With Axios Systems introducing Artificial Intelligence (AI) chatbot integration capabilities into its next release, we look at how chatbots can improve your overall IT Service Management (ITSM) efficiency and productivity. Intelligent CIO talks to Craig Whytock, Solutions Consultant, Axios Systems, about how you can improve your ITSM delivery with AI chatbots.
What is driving the demand for chatbots?
When I chat to my customers, they’re telling me that real-time engagement is key to delivering a greater experience for their businesses. They are looking for 24/7 availability, faster ticket resolution and if possible, an instant response. Adding more agents to the issue is not a practical solution at all, so we need to look at another option.
Only 32% of the companies surveyed by Axios Systems believe that spending on IT would increase in the next 12 months. So, most service desks are being asked to improve their performance, but with fewer resources.
What if, we can get you to a point where you can eliminate the simple issues and requests through self-service and full automation capabilities with a knowledge base that’s presented to your users right at the point of when they need it?
Gartner predicts by 2019, IT service desks using AI capabilities will free up to 30% of support capacity and by 2020, 55% of all large enterprises will have deployed at least one chatbot in production. When fully adopted and driven by senior management, the benefits of using AI capabilities for IT service desks will justify the time and effort you’ve put in – from 24/7 availability, faster ticket resolution, proactive engagement and cost optimisations.
Before we can even consider a chatbot, you definitely need to have a solid foundation in place first. I’m talking about a self-service portal that’s been adopted by the business over the last year or so, one that is familiar to everyone and already has high usage in terms of day-to-day enquiries and requests.
The Axios Systems self-service portal can be the first port of contact for your business users instead of them trying to contact you over the phone or email.
The first thing they would see is a range of business services they can pick from.
The portal can be tailored to meet the needs of a business and be expanded to include other departments like HR, facilities, properties and travel so it becomes your one-stop-shop for Enterprise Service Management.
You could also re-brand the whole portal to your corporate look and feel which always helps with user adoption.
There’s a search engine, so staff can search for knowledge FAQs to help them fix their own issues. assyst tracks the usage of these FAQs, so that over time you can improve the content based on what’s working and what’s not. Your staff can even feedback and rate each FAQ. Knowledge Managers can then search on the one star and two star rated FAQs and replace them.
As users navigate around (the service catalog), it feels like a shopping experience – I once heard Fife Council say to its business that assyst is your ‘Amazon for IT’ and its staff quickly adopted it across a range of shared services.
It certainly feels like a shopping experience as you can include shopping baskets if you want and FedEx like tracking progress bars for workflows in the Service Requests.
Live chat agents
Built into the self-service portal is the live chat agent. Just like any commercial websites, you can speak directly to an agent. The agent can ask more pertinent questions, capture better information up front and the system will automatically log an Incident in the background capturing the full discussion.
So the self-service portal, the service catalog (‘The Amazon for IT’) and the agent chat – this is the foundation you need for a successful chatbot.
Chatbot integration capability
As I mentioned before, the next release from Axios Systems will include a chatbot Integration capability.
In terms of your use cases for chatbots – the sky’s the limit, especially if you involve your DevOps team too.
- Allow your users to find the right service request in your catalog first time without searching around
- Link your users’ questions straight to your knowledge-based articles
- Provide step-by-step instructions on some of your simpler high-volume type tickets
- Automate the provisioning of passwords or access requests
Agents also need the right knowledge base to help them and they may not always know the right questions to ask to find the right information. For example, AI could be used to help identify seemingly unrelated Incidents for Problem Management and would suggest that the agent then opens a Problem record.
In assyst, a similar search engine will use the event information to proactively push incidents/problems/known errors, and changes so you don’t need to re-search the knowledge base. AI could vastly improve this capability.
AI could also start to forecast issues based on previous experience. That would be quite interesting as the engine could start to do an impact analysis before users are impacted and escalate this information to Service Desk Manager.
In assyst, we currently capture all your assets and configuration data together with all previous history of incidents, problems and changes/releases in the one database.
So, AI could be applied to this data to predict future Service-Level Agreement (SLA) performance and escalate when it thinks you as the Service Desk Manager need to allocate maybe extra resources to focus on a particular service before it becomes an issue to your business.
The benefits of metrics
The biggest gain from a chatbot is the metrics it can provide to you. You might be struggling to track these metrics right now from your current knowledge base usage.
For example, the benefits include:
- Coverage – the number of questions that the chatbot can answer correctly – you don’t tend to get that from a standard search engine unless there’s an option like in assyst such as ‘yes, this fixed my issue’.
- Accuracy – the usefulness of the answers the chatbot provides, so you can review the quality of the responses for each and every question asked to the chatbot. You get a report, so you quickly see questions that are not being addressed properly. Again, you don’t tend to get that with a standard search engine.
- Deflection – the number of contacts to the service desk the chatbot can handle without human intervention, one of your main justifications for doing this in the first place.
- Experience – The average customer satisfaction scores of this channel compared to your other channels. Let’s face it, employee satisfaction should be the primary KPI for all AI strategies.
Training the chatbot
One thing you must consider is the bandwidth of your teams and their availabilities to start to train the chatbot. This shouldn’t be underestimated.
The chatbot will only be as good as the quality of training you give it. So, when you research the different chatbots that are out there, it’s worth reviewing how much effort it’s going to take and ask for a customer reference.
Also, going forward, who in your organisation will be reviewing the chat Q&As to ensure you’re providing the best responses? In the first few weeks, that could be a part-time or full-time position for someone.
How do you go about defining your requirements?
So where do you start:
- Review your own resources and bandwidth to implement and maintain the chatbot
- Ask your customer base what their expectations would be
- Involve customer advocates at the earliest opportunity to help you define the expectations
- Research each chatbot to see which provider will suit your services and what API does it have
- If it’s possible, build the chatbot in line with your service catalog (Offering by Offering) and knowledge base (FAQ by FAQ)
If you’d like to learn how an AI Chatbot integration can transform your ITSM, you can watch my webinar free, visit hubs.ly/H0gyXkV0