value of the opportunity. A good automation strategy can be as important as the tools themselves.”
A good example of this occurred in the British NHS, when its blood and transplant division needed to allocate more life-saving donor organs more quickly. The organisation turned to IBM to automate a cumbersome 96-step process, a decision which not only improved the system, but also freed the agency up to refocus on providing improved care.
“They recently reached a critical milestone, the first transplant of an organ—a heart—that was allocated according to a scheme developed using IBM Business Process Manager on Cloud for process optimisation, and IBM Operational Decision Manager on Cloud as the rules engine for the allocation scheme itself,” said Kennelly.
Intelligent automation is one of the most disruptive developments facing the tech sector.
“The market development is still nascent, but we will see exponential growth set in within the next 12 months. However, the market communications about and the perception of the progress of IA are blurred, because the usual stakeholders who tend to educate and influence the market are on the sidelines,” he said.
“In particular, the supply side is anxious about the potential negative impact on profits. There is a lack of clarity about the best commercial models for dealing with the disruptive threat. Thus, market communications are dominated by automation tool providers with tiny marketing budgets and a vested interest.”
Xerox is best known as a document management company, and its business process automation activities mainly revolve around helping companies manage the way they interact with paper documents.
“In many Irish business environments, there’s a need for efficiency in how documents and especially paper documents get handled,” said Eamon Hardy, head of professional services and technology for Xerox Ireland.
“We use workflow assessment services to understand document workflows across organisations and to help transform paper-intensive processes using automated document capture, process automation, secure storage and access from the critical systems.”
Hardy argues that in most Irish companies, there is lots of potential to improve efficiencies by reimagining how documents are dealt with.
“There are usually many opportunities to increase accuracy, reduce cost and speed up the processing of documents to and from disparate sources like email, web, paper, fax and content management systems,” he said.
Xerox has worked with companies in a wide variety of industries to help rethink how documents flow through each organisation, from banking to financial services, to human resources and more.
“With banks, we’ve worked to improve customer service through enabling faster responses to requests such as opening accounts and loan applications,” he said.
“For example, with one client we used our workflow automation services to feed sixty separate bank applications through this process, covering mortgages, loans and life assurance. That allowed us to just insert documents directly into the bank’s database, further improving service.”
Document-heavy processes such as opening new accounts or change-of-address requests were sped up from five working days to being dealt with in hours. This in turn had a knock-on effect of delivering faster and more responsive customer service.
“It allowed the bank to make good on its ‘24 hour promise’ commitment to customers. With mortgages and loans, speed of response is key in terms of how much business they’re going to win, and as banks automate, that’s going to be critical,” said Hardy.
Hardy points out that according to Forrester Research, the average human resources department spends more than 50% of their time processing employee information and responding to inquiries.
“Effectively managing content can help free up time for HR people so they can focus on what they do best – serving the needs of employees and employers. Organisations can better control information management costs, support internal compliance efforts, enhance the employee experience and streamline HR functions,” said Hardy.
The future of business process automation is likely secure but despite this, there are some things that will probably remain best done manually by human beings.
“Workflow automation enhances our lives by providing quicker access to the products and services that we use every day. It increases the speed and accuracy of document processing in healthcare, banking, government and retail and we all benefit from this technology,” said Hardy.
“And although all processes can benefit from improved accuracy, security and speed of workflow automation, I expect that there are some things, like the reassuring smile of a healthcare professional, that will be here for the foreseeable future.”