Following the launch of Teleopti’s Artificial Intelligence (AI) initiative, infusing core WFM with machine learning to make it more efficient and adaptive, Jeremy Hamill-Keays looks at the reality of AI for WFM. What is AI really and what effect is it having on WFM software and processes?
There is a lot of news, even hype, around AI and machine learning. The long-term promise held by AI could lead to all manner of developments, from the cure for cancer to flying taxis. To many it can seem like magic, but AI is more than that, it is a real “thing” built on advanced algorithms. So what is the reality of AI, beyond the magic?
AI and machine learning are terms encompassing a very wide range of different mathematical and SW methods, and these are often used in combination.
Rules-based systems are popular in modern business and life; a good example is an auto landing system for airplanes which provide a “path” for the plane to follow. If the plane starts to deviate from the path, it corrects itself using a simple rule. In contact centers rules-based systems are also common, for example for handling agent absence requests by having a service level rule. However, with rules-based systems, changing rules often mean reconfiguring the whole system, but AI goes beyond this by introducing the ability to adapt freely to work towards the desired outcome. For instance, Teleopti’s new AI tool, which uses machine learning, automatically reprograms shift category rules nightly. It recalibrates the shift category sets based on the contact center’s current data, - avoiding manual reprogramming but aligning to the contact center’s most recent historical data and situation. The result being that a complicated process is simplified, user experience is cleaner, and there is higher overall accuracy of shift data.
There are lots of other sophisticated techniques, some based on statistical methods, others using software neural networks and so on. Rather than look at how these work, I would like to take a look at what their usage means for the future of contact center planning. In particular, how AI impacts Workforce Management (WFM) software and processes.
One thing I am fairly certain about, is that demand on contact centers will increase. The world is moving to an online economy, delivering convenience and amazing choice to consumers. This leads to an ever-increasing demand on, and complexity within, contact centers, and there is an obvious need to automate simpler processes to provide time and space for WFM planning teams to focus on effective management and scheduling. In addition, clarity and understanding must be brought to large multi-dimensional data sets – these data sets will only increase with the rising online demand on contact centers. At Teleopti we see that AI will have a big effect on the development and performance of the WFM software itself. This is both exciting and necessary when it comes to contact centers planning efficient customer meetings to fulfill such a growing digital demand.
There is a huge diversity and range of needs in each contact center. As an example, request approval for an hour off will be different depending on the employees’ contract: full-time fixed hours, full-time flexible hours, part time rotation etc. Just when we think we have got a grasp on one particular setup, everything changes and new absence rules are brought in. In essence, a modern-day WFM system must be capable of quickly meeting a wide variety of needs and be able to quickly adapt.
The nature of AI means that it doesn’t rely on fixed and pre-defined equations, so it is able to adapt and react to data that is continuously changing. AI trains on new data as it comes in. This also means that it creates solutions that are specific to each contact center – the AI is the same in the WFM software but the results are unique and different for contact center A, B and C and even different between customer service teams within the contact center.
Intelligent WFM software has existed for a long time using advanced statistical methods, with some conditions: the more historical data the better the results. By this, I mean that each contact center, even each skill, has unique data and attributes. AI allows each of these to be learnt and the corresponding results are improved, specific to that data/those attributes, e.g. for a skill. As time passes and more data is collected, not only does the system’s accuracy improve, but if there is an underlying change, an AI system will more easily adapt to that change – far more easily than statistical methods which have less flexibility to handle incoming change. This is where machine Learning and AI become blurred. True AI can adapt it’s own algorithms in each neuron to improve itself. Sounds scary, but neural networks are just complicated interlinked nodes (pieces of code) that can inter-work to reach the correct answer by trial and error with self-adjustment, eg IF x>2 then y = 5. More detail on how they work is beyond the scope of this article! In today’s dynamic world, consumer behavior changes month by month, so this ability to adapt is hugely attractive.
In addition, the more complex the situation, the harder it is to understand. Pivot tables can only take you so far. AI software can scale well while maintaining such combined accuracy and adaptability.
Contact centers hold a multitude of complex situations that need to be analyzed and balanced all at once, in combination. Humans are very good at this, and often don’t even think twice about it, bringing in experience and knowledge without consciously thinking of it. However, humans can only handle a limited number of tasks at one time. What we really need is an intelligent assistant that can do much of the analysis for us and make recommendations. AI can be very good at spotting trends that we may miss, for example Bob is always 30 minutes late back from lunch on the second Tuesday of each month. AI can spot such a pattern and help us schedule him a longer lunch break, as he would’ve come back later anyway, but then schedule him to work until later to keep to his contracted hours.
As mentioned in my intro, there are AI techniques which can greatly help us simplify user interfaces, and in doing so make us more efficient and productive. It does this by identifying the important factors and hiding any information not needed. A good example of this is Teleopti’s machine learning functionality, automating shift categorization rather than having manual shift category selection menus. The scope is huge and even encompasses complex voice recognition chat systems. We fully expect AI will soon be very noticeable in WFM solutions, making life easier for those using the software.
In the modern contact center it is recognized that techniques targeting employee satisfaction result in better efficiencies and happier customers. However, the demands on a modern contact center can make this a difficult challenge. With the best will in the world, resource planners are often too busy to handle employees’ requests. At worst this can lead to resentment along with a lack of flexibility when both the center and agents most need it.
Automation features, driven by AI, remove administrative, non-strategic work to change schedules and deal with agent requests for time off, break moves or shift trades. Not all items can be handled by AI, a request due to a bereavement for example should be handled by a manager, but the majority of requests along the line of “the sun is shining, can I head off early” are easily handled by AI systems, without any negative impact on service levels. This is especially useful where agents are scheduled for a minimum amount of time due to contracts and can result in better optimization of the center, not worse. AI also allows negotiation in the form of “take an hour off now, work extra where we have an issue next week”. For multi-skill, large contact centers this ability to spot these opportunities would be highly welcomed.
Rather than fearing the advent of AI and seeing it as a way to cut back on staff, it should be welcomed as a chance to work with more intelligent software, challenge WFM practices, support agents in the contact center, and expand business by improving customer service.
Teleopti sees AI as an exciting opportunity to enter a brave new world of WFM and customer service excellence, read more about Teleopti’s AI initiative and recent machine learning release here.
Publish Date: May 25, 2018 |
| 2.) |  SPEED & PLUS The SPEED solution solves for service level issues while cost optimizing the environment with automation. Provides an enhanced way to speed up & optimize invoking temporary agent skills configuration changes into the contact center environment. Speed allows you to schedule both future changes & temporary changes that auto-revert back to the original state when scheduled time expires. Speed features: *Automated / Scheduled Temporary Agent Skills Configuration Management *Immediate Temporary or Reoccurring Schedule Skills Configuration Changes *Easy to Use/Operations Administration Focused *Descriptive Monitoring Activity Dashboard *Detailed “End to End’ Audit Trail and Perfor... (read more) |
LOGIN WITH YOUR CONTACTCENTERWORLD.COM ACCOUNT
I am checking out all the amazing and daily updated content on ContactCenterWorld.com and networking with professionals worldwide
Send To Friends Post On My Wall