Chatbots - When and How to Use Them

Is your business in the business of helping other businesses? Or providing a service for customers?

There are many methods for your organization to reach your customers: email, TV and radio ads, paid internet ads, social media… but what channels do your customers have available to reach you?

Often, there’s two ways for a customer to reach out to an organization: over the phone, or via a “contact us” box somewhere on a website.

There are, however, better ways to facilitate communication between you and your customers — and best yet, they’re highly scalable. After reading this blog, you’ll have a roadmap for how you can offer the best possible customer communication experience on your website.

CHATBOTS

You may have heard a lot of buzz around a certain ChatGPT — but don’t worry, because this isn’t about that.

Did you know that most web services like HubSpot, Wix, SquareSpace, and Wordpress all support templates with chatbots? Did you also know that Chatbots are currently able to manage about 70% of all customer interactions?

Below, you’ll find a step-by-step approach you can take to create the perfect chatbot for your customers.

STEP 1: Determine the best type of chatbot for your business

Before setting up your chatbot, it’s important to determine what the ultimate goal of your chatbot is: do you want to provide automated customer service 24/7, or do you want to increase your lead generation?

Keep in mind that chatbots can do whatever you want them to (within reason), so you technically could build a chatbot that both generates leads and answers customer inquiries. However, your chatbot will require backend infrastructure to provide the logic framework to actually function — so the more you want your chatbot to do, the more you have to build.

STEP 2: Create achievable goals with measurable results

Implementing and managing a chatbot takes work, and work takes time, and time is money. So, in order to measure the impact your chatbot has on your organization, you’ll need to set and measure KPIs centered around the aspect of your customer interaction you want to automate.

For example, if you were implementing a chatbot to facilitate customers ordering supplies from your website, you could measure how many orders are placed through the chatbot, versus how many are placed manually or over the phone.

STEP 3: Create your user persona

It’s not only important to figure out what your chatbot needs to do for your customers; you also need to determine the kind of person who will be interacting with your chatbot.

Are they on their lunch break, trying to figure out how to place an order on your website? Are they attempting to contact someone from your sales department, and don’t know how?

When users are interacting with chatbots, they want answers, and they want them quickly. When creating this user persona, think about it like your companies mission statement. For example, your user persona could be a few sentences describing the situation the user finds themselves in, and ultimately, what they should expect to receive form interacting with your chatbot.

Make sure that the first thing your chatbot does when interacting with users is identify itself as a bot. This will prevent any confusion or ill feelings on the user’s part.

STEP 4: Create your chatbot personality

No one actually wants to talk to a robot — we’re social creatures after all. This is why it’s important to give your chatbot a personality. 

Think of this as an opportunity to create the perfect customer service representative who is always ready to assist your customers — but who is also the perfect brand ambassador. Your chatbot will follow your brand principles to the letter every time, as long as you set it up to do so.

This includes coming up with a name for your bot that matches your brand (you may have already interacted with our bot, Cobby), making sure your chatbot employs casual language or even industry-specific slang, and a friendly profile photo for your robotic helper.

STEP 5: Design your chatbot’s logic

Now is the time to determine what happens during your customer’s conversations with your chatbot. This sounds a lot more daunting than it actually is. 

The best way to think about this is a simple logic gate: if the customer is presented with three options, and they choose option 3, they then go down option 3’s logic path. The second step in option 3’s path would branch into two paths depending on the customer’s selection, and so forth.

You can map this out using sticky notes, on a whiteboard, or any digital program you prefer. In the end, however, you should have a visual representation of branching paths that ultimately lead to the destination the customer wanted to find.

STEP 6: Design your chatbot’s UI

This is where the fun begins. Those website builders mentioned above — like HubSpot, SquareSpace, and Wordpress — allow for your to change the look and feel of your chatbot’s user interface.

This can be as simple as choosing what color the pop-up window displays, the typeface it displays, and what intractable elements look like, as well as how they function.

This stage of development is more than just window dressing, however. Using your previously-mapped out logic, you can now design how exactly the customer will interact with your chatbot. 

Does the customer type in their responses to questions from your chatbot, or do they click a button that leads to a page or a follow-up question?

STEP 7: Implement and test

If you’ve ever worked on any digital, customer facing product — like building your website or a customer portal — you’ll be familiar with pre-launch testing. If you aren’t, don’t worry. You’ll find everything you need to do below:

To begin testing your chatbot, implement it on a page only you and those in your organization can access. Then, simply begin going through the path a customer would take to achieve the goal they want to accomplish. 

If that goal is to renew their subscription to your services, follow that chat flow until you have completed the process. Do this for every path available.

Keep in mind that you may not be the best person to do this testing. Since you have intimate understanding of the chatbot’s logic, you’ll know exactly how to frame queries, as well as which buttons to press to reach your intended destination.

Find an employee (or better yet, employees) who have played no role in the development of your chatbot. Then, give them goals to achieve using your chatbot. Then, set them loose and watch from afar.

It’s important during this step to not influence or direct your testers in any way. You want to get a clear picture for how someone who has never been on your site would interact with your chatbot.

STEP 8: Create your chatbot’s activation parameters

This is the step where you will determine when your chatbot enters into your customers’ website interactions.

There are many different parameters you can use to trigger the activation of your chatbot: spending a certain amount of time on a page, visiting a particular page, clicking a CTA, pressing the “back” button on their browser, or even if they are on a particular type of device, or located in a specific geographical location (if you have cookies turned on).

Don’t feel like you need to make this overly-complicated. If there’s only one activation parameter needed for your chatbot, that is fine. In fact, the less activation parameters it has, the easier it will be to implement, and the less likely it is to annoy a customer.

If you chatbot is constantly popping up and interrupting a customers interactions with your website, it won’t be viewed as the helpful guide you intended it to be. There’s nothing more annoying than a robot offering help when you don’t need it.

STEP 9: Implement and Measure

Now, you can finally implement your chatbot site-wide! This is why the goals you set in step 2 are important. In addition to those goals, there are a few data points you’ll want to measure:

1. Activity: How many customers are interacting with your chatbot in a given time period?
2. Rate of completion: How many customers are ending their conversation with your chatbot without finding the answer to their question?
3. Effectiveness: What is your ratio of customers who found what they were looking for, versus how many needed to speak directly to your customer service?

If you have the chance, it’s highly valuable to collect qualitative data from your customer’s interactions with your chatbot. Did they enjoy interacting with it? Was it useful? Would they use it again?

STEP 10: Optimize

Based on the data you receive from step 9, you’ll need to begin refining your chatbot’s logic, as well as optimizing the experience for your users.

The interactions your customers have with your chatbot will inform you on what steps you need to take next, but make sure to prioritize any issues that causes a pattern of customers stopping their interaction with your chatbot.

For example, if 50% of customers stop at the same exact point in your chatbot’s flow, you’ll want to revisit that part of their conversation to determine why. Look closely at the phrases and tone your customers are using, and then determine what you could do differently to improve their experience.

And now, something completely different:

Here at Cobb, we take customer service seriously. It’s why BEI/NEXERA, an independent customer satisfaction survey, ranks our customer service higher than companies like Amazon and Costco.

To learn more about what makes Cobb special, click here.