Digital transformation relates to the change associated with the application of technology, but it isn't just about adopting new technologies. Digital transformation is also about how these new technologies affect the way we work, who we work with, and how we create value.
You probably already have some form of digital transformation going on in your organization. Process automation is one of the first places businesses explore change, but the key is starting small and then building on these initiatives. So where should you start?
Where to Start With Digital Transformation?
Before you jump in, it's best to assess where you are. A digital transformation roadmap like the one shown below can help. Depending on the maturity level of your organization, some processes may be significantly transformed already, and others not at all. Areas of your company, like operations (ops) may have already been working to optimize some of their processes, and in doing so may have introduced new technologies, like a bot, artificial intelligence (AI), machine learning or augmented reality (AR).
After identifying the processes that have already been digitized, look for a process and department that could provide a big win. This often requires collecting a lot of data. Find three candidate processes where you have access to large amounts of data, such as processes involved in your meetings: who goes to what meetings, when and what room did they use?
Since most companies have far too many meetings already, eliminating status meetings is a great candidate for digitization. By ending status meetings you can increase productivity, all it requires is a common project work area, where everyone on the project can not only see task progress whenever they want, but also look at issues and online discussions. Some issues may need higher bandwidth interaction, making those conversations more appropriate for meetings.
Related Article: A Multifaceted Approach to Digital Transformation
Introducing Chatbots to the Help Desk
Another area that's ripe for automation is the internal help desk. By introducing a chatbot to this area, businesses can reduce costs and increase the efficiency of the help desk without anyone losing their job. For example, the most frequently asked question for the Help Desk is about changing passwords and login. In some cases it can be up to one third of all queries. This routine process is an instance where a bot may be able to perform better than a person.
Cable company Charter Communications implemented an external chatbot to its customers and found it decreased the live chat volume by 86 percent. Users were able to reset their passwords 50 percent faster, freeing up the help desk team to answer more challenging questions.
Implementation of the bot was easy, what is challenging is having enough data so that you can see over time this was the most asked question, and that an inordinate amount of (people) time was spent on it. It was a very specific problem, which is where Bots excel, and by using the Bot you are automating or digitizing the process. There is also the challenge of dealing with those that work on the support desk that fear they will be replaced, so there is some change management work to be done there to explain that the Bot is relieving them of a bothersome (but necessary) task, so that they can deal with more challenging questions.
In this case collecting the right data, determining the pattern, identifying the opportunity for automation or digitization, finding the right tool or technology for the job, creating and testing the bot, and dealing with the other Help Desk people all played a part in the digital transformation.
Related Article: 4 Questions to Ask Before You Send in the Chatbots
Identify Critical Processes
Both of the examples above show critical processes that can drag a company to a halt if ignored. Not paying employees, not finding customers, not servicing customers, not developing new products — all of these would fall under critical processes. Your challenge is to find a group that owns one of these processes, who is angry or frustrated enough to change.
The next step is to look at these critical processes and determine which ones should get the highest priority. The next example I'll use is recruiters for human resources (HR), because if you don’t keep hiring talented people, you will soon be out of business. At a time when a data engineer two years out of college can command a $300,000 salary, you know the war for talent is white hot.
Related Article: Think Digital Renovation, Not Digital Transformation
How Identifying Areas for Automation Works in Practice
We surveyed everyone in the HR group of a 1000 person company. In one of the questions we asked them what they did specifically, and how much time they spent doing that task or service. We also asked them how important they rated that task or service. The company had 92 people working in HR, 26 of them working as recruiters. From that sample, we created four personas (recruiter, recruiting consultant, recruting support, recruiting manager) and assigned a persona to each task listed. We then identified the simplest tasks (mostly in recruiting support), which we had found case studies showing where bots had done the task quicker, better and at a lower cost.
We summed the estimated number of hours for all 26 people to get the total time spent on thse simpler tasks. We did some additional research to find out the average cost per task. We then took the sum of the hours for that task for all of the recruiters and came up with a current cost to do that task in this specific company (within a specific persona). We then looked at a number of case studies of bots doing that specific task and looked at how much it would cost to create the bot, how long it took the bot to do the task and the ROI. In many cases the ROI was quite high.
For example, in one case a recruiter had to interview 270 people to get three qualified people to the hiring manager. All applicants were asked to submit a video of themselves answering a specific set of questions. It took a total of nine hours for a recruiter to view the videos. It only took the bot 45 minutes to review the same videos, so in this example the savings or ROI was about 10:1.
We applied this methodology to all of the potential bot tasks to help determine the priority of which tasks to automate. Of the HR sample we collected data on, doing “mass recruiting” took a total of 300 hours across all recruiters. If the average cost per recruiter was $100 for a specific task, that means the company was currently spending $30,000 to do mass recruiting, and if the bot cost $5,000 to be built, that would be a 6:1 return (within the first year). If mass recruiting took up an average of 20 percent of the recruiters time, then by letting the bot do it, recruiters had an addition 20 percent of their time to devote to special hires, or jobs that were very hard to fill.
By taking a revenue-based approach based on survey data, we were able to create a roadmap of which processes — and when — to automate. This process obviously isn't limited to HR, but can apply to all aspects of your organization.