Do you really need a Chatbot for your business?

Chatbots are at the peak of their hype-cycle, conversational commerce is the new silver bullet that will save the world from bad UI and and let us all undertake any task by pretending we have a friend on the other side of the screen, but wait, if chatbots are really that good, why hasn’t every company integrated a digital personal assistant into their customer facing business?

Well, even Facebook’s M platform has its limitations, it’s all well and good being able to attract a portion of Facebook Messenger’s 1.2 Billion users, but the actual service is more of a decision tree than real, interactive conversational commerce. I recently worked with a large multinational retailer that developed a chatbot for messenger, and NLP was kept to a minimum to try to protect the customer experience. I’ll let you digest that again, natural language processing, the most integral piece of UX when it comes to conversational commerce, was kept to an absolute minimum. Why? Because NLP is hard, NLP allows users to be let loose in the sandbox, and when that happens, people break stuff (think Microsoft’s ‘Tay’). We’re in a place where traditional brick and mortar companies see the appeal of self serving customer queries, but they’re either too scared to impact customer satisfaction in the short term or they simply don’t know how to approach the problem. To an extent I agree with the approach, with call centers you can assure a minimum level of service, if an agent is consistently giving poor service, you can remove them from that position and employ someone else, in addition to this the impact is only on a small volume of your total customer base. If there is an issue with a customer service chatbot, if something doesn’t work the first time customers use it, if a key feature ‘breaks’, if the bot is giving ‘bad’ output, then UX suffers in a big way. So, what is the solution? If you are a stakeholder in a large business (or even an SME) how do you approach the issue of self service chatbots in a sustainable, exploratory way without damaging that vital customer journey?

Well, even Facebook’s M platform has its limitations, it’s all well and good being able to attract a portion of Facebook Messenger’s 1.2 Billion users, but the actual service is more of a decision tree than real, interactive conversational commerce. I recently worked with a large multinational retailer that developed a chatbot for messenger, and NLP was kept to a minimum to try to protect the customer experience. I’ll let you digest that again, natural language processing, the most integral piece of UX when it comes to conversational commerce, was kept to an absolute minimum. Why? Because NLP is hard, NLP allows users to be let loose in the sandbox, and when that happens, people break stuff (think Microsoft’s ‘Tay’). We’re in a place where traditional brick and mortar companies see the appeal of self serving customer queries, but they’re either too scared to impact customer satisfaction in the short term or they simply don’t know how to approach the problem. To an extent I agree with the approach, with call centers you can assure a minimum level of service, if an agent is consistently giving poor service, you can remove them from that position and employ someone else, in addition to this the impact is only on a small volume of your total customer base. If there is an issue with a customer service chatbot, if something doesn’t work the first time customers use it, if a key feature ‘breaks’, if the bot is giving ‘bad’ output, then UX suffers in a big way. So, what is the solution? If you are a stakeholder in a large business (or even an SME) how do you approach the issue of self service chatbots in a sustainable, exploratory way without damaging that vital customer journey?

Gather your data: What is the cost of your customer service overhead currently? Whether it is replying to emails, answering the phone or people coming in to your stores/offices to ask questions. Cost can always be estimated in a financial way, this shouldn’t be your main driver, the customer journey should, but that financial data can act as a key factor towards your decision.

Start small: You don’t need a fully fledged AGI answering everything that a customer throws at it including the distance to the moon (384,400 km) and the current age of Barack Obama (55 at the time of writing), you should aim to solve the biggest automatable queries that your customer-base (and potential customer-base) will need answering. If your biggest time drain is customers calling you to check if an item is in stock, then that’s what you should think about automating.

Will your customers use it: Now I don’t want to sound ageist here, but there is a high chance that if the customer demographic for your 70’s+ bingo hall doesn’t generally own a smartphone/ tablet/ PC, your bot will probably not have the sweeping impact that you wanted it to. Do some customer research, actually talk to the people that you want to use it, they are stakeholders in the product as much as you are. If the consensus is that they wouldn’t use it, then the phrase “if you build it they will come” may not fit your business as well as it did ‘Field of Dreams’.

Don’t build it if you don’t need it: Something you should ask yourself before you even consider putting the time and effort (and money) into producing a chatbot for your business is, “Do I really need this?”. If you know the questions your customers have, and you know the answers, would an interactive FAQ be more easily accessible and a quicker win for your business? Could a change in process fix the issues that your customers are facing? If you can limit the impact to the point that it is negligible, the investment necessary to create a really good customer journey using a chatbot may not be worth it. I mentioned at the beginning of this article that chatbots are at the peak of their hype cycle, with the numerous frameworks available and the tumultuous noise gravitating around bots currently it’s easy to get sucked in, quite simply, don’t.

My final suggestion is to really consider the issue you’re trying to solve, even if a sensitive issue is easily solvable with an automated response, telling somebody their mortgage application has been denied, or giving medical advice using a chatbot isn’t a good user journey (it can also be quite dangerous). Also as a rule of thumb all complaints that don’t result in an immediate resolution in favor of the customer should be handled by a real person. A bad bot experience when a customer is emotional can be a gargantuan UX killer; bad UX will lead to bad customer journeys, and bad customer journeys lead to the dark side. Repetitive queries can be automated, bespoke queries should be handled IRL. More

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s