Getting Started with a Chatbot – How Difficult is it?
As we see wider adoption of chatbot technology in the fields of customer service and employee engagement, many organizations are considering if a chatbot might solve a problem for them, too.
While the idea of using a chatbot in general may appeal, the thought of what it might take to get started with a chatbot can be overwhelming and, consequently, off-putting. But how difficult is it, really, to get started with a chatbot?
The most basic questions that must be answered before starting work on your chatbot is, what do you want it to do and what are you trying to achieve? We always recommend that a staged approach is taken to building and launching a chatbot; choose one or two areas to train the bot on first and, when you are happy with performance, add more questions and answers over time.
The good news is that, In GeckoEngage’s experience, getting over that first hurdle to having a working chatbot can be as easy as 1, 2, 3.
Start with the A, B, Cs (or F, A, Qs)
Even if you don’t have a specific Frequently Asked Questions (FAQ) section on your website, you’ll still be very aware of all those routine inquiries that just keep on popping up. Whether it’s someone asking for the price of your product, a query on your physical locations, or more detailed information on the service you provide, every organization has a bank of questions that website visitors are looking for answers on. Use them!
Import the questions and answers to your chosen chatbot platform and, once launched, you’ve got a functional chatbot feature for your website that provides the answers to your most frequently asked questions. This can be done in as little as an hour and provides you with a straightforward, scripted chatbot.
To make things a little more conversational, put aside some extra time to word each frequently asked question in slightly different ways. We recommend ten versions of the same question as a starting point. For example, a university using a chatbot for student enrollment events might include a question such as “How do I get to the campus?”. Similar questions to add on to this FAQ might be: “Where is the university campus?”, “Can I get directions to the campus?” or “Where is the campus event happening?” and all of these questions can point to the same answer which details the address and a link to a location map.
Train that chatbot to do people’s bidding
To create a more intuitive, intelligent and sophisticated chatbot, use the collective knowledge and time of your team to build it out further.
Get each team member to set aside an hour a day for a couple of weeks to train the bot by asking it questions and testing the answers. If the chatbot doesn’t provide the right answer, or can’t provide an answer, to the question asked, add it to the data bank. If the bot isn’t sure, but does provide a suggestion – and it’s correct – log in to the back end and add that question variation to the data bank.
Over time, the chatbot will learn these new questions and variations and will use that experience to better answer your visitors on a wider and wider range of topics.
Going back to the previous example of university campus visits, say, for example, a prospective student asks if there is car parking. The bot will have been trained to associate this request with transport/directions and will be able to point the user to the address, location map and where car parking is available. The longer your intelligent chatbot is in use, the better it gets at answering user questions.
Keep it human
Of course, there will always be times when an actual human will want to step in to the conversation. And triaging conversations at the right point, to the right information source is simple when using a chatbot.
There are multiple routes to channelling those more complex requests through to the appropriate team at the appropriate time. You can train the chatbot to respond in a specific way to certain keywords and funnel that request through to start a live chat with the right person. You can alert teams when a user has started a chat on a specific page of your website; a page that you know will spark a more complex information request. If there comes a point in any routine chat where the bot can’t answer, it will automatically ask the user if they would like to speak to a human or ask the question in a different way.
This kind of simple, staged approach to launching a chatbot breaks the process down into manageable chunks and helps you to get started quickly, without feeling overwhelmed. And the very real benefit of an intelligent chatbot is that, while training will be needed in the early stages, the bot will learn over time, maximising on its collected knowledge to improve day after day.
On top of simplicity of launch, it’s also very straightforward to get your chatbot integrated with existing data capture systems, such as your CRM (Customer Relationship Management) tool. Using GeckoBot as an example, a new organization deploying the chatbot would simply have to choose the CRM system they use on the chatbot dashboard, log in and match up the fields that they would like to sync. Boom! Data captured via the chatbot is logged in the CRM system and the full transcript of the chat can also be saved to the CRM record.
So, training and launching a chatbot for your organization is perhaps not quite as daunting as you once thought. And, over time, it becomes more and more efficient; freeing up staff to focus on higher impact work.
GeckoEngage offers chatbot, live chat and event management technology. For further information on how we can help your organisation, schedule a free demo.