What Your Employee Self-Help Chatbot Can’t Live Without
It’s relatively easy to know what your help desk agents can’t live without. They need the tools—both software and hardware—to do their jobs. They also need training to ensure they are able to answer questions and resolve issues effectively.
However, with an average attrition rate of 41% (according to HDI), it is very difficult to ensure that agents will get up to speed quickly on all of the technology required to optimize employee productivity. CIOs report that throwing more money at this problem is the wrong approach and are turning to automation instead.
Given the relative newness of the employee self-help chatbot space, however, most organizations are not quite sure what chatbots can’t live without. This is why we are seeing a number of RFPs that are literally all over the place. And, honestly, most of the ones that we see don’t have the key fundamentals covered.
This inspired me to write a blog on what organizations should be looking for in their automation solution:
- Intuitive, consumer-like interface
According to Conner Forrest, senior analyst at 451 Research, 91% of candidates say that the availability of productivity tools and services impacts their decision to take a job. Today’s consumer mentality around increasingly personalized service and instant gratification has bled into the workplace, which means employees expect an Alexa-like experience. So, your chatbot needs to deliver answers within 3 seconds that are personalized to their location, role, and equipment. And it needs to be easy to use or they won’t come back.
- Workflow-based experiences
The reason you are investing in an employee self-help chatbot is most likely related to budget reduction, ticket deflection, and/or workforce productivity. However, by definition, most chatbots are simply providing responses to basic questions, which means only a class of employee requests can be automated. Advanced chatbots, on the other hand, are capable of automating business processes. When you automate password reset, laptop refresh, mailing list management and more, employees will keep coming back because it is so easy. And speaking of automation, your chatbot needs to be able to integrate approvals too. Sending employees elsewhere for approvals will not work.
- Omni-channel accessibility
Employees have been using email and calling the help desk for years. Most chatbots have a limited conversational AI interface, which means your employees need to change their established behavior in order to maximize adoption. You cannot simply offer a Microsoft Teams interface and claim victory with 100% employee adoption. To truly maximize adoption, a chatbot needs to be accessible on any desktop or mobile device, deflecting employee issues across multiple interfaces including email, phone, service portal widget, collaboration tools, web, and native apps.
- Understand employee language on the first day
Declan Morris, former CIO of Splunk, recently commented on the importance of employee adoption by saying, “It’s all down to first impressions. If you roll something out to the broader community and it does not deliver on its promise, you don’t get a second chance.” Most chatbots require significant training of language models in order to understand what your employees are asking – but where does the training data come from and how much is enough? Many chatbots have 10 or 20 templates to start from, which is quite limited and requires significant training to be successful. Since you cannot guess in advance what your employees will be asking, we believe that in order to meet employee expectations a chatbot needs to understand literally millions of employee phrases the day that you launch it.
- Intelligent integration of knowledge articles
For organizations that have a strong knowledge management practice in place, your chatbot needs to be able to ingest and leverage your knowledge articles. What is important is that supporting knowledge does not mean a chatbot should be delivering the same old knowledge experience that your employees have voted against. No one wants to have to parse through multiple documents in the hope that an answer to the problem will be found. The chatbot has to intelligently map articles to the language model so that it can match articles to employee questions, even if the articles use different terminology. Keep in mind that this won’t happen if the chatbot is using keyword search.
- Intuitive and fast content updates by anyone
One of the many lessons that we learned when COVID-19 struck is that chatbot content needs to be able to be updated daily, if not hourly, as policies and processes are constantly changing. This could never occur if all content updates have to be handled by IT service management agents or by the chatbot vendor. An enterprise grade chatbot is designed to enable your subject matter experts across your organization to teach your chatbot responses, without requiring any technical skills. Further, it won’t help if you need an admin or developer to retrain the language model anytime someone makes changes to your content. Make sure the solution you are looking at enables non-technical subject matter experts to quickly and easily update content and publish it based on time, role, location, etc. This will be critical in the “new normal” when you have employees transitioning between work from home and work in office and back.
- Advanced natural language processing
I saved the best for last—or should I say the most critical for last. None of the above will work if your chatbot does not have advanced natural language processing (NLP) that is optimized to understand the language of employees. Many solutions on the market were created to address any and all problems, and not optimized for use with employees. Further, most NLP solutions operate off keywords, which means they can only understand phrases if your employees used the correct words. Unfortunately, humans are largely unpredictable, and success requires a solution that can understand what your employees are saying—not the other way around. With advanced NLP, your chatbot will be able to understand what employees are asking no matter how they phrase it. If an employee asks, “I have an issue with my laptop,” your chatbot needs to know that is an IT incident. Yet if they ask a question that uses the same words in a different order like, “issue me a new laptop,” your chatbot needs to know that is a service request.
As IT, HR, and Facilities leaders prepare for the “new normal,” employee self-help will be critical. That is why we are seeing so many organizations issuing RFPs for intelligent automation of their help desks. Take time to learn more about what your employee self-help chatbot can’t live without. We have an eBook that goes into detail on the 7 topics I just covered briefly, with suggested questions to include in your RFP. You can access that eBook here.
Let me know if I can help as you navigate through this new time.