With three billion users per month worldwide, chatbots are a valuable digital asset.¹ Would you like to kick your company into high gear by equipping it with a conversational agent? If you feel a little lost, don’t panic! Here are seven essential questions to ask yourself to ensure that your chatbot will be… a smartbot!
1) Why do you need a chatbot?
If you’re thinking about deploying a chatbot, it shouldn’t be just to keep up with the Joneses. From the outset, you need to identify the reasons behind your decision. A chatbot can respond to various, often complementary challenges:
- Economic challenges. A chatbot can help you reduce costs and secure your margins: by automating certain conversations directed to your call center or online, you’ll avail of a sort of low-cost concierge service.
- Service quality challenges. A chatbot improves the user experience by offering an instantaneous response that’s accessible 24/7.
- Marketing challenges. The implementation of instant messaging or an intelligent avatar reinforces your brand image. The relationship between your brand and your customers becomes fun and interactive, and personalization of the interactions with the chatbot represents a competitive advantage.
- Internal communication challenges. Deployment of an internal chatbot can considerably lighten the load of your HR department. Find out how!
Depending on your objectives, identify the KPIs that will allow you to verify the efficiency of your approach and make continuous improvements. That’s the key to success.
2) What should your chatbot be able to do?
You’ve precisely defined the challenge or challenges your future chatbot will meet. Now, you need to think about its education and training: its knowledge and dialogue capabilities must be a direct reflection of the goals of your project, as defined upstream. For that, you’ll need different types of information:
- Proprietary knowledge that will be “taught” to the chatbot
- Access to various internal or external information systems that will allow the chatbot to provide personalized responses or carry out automated treatment (to verify a delivery date, record a request for time off, recover a password, etc.)
One warning, however: it’s important to pay close attention to the scaling of your chatbot. If the goal isn’t technically viable, be more moderate and limit your ambitions to that which is achievable.
Read also: How to Reconcile Chatbots and the GDPR
3) Who will bring your chatbot to life?
You’ve identified the need for a chatbot, you’ve qualified its objectives and you’ve defined its scenarios. Now, you need to decide who will be the human behind the chatbot! Even though chatbots are developing increasingly strong capacities for self-learning, they must still be trained upstream, but more importantly, they must be constantly supervised in order to analyze their performance and enrich their knowledge.
So your chatbot must have a designated manager. Assign it a “tutor,” who will play a very important role during both implementation and operation. This manager must train the chatbot in its profession, bring it to life and allow it to evolve as needed, according to feedback from the field. That’s why managing the chatbot must be a true part of the employee’s job and not just an extra duty that’s performed from time to time!
4) Where will your chatbot be available?
Your chatbot now knows what it must do. But it still needs to be utilized! Which channels of communication will it inhabit? This is a crucial question: choosing the right channels, i.e. the ones where potential users of the chatbot are found, means ensuring satisfactory adoption of the service!
So, smartphone or computer? Website or mobile app? To guarantee the achievement of your goals, make sure that the chatbot becomes your users’ first channel of contact.
5) Should you give your chatbot an appearance? How do you choose one?
Another factor in engaging and retaining users is the personality of your chatbot. If you want your chatbot to differentiate itself, it must provide a full-fledged relational experience. The identity of your chatbot rests on several characteristics:
- Its avatar: what will it look like?
- Its voice: will it have one? If so, will it be male or female?
- Its tone: will it have a sense of humor or should it remain formal?
- Its reactions to questions that are original or not in the script.
- Its relationship to the brand image you want to convey: don’t forget that your chatbot is a full-fledged representative of your company.
With a true personality, adapted to the users’ expectations, your chatbot has better odds of inspiring trust, convincing and… transforming!
6) What happens when your chatbot doesn’t know?
Your chatbot has a purpose: to automatically provide responses to user inquiries. However, it doesn’t always know how to help the user for a variety of reasons, including:
- Lack of knowledge;
- No connection to personal information;
- Poorly formulated question
The most important thing is to guarantee that the user’s request will be handled. This is why the chatbot is trained to “step down” to its human colleagues by offering to open a chat or telephone conversation or by creating a help ticket in the client’s IS.
7) What solutions are available for creating your chatbot
With the explosion of uses for chatbots, a plethora of solutions have been developed. It isn’t easy to find your bearings! The solutions are divided into three major families:
- The APIs of the American internet giants (the Big Five). These solutions, which are basically provided in SaaS mode from the United States, allow users to develop their own chatbot. In their “company” versions, these solutions, with their powerful capacity for calculation and innovation, require development skills and significant investments. Furthermore, you’ll have no control over the use of information, as it will not be your property.
- Personalized solutions proposed by startups. Specific chatbot developments are launched every day: for HR, for IT support, for online sales, etc. These solutions are proposed by startups and are generally supported by the APIs of the Big Five (Google, Apple, Facebook, Amazon and Microsoft). They address a precise need. More integrated, these solutions generally include professional support and guidance: configuration is delegated to a specialized third-party team. However, they also have a limited level of technological independence, leaving little or no room for personalization.
- A general platform dedicated to chatbots for company use, like Living Actor. This solution brings together all the elements that are necessary to bring a chatbot to life, regardless of its role (sales, support, recruitment, etc.). It relies on proprietary developments, but also on the latest innovations in the market. Most importantly, it provides a console for creation and administration of intuitive knowledge, enabling professional teams to manage their chatbot directly. Hyper-evolutionary, it can be connected to unlimited information systems and useful web services in order to carry out its missions.
If you have to take one thing away from this synopsis of the questions to ask yourself before deploying a chatbot, it’s that the chatbot should constitute a genuine company vision. And as such, it merits your time. In particular, you must ensure the total involvement of all concerned personnel. Would you like guidance and support in this challenge? Living Actor can help you structure your reflections and communications and can guide you with internal training to help navigate change management. Contact our teams!
This post is also available in: FRENCH