Businesses are finally starting to get value out of using chatbots. Companies like JPMorgan Chase, Bank of America, Marriott, and Coca-Cola are starting to see massive returns.
● Rodgers Wireless had a 60 % improvement in customer service satisfaction.
● Marriott use of chatbot services has grown 85 % month over month, since the technology was launched via Facebook Messenger.
● Swedbank chatbot, Nina, handles 40,000 conversations a month and resolves 81 % of the issues.
All in all, bots are finally starting to win businesses over and, according to an Oracle survey, 80 % of businesses want chatbots by 2020.
Here’s a look at where the biggest opportunities are and how to take advantage of them.
1. Pick use cases with an instant ROI
If your company is looking to leverage chatbot technology, you’ll need a good place to start, and one of the biggest tips is focusing on ROI.
The following use cases have a near instant ROI:
A. Customer service.
Companies are saving as much as 30% on customer service by deploying bots , according to IBM. BI Intelligence estimates that chatbots will produce $23 billion in savings from annual salaries. Additionally companies can automate 36% of sales representative positions, resulting in total annual estimated savings of at least $15 billion from salaries.
B. Sales and Marketing.
Chatbots have quickly become a very powerful content marketing tool. Brands like 1–800–Flowers.com, Sephora, and Nitro Café have received massive value from their bots.
- 1–800-Flowers.com reported that more than 70% of its Facebook Messenger orders were from new customers.
- Sephora increased its makeover appointments by 11% via its Facebook Messenger chatbot.
- Nitro Café: increased sales by 20 percent with its Facebook Messenger chatbot.
C. HR and Recruiting.
According to a survey conducted by the Society for Human Resource Management, 38% of HR respondents said the primary challenge they face is “maintaining high levels of employee engagement.” Part of the problem is that most employee self-service HR systems are antiquated and hard to use, which results in employees bringing even their simple requests directly to the HR department. Bots are solving this problem by becoming a middle layer that can integrate with existing systems, while offering a much easier user experience for employees. As a result, bots are improving engagement while decreasing the amount of time HR reps spend on requests by 30–50%!
Let’s look at a few examples of winners in the space:
- Maya: Automates recruiting steps, such as sourcing, screening and scheduling, saving up to 75% of a team’s time.
- SGT STAR: US Army chatbot that helps to answer questions and enlist future soldiers.
D. Backend and business-to-employee (B2E) bots.
These are perhaps the biggest use cases no one is talking about. Chatbots are being used to expose and work in conjunction with core enterprise systems like robotics process automation (RPA), service management, field force management systems, and CRM. This automates a lot of low-level tasks and allows employees to interact with outdated enterprise software in a more intuitive and efficient way. Here is one of my favorite examples:
○ JP Morgan Chase has numerous bots, including COIN, which has saved over 360,000 hours of manpower.
Its simple bots work by parsing emails for employees, granting access to software systems, and handling common IT requests, like resetting passwords. Bots are expected to handle 1.7 million access requests this year, doing the work of 140 people.
2. Start small and focus on ROI
Building bots is a lot different from building apps. The key difference is that, with a traditional application, the user has a limited number of actions he or she can take — they can click a button, zoom in or out, fill out a form. However, with a chatbot, the user can request anything — the inputs are limitless.
To simplify this issue, you need to start out small. Think of one thing that your chatbot needs to do really well, and build the bot using a very lean minimum viable product (MVP) approach. As people use your bot, you will learn what the most common requests are and will then be able to decide how you want to continue building it.
3. Data is key
Every time a user chats with your bot, they give you an enormous amount of information about who they are and what they want. The key to getting the right data is knowing which questions to ask. So consider, what information you need to know most about your customer and how can you design your bot in such a way that the user wants to share that information. You can learn more about conversational marketing here.
4. It’s all about the relationship
The most successful bots build relationships based on trust. Consider these four pillars:
- Learn: Every time a customer uses your bot, they are making decisions. Each decision gives you very valuable data on who that person is and what they want.
- Personalize: After you know who the user is and what they want most, you can give it to them. Bots allow your brand to deliver personalization at scale.
- Deliver value: Your bot should improve the customer experience and allow them to do things that they previously couldn’t. At minimum, your chatbot should reduce friction.
- Build a relationship: Like all relationships, conversations are vital. Focus on how conversations can help you learn even more about your customers, build trust, and develop a relationship with your brand.
5. Customer journey
Your brand’s chatbot can automate each step of the customer journey and can bring a smile to each customer’s face. So consider how your bot can help the customer along the way, while also showcasing your brand’s story.
6. Errors, failing, and continuous training
At some point, your bot will fail, and that is a good thing! Language is far more complicated than pressing a button, and there are hundreds of ways to make the same exact request. Additionally, current natural language processing (NLP) technology lacks the ability to understand context and intuitive meaning. For example, if I ask you, “Should I bring an umbrella?” chances are that I am hinting about the weather. A chatbot would have no way to make sense of such a hint.
Why bot failure is actually good thing
When bots fail, it is because they are unable to understand or to help customers get what they want. In essence, users are always telling you exactly what they want, which gives you the information you need to be able to give them a magical experience. In this way, bots are kind of like a mass survey. After you know what it is that the customer wants, it is up to you to decide whether you want to support it.
Once you get your bot going, understand that it will continue to evolve and it will need continuous training so it can improve and delight your customers.
7. Trust and security are key
Enterprises take security seriously, and this includes the security of the content and the transactions the bots have with users. It is important to understand which messaging channels are compliant with information security standards, such as GDPR, HIPAA, SOC 2, and others.
Additionally, you want to start on the right foot with your customers and prepare them that they are about to use a chatbot.
One of the best ways to do this is by using the .BOT domain. When users see that they are going to a .BOT domain, they have the expectation that they will be interacting with a bot.
The .BOT domain was recently launched by Amazon and is reserved for chatbots only. Amazon’s .BOT helps bots earn trust because bots go through an authentication process during domain registration. Additionally, using .BOT offers big SEO and branding advantages to enhance your bots identity and discovery.
Building Trust, Discoverability, and Standardizationchatbotslife.com
BONUS: How to get started with chatbots
Creating a chatbot from scratch can be expensive, and managing it can be a nightmare. One of the biggest bottlenecks is that bots need constant iteration which means that your product managers, marketers, designers, and copywriters all need access to the bot. As a result, there are a lot of chatbot building platform solutions, however most of them are not secure and are expensive to integrate with.
One of the best solutions for this is Smartloop, an enterprise-level chatbot management software which can be hosted behind your firewall. It uses Rasa (open source and hostable behind your firewall) as it’s NLP engine , which means that you all of your data will be yours, behind your firewall and compliant with HIPPA, GDPR, and SOC.