There is that old line about IT vendors using their own technology to run their businesses—depending on your perspective, they are either “drinking their own champagne” or “eating their own dog food.” One could make the same remarks about IT organizations within companies, who should probably employ key technologies on their own processes before asking internal user groups to consume them.
That scenario appears to be happening with automation technologies. I’ve heard recently about several IT functions that are trying out these technologies on themselves before embracing them for business processes outside of IT. This makes a lot of sense, because IT organizations can learn the strengths and weaknesses of the technology before undertaking business-wide rollouts. And as IT becomes a bigger component of many organizations’ businesses, the productivity improvements that automation technologies can fuel can make a big difference to the companies’ bottom lines. Since these technologies are pretty sophisticated, I think “champagne” is a more suitable term than “dog food.”
Of course, automation technologies bring fears of job loss. I believe that when an organization adopts these tools, it’s a bad idea to put the primary focus on eliminating human jobs. The successful examples I’ve found thus far all have a different primary focus. Many of them involve reallocating IT professionals to other roles. Some jobs may eventually be lost, but the loss is usually through attrition.
DIY Automation
A few of these examples, primarily in online Silicon Valley firms, involve IT groups building the automation technologies themselves. At Facebook, for example, the proliferation of customers (over 1.7 billion) and servers (no numbers are released, but it’s clearly in the hundreds of thousands) led the company to develop more automated approaches to server management. Jay Parikh, Facebook’s VP of Engineering, told my coauthor (for the book Only Humans Need Apply) in an interview last year that the company had several approaches to automating server management, including “Facebook Auto Remediation” to detect and fix basic hardware remediation tasks, automated startup of new server clusters, and break/fix server software automation. The most astounding fact at Facebook is the number 25,000—that is the number of servers for every network engineer. Most organizations manage a few hundred servers per engineer at best. Parikh says the goal is not to eliminate employees, but to free up “a smart person in tech” to “go think about and do things for the future”—which will in many cases mean creating more automation.
Examples of IT drinking its own automation champagne will probably be more common using vendor technologies. One example is IPSoft’s Amelia, which is intended to automate customer interactions. Amelia was developed primarily to handle interactions with external customers, but IPSoft has had a long history of IT automation software (with its IPCenter suite), and Amelia can also be used in IT service centers for internal customers. That’s the approach that one Swedish bank, SEB, took in its adoption of Amelia. The company, which had been an IPCenter customer for several years, plans to use Amelia for external customer service applications, but started with the IT service desk. In the first three weeks of the pilot, Amelia had over 4,000 conversations with 700 IT users and resolved the majority of the queries independently, allowing employees to get consistent and rapid IT support. I kiddingly asked some of the bank’s IT leaders whether the fact that the Amelia avatar appears to be Swedish (see photo; she is actually based on a Los Angeles college student) influenced her adoption; they agreed that there were many user comments on her Scandinavian appearance. SEB is establishing an overall center of competency for cognitive applications, of which Amelia will be one component. 
IT outsourcing and services firms are another potential vineyard of automation champagne for their clients. Several India-based outsourcing firms are creating such capabilities, but most have been slow to be introduced to the market. One exception is HCL Technologies’ DryICE, which is already being used at 200 client sites. DryICE is a broad platform that performs a variety of automation and IT process orchestration tasks, using a variety of technologies. It incorporates machine learning (for analysis of machine data), robotic process automation (for IT process orchestration), and natural language processing for communication with human users.
Streamlining IT Operations
Energy Futures Holdings (a collection of Texas-based utility companies) is one of the HCL clients using DryICE. The company’s goal, according to a presentation at the 2016 Outsourcing World Summit, is not simply to eliminate employees (or reduce the cost of outsourcing), but to deploy more resources away from “running IT” to innovation and value creation with IT. Using DryICE and other tools, the company has already reduced IT operations costs substantially, and has added to its innovation and cybersecurity budgets.
HCL’s primary focus is on IT process outsourcing and operations, so I don’t know that the company will move toward automation technologies in other non-IT domains. But I suspect that many of its clients will, and their experience with IT process automation will help with the process. The outsourcing industry faces an uncertain future with the growth of these automation technologies, but companies like HCL that are willing to risk some cannibalization of their existing services have, I believe, the greatest chance of surviving this disruption.
I am not necessarily an advocate of starting with IT processes as a company explores cognitive technologies. The greatest value to the business may well lie in areas outside of IT. But if you’re running an IT function or an IT-driven business, it makes perfect sense to jump onto the cognitive bandwagon at an early stage. You have nothing to lose but boring, repetitive tasks. A champagne toast to getting rid of those!
Tom Davenport is the President’s Distinguished Professor of Information Technology and Management at Babson College, a Fellow of the MIT Initiative on the Digital Economy, co-founder of the International Institute for Analytics, and an independent senior adviser to Deloitte Analytics.
This blog first appeared Nov. 10 on DataInformed here.