Posts

The Success of a Chatbot is in the Details

By Mandy Reed, Global Head of Marketing

Last year many of us were spending much more time at home than usual as we did our part to slow the spread of COVID-19 in our communities. This created a surge of another kind, though: a wave of do-it-yourself project attempts. If you’re at home anyway, why not try to tackle that project yourself?

As many DIYers quickly discovered, the devil is in the details. No matter how easy something looks in the YouTube video or how complete the step-by-step instructions may appear, some projects are better left to the professionals. Sometimes recognising the small details needed to really succeed requires an expertise that comes only from experience.

This is most certainly true when it comes to creating chatbots for customer self-service, and too often organisations fall into the DIY trap. Some companies delegate their project solely to an inexperienced internal team because they underestimate the amount of expertise needed to build and deploy a working chatbot. Some companies treat their chatbot as an unimportant DIY side project not worthy of dedicated resources and investment because they fail to recognise the importance of conversational AI in modern customer experience (CX) strategies. Whatever the reason for a DIY approach, these organisations soon discover that they should have collaborated with a chatbot professional.

A good example of an important detail for chatbot success that is often overlooked comes from Maria Ward in a conversational AI vendor selection guide. Maria, an industry expert with over 15 years of experience working with chatbots and virtual agents, shares this insider tip:

“Avoid using ‘Yes’ and ‘No’ at the beginning of virtual agent answers as it may not fit with the many ways a question may be asked. Adding the subject within an answer also gives the user confidence that they have received the correct answer. For example, when answering ‘Can I have 3 slices of cake?’ instead of ‘Yes, you can.’, use ‘You can eat as much cake as you like!’”

 Besides making me hungry for cake, this quote also illustrates an important point about the importance of the details in chatbot creation. When you read Maria’s tip, you might think that structuring the chatbot’s answers in that way should be common sense. It sounds so logical. But, if you’re not experienced with creating chatbot content, is that something you would automatically know? Probably not.

Really understanding how to structure answer content to work with the many ways users may ask the same question is just one skill that comes with experience. It’s a detail that’s crucial to delivering a quality self-service experience and building user confidence in your conversational AI tool. It’s also a detail that could very easily be overlooked by an inexperienced DIY team.

When it comes to creating a successful chatbot and delivering a positive customer service experience with conversational AI, the devil really is in the details. Unless you have a highly experience internal team, taking a DIY approach is not worth the risk. CX is a critical competitive differentiator for organisations, and a poor performing self-service tool will quickly erode customer loyalty and your brand reputation.

Will Old Internal Systems Destroy your Conversational AI Dream?

By Mandy Reed, Global Head of Marketing

Microsoft announced this week that they will stop supporting Windows 10 in 2025, with no new updates or security fixes being released after 14 October. This news comes as the company plans to reveal a new Windows operating system later this month. Screenshots of the new Windows 11 have been leaked online ahead of that virtual event – you can check them out here.

Whether you’re a Windows user or not, this type of announcement can highlight the need for wider discussions about the technology and systems being used internally at your organisation. Are you one of the many companies that relies on old or out-dated solutions? These legacy systems can have a knock-on effect, sometimes creating issues when the restrictions they create aren’t acknowledged at the start of a digital project or even keeping organisations from embarking on new projects at all.

For example, this can be the case when deploying an internal conversational AI solution for employee support. As virtual agent and chatbot expert, Peter Studd explains in a conversational AI guide:

“It’s very important to be aware of your users’ technology when you’re implementing a virtual agent to ensure it will be compatible. For example, it’s not uncommon for large companies to have very old systems and browsers that are still being used companywide. This can cause issues if not taken into consideration when developing an internal virtual agent.”

Constraints created by old systems and browsers don’t necessarily mean you need to abandon your plans for an internal virtual agent. However, they should be identified and discussed at the beginning of your project to avoid any unwanted surprises. If not, you could find that you’ve invested time and money into a solution that can’t be properly integrated with an important legacy system or isn’t user friendly for employees using older browsers.

This is just one of the downsides of taking a DIY approach to building and deploying a chatbot or virtual agent. If you don’t have people with the right experience on this team, it will be easy to miss or overlook older systems that could create project roadblocks. You end up with a case of ‘we didn’t know what we didn’t know’ and a tool that doesn’t meet expectations.

An experienced conversational AI provider will know the right questions to ask to help you pinpoint any potential compatibility issues upfront. They’ll provide guidance on any changes or accommodations you may need to make to your plan. They’ll also be able to offer suggestions for ways to work with potential constraints based on a deep knowledge of their technology and prior experience with similar challenges.

In an ideal world, all of the systems and technologies being used by your organisation would be up-to-date and easy to incorporate into new conversational AI projects. Since that’s not likely to become a reality any day soon, the best course of action is to work with a provider that can not only help you identify potential challenges but also provide the guidance on how to overcome them. Old internal systems don’t need to destroy your conversational AI dream.