8 Ways to Know if You Need a Chatbot For Your Business

Choo-choo the chatbot train is chugging along at full speed and if you’re not on it, you’re missing out!

At least that’s what it feels like when you read articles about why your company needs a chatbot. But chatbots are a big investment and aren’t necessary for every business. If you jump on without doing your research, then the chatbot train will start to feel less like the Polar Express and more like the Trans-Siberian.

It’s true that chatbots can help make customer service faster, cost-efficient and modern, yet they require a surprising amount of upkeep.

But they’re popular, they look snazzy on a website and they’re a viable option for boosting your customer communication. Do you really need one? This post will help you find out.

Chatbot options

Today’s bots come in many shapes and sizes. A lot of them are even free and ready to go right out-of-the-box. But is it enough to fit your needs?

Short answer: No. At least not for customer service. Many businesses do use the likes of Facebook Messenger and other bot building platforms for their chatbots, but users are left to click buttons and pick topics from predetermined menus.

image of pyramid graph showing that the quality of user experience is higher the more advanced a chatbot is

But businesses usually want more than this; a bot that is cute, intelligent and human-like to help cut service team costs.

In order to create a chatbot that can match your team’s service quality, you have to make a big investment. Creating this caliber of chatbot is a major, long-term job that requires developers and a fully customizable platform like IBM Watson .

But this is something many businesses don’t consider. Now half-ripe chatbots populate the internet frustrating customers, giving other sophisticated chatbots a bad name.

So is it worth investing in a complex and highly-developed chatbot? There are a few things you need to assess.

Will a chatbot add value?

Before you start shopping around for chatbot platforms, thoroughly examine your processes. For example, chatbots can help improve:

  • Reducing cart abandonment by offering to help with the checkout process or give a discount code.
  • Boosting sales leads by asking visitors necessary questions.
  • Automating common tasks like scheduling meetings, collecting feedback and sharing promotions.

Using live chat as your primary service channel may be a better option. It’s a direct line of communication that can help with long waiting times, slow responses, missed leads and other pain points of phone and email customer service.

This was the solution Frankfurt School of Finance & Management adopted. The school first considered adding a chatbot to improve its online presence but found that live chat suited their needs better.

Building a chatbot infrastructure is a costly matter. Before investing in it, we wanted to assess how students respond to chat as a contact channel.

Florian Fürst, Frankfurt School

Take a look at our post "Live Chat Benefits in Theory and Numbers" to get a better idea if live chat is right for you.

Can you closely maintain your chatbot after launch?

Giraffes can stand and walk within minutes of being born. Thanks to their long gestation period of 15 months , giraffes are already highly developed when they exit the womb. They quite literally hit the ground running.

Think of your chatbot as a baby giraffe. It takes a long period of development until they’re ready for your website. Once implemented, they also hit the ground running. However, it’ll take weeks of observation, extensive “training” and learning for your chatbot to be self-sufficient.

Take Capital One’s chatbot “Eno” as an example. The banking assistant was built from scratch in three months using real customer conversations from chat logs.

image of a conversation with Eno

Eno understands over 2,200 terms and emoji and speaks conversationally, not via commands. Its engineers used a three-step process to train Eno:

  1. Teaching Eno the meaning and similarities of words.
  2. Feeding Eno the tens of thousands of utterances customers made in chat logs (like “activate” and “make my card work”).
  3. Training Eno to understand new terms customers haven’t used before based on those utterances.

Texting with Eno is simple, but, of course, the easiest-to-use technologies are the hardest to build.

Margaret Mayer, VP of Software Engineering, Capital One

Capital One is the 10th largest bank in the United States, so it’s unsurprising that it could afford to create such a refined bot in a short amount of time.

The lesson is not to be discouraged from building your own chatbot from scratch, but to take a closer look at the reality. That’s why chatbots are like children. Making a baby often isn’t pricey (sometimes even a bottle of Rosé might do) but raising a child until they’re 18 can cost parents up to $234,000 .

An intelligent bot like Eno can have starting costs ranging from $20,000 to $50,000 . Companies that took their senior developer salaries into account reported spending over $300,000 for their custom bot. Not to mention the time and money spent keeping it running smoothly.

It may be in your company’s best interest to stick to human agents if you don’t have the resources to “raise” your chatbot.

Are chatbot functionalities too limited?

Even the best bots have their limits. Eno may be programmed to understand thousands of responses, but it will always lack empathy.

Chatbots can’t bend the rules or handle individual use cases. They also have difficulty detecting when a customer is becoming increasingly angry or irate. Its social abilities are entirely in your hands.

Check what types of questions your customers or users ask. Are they repetitive and something a chatbot could answer easily, or are they usually specific?

Will a chatbot cheapen your business?

Chatbots excel in service-heavy industries. That’s why you’re bound to meet a bot when browsing for flights or clothes.

They can make product suggestions, reserve tickets, forward users to certain webpages, suggest dates and times — tasks that are easy to predict when writing a script.

But luxurious services and industries risk looking cheap by using chatbots. Imagine walking into the fine dining restaurant Eleven Madison Park and using an iPad built into your table to order food. Excessive technology just doesn’t fit the environment.

Chatbot conversations can also quickly go south. Replika , one of my favorite chatbots, has trouble understanding basic questions despite being a highly-developed conversational bot.

screenshot of conversation with Replika gone wrong
Some users say that Replika’s speech improves over time, but others disagree . Chatbots are too unpredictable.

High-profile businesses tend to use curated language when speaking to customers or potential clients to uphold an image. My former director at the university I worked for would go over emergency text messages to the student body with a fine-toothed comb. The university has a reputation to uphold, making it difficult to trust a chatbot to communicate with the same level of care.

Is a chatbot right for your target group?

Your customers may not be the right demographic for chatbots. According to Chatbot Magazine’s “Chatbot Report 2019: Global Trends and Analysis,” a survey found participants aged 18 to 34 were twice as likely to talk to a chatbot when shopping.

infographic of messaging support in the adoption curve, a major customer service trend
From Facebook’s “Insights to Go”

Additionally, women speak to chatbots or live chat while shopping online more often than men.

If older men are your target audience, you may want to reconsider your chatbot idea. But there are still a couple factors to consider.

People prefer to communicate through text rather than phone , which is great for the future of chatbots. But if your company is dependent on phone customer service, then the sudden switch to chatbots may feel odd.

Many brands also use emotions to sell products, a chatbot’s greatest downfall. Rational compassion and appealing to a customer’s personal experiences is difficult to program.

For example, the German prosthetics company Otto Bock uses empathetic language on their website to sell their products:

screenshot of Otto Bock's website

Otto Bock representatives likely speak to customers who are struggling to cope with their situation on a daily basis, which is evident in their copy.

If a new customer were to contact their support team and speak with a bot that starts spewing tech specifications and generic condolences, it would hurt their image.

Is your team ready to be a safety net?

Sometimes questions or issues are too complex for a chatbot to handle by itself. That’s why participants in our chatbot survey agreed that speaking to a bot directly is okay as long as it’s easy to escalate the situation to a human agent.

Userlike makes it easy to integrate existing chatbot frameworks , which means you can pair your fully customized bot with our chat solution.

infographic of messaging support in the adoption curve, a major customer service trend
Henry greets you on LogoEnergie’s homepage when you select "chat"

If your chatbot is proactively messaging visitors on your site, you may notice an increase in chats. Make sure you have enough agents to help cover these handovers.

Are you prepared to adhere to privacy laws?

A chatbot is only as complex as its infrastructure and the data it can access. But this may require gathering private details about your customers.

Data privacy is an important development factor to consider. If your chatbot can access user accounts and collect customer information, then it needs to follow data protection regulations set by your country.

In the U.S., there is no single overarching data privacy legislation. Instead, the country takes a sectoral approach to data privacy, relying on a mixture of sector-specific laws and state laws.

In the European Union, however, the laws are rather robust. The General Data Protection Regulation requires controllers of personal data “to put in place appropriate technical and organizational measures to implement the data protection principles.” A service provider may only collect or access personal data on a contractual basis.

What are the laws in your country? Are you already well-versed in data protection because of the nature of your company? Then maybe a chatbot is no big deal. But if not, i-Sight has a brief overview of each country’s legislation.

Want to talk to a human about chatbots?

We'd be happy to answer any questions you have about using chatbots in your business and how to get started.

Book an appointment

Do you have the patience?

AI improves over time, but this means customers may have a less than ideal support experience for the sake of trying something new.

Chatbots are an ongoing project that requires a dedicated team. It needs to be monitored and tweaked to keep it running smoothly. Can you afford the commitment and attention?

Implementing a chatbot means asking for your customer’s patience as well. There’s no guarantee that they’ll be forgiving of a bad support experience, especially if they’re used to receiving high quality service.

A chatbot may not be worth losing customers for, especially if your company is still growing.

It’s better to have no chatbot than a bad chatbot

When social media started to grow in popularity, brands immediately created Instagram and Twitter accounts without any clear purpose or strategy.

Chatbots are experiencing the same phenomenon; businesses are throwing a crappy chatbot on their site to keep up with trends but customers are fed up with them .

An inefficient, useless bot can hurt your customer perception and drive new visitors away in a heartbeat. Why risk it?

The internet needs more good bots to show that AI can actually be helpful. So if you’re prepared to pour money, time and love into developing one, your customers will appreciate you for it.