See the agenda in the Events list on the sidebar.
Are you using a Workforce Management System for Support organizations?
We are currently looking at finding a more efficient way to route our incoming cases to our support members and would like to take several factors in consideration:
- Workload distribution
- Work hours
We are starting to look into our requirements and we are wondering who is using a workforce management system and/or which approach you have taken to achieve this, whether it is building an in-house solution or taking something off the shelf.
Hey Patrick. Used to use automation for tickets triage and routing; and overall staff utilization, but in-house development. What is your weekly workload like? Multiple channels?1
Our workload averages about 100 cases per week. All of our case volume (all but a few exceptions) are submitted via our customer community portal. We do capture some information up front that would allow us to route accordingly. Presently, we are using a Slackbot that uses a workload score to identify who the case should be assigned to. The agent can then claim or reassign the case via the Salckbot. It works rather well for now, but with growth, new Support channels on the way (Chat), on top of having more resources working different shifts, we are coming to the point where we need a solution. I have been researching options for some time now and haven't really found the best for use. In a previous life, we used Verint, but this would be overkill for our current and future needs.
Let me know if you need more information. Thanks for your response.1
We are using ServiceNow - currently moving from ITSM to their CSM module. Once we have done that we want to implement a similar automation you are talking about. ServiceNow also offers some AI which should help. At least ServiceNow support management told me that they are using the AI for triaging and the success rate is greater 90%. Looking forward on some other comments - very interesting topic!1
@Alexander Mundorff Do you know does ServiceNow define their success rate?
@Patrick Martin Overall a low volume. What level of activity do you expect once you turn on chat?0
@David Perrault How to define success rate is an interesting question as one can imagine several ways to measure this. One that comes to my mind is a request successfully resolved within the target, but this works better for support narrowly defined rather than success widely defined.
In a lot of these AIs success is defined in terms of the 'match' rather than the 'outcome.'0
@David Perrault it's hard to say right now for chat. Our goal is to have a chatbot that will have a positive impact to our self-service metrics, and have the conversation routed to the right agent via the same rules as an incoming case via the portal. That's why we would like a solution that can easily handle multiple channels as we are looking for a solution that will scale with our evolving business needs.
We are looking at building something in house, where we could use our own solution to identify who is the most skilled agent to take the case, but the man hours required to build this are quite significant, and my team, being composed of front line agents, doesn't have the bandwidth to build such a solution. That's why we are looking for existing solutions. We use Salesforce for our case management, and we did see several apps that can do this. We are just starting to build our requirements so that's why this conversation is helpful as it helps us see what others are using, whether it's an in-house or off the shelf solution.0
Based on your current workload, I am honestly not sure you are going to get much benefits from automating the routing. My experience with automation for workload distribution on much larger volume has been on and off, though clearly AI has improved significantly in past few years. It becomes a real requirement when you manage 100s of new tickets a day. Until then, in my experience, nothing beats the speed and effectiveness of a manual triage when you are supporting complex SW solution. To be effective automation also requires up-to-date information not just on current workload distribution but also complexity of the issues being worked on, up-to-date knowledge and expertise of the agents, etc.... A solid experienced "call catcher" can do triage for low volume Wirth a better success rate.
Success rate for triage/workload distribution for me is measured by if the agent who got the ticket, resolved it as well.1
Thanks for the insight David. I understand what you mean in regards to manual dispatching. My challenge is that I don't have the business need (or workload) for a manual dispatcher working on 3 shifts (day, evening and night) 5 days a week. That's why we moved to a Slackbot in the first place, which has been working great for us, but it has its limits. I understand that the dispatcher can perform other tasks during slower periods, but I don't see this happening over the 24/5 support window, hence why we need an automated solution.
Our goal to automate is so that we can easily identity the best resources for the incoming case through AI. I understand that a manual triage could be efficient, but I believe that there is definitely a tool that can help in making this process transparent and automated for us, even though our volumes are low.0
I would argue with that workload you don't need to invest in either a dedicated manual dispatcher or an automated tool. Currently, we rotate some key Support agents into that role on a weekly basis and we have some also shifts that are very lightly staffed. This is just based on my experience, and if anyone has been successful at deploying an AI tool for low volume in a cost effective way, then I would indeed also be interested to hear about it.0
Thanks David. The idea of rotating the responsibility makes sense. My concern here is that I think we would be taking a step back in comparison with what we have today. Our Slackbot will automatically identify who should take the next case, but the criteria are very limited. It's a low cost solution that has worked well for us so far. I'll definitely bring this insight back to my leadership team to evaluate our options.0