Bots offer a new interface to an already established human interaction:
the conversation

Conversational interfaces are introducing a paradigm shift

User Interfaces developed from batch processing of punch cards and command line interfaces to graphical interfaces using metaphors like the desktop and dustbin. The rebirth of artificial intelligence, advances in language technologies and the emergence of the semantic web are the groundwork for a new kind of user interface: the conversational interface.

Conversational interfaces are a paradigm shift in the way users can interact with a system. Instead of communicating with a computer on its own inhuman terms – by clicking on icons and entering syntax-specific commands – users can now interact with the system on their own terms, by just telling it what to do.

Spread the “word”, check your data and level up your recommendation system

Gartner predicts that bots take over and that consumers favor voice and visual search. This means that the way customers will access your data and services is changing, providing you with a couple of challenges your company needs to face:

  1. 1. The interaction shifts from push to pull
    You no longer provide a monthly email with the average spending on some categories, customers just say “what’s my average spending on groceries in the last 6 months” and will be given an answer.
  2. 2. Your bot must be context aware
    While booking a flight you need to be able to also answer the question about the status of an already booked flight without losing the context and without losing the customer during the conversation. You also need to be aware about all interactions your customer has done with you.
  3. 3. You must have the data to answer the questions
    When using voice, customers tend to be more precise than when typing the search terms. You must have the data to handle those questions. Visual search can be of help here e.g. when your customer is searching for a "yellow shirt with a black logo on the left chest".
  4. 4. Customers expect answers right to the point
    When asking a chatbot to search for some product – especially in a voice only scenario - you must provide a close to perfect match search result within your answer. You cannot read more than 3 search results to the customer or you will lose. Therefore, you need to level up recommendation system as well.

Start small, but get going

Major internet companies like Google, Amazon and Microsoft are working on the democratization of artificial intelligence, providing toolkits to build a chatbot using Natural Language Processing (NLP) and Natural Language Understanding (NLU). These tools make it easy to get started with a chatbot proof of concept. But once you have the proof what are the next steps? What needs to be done to have a production ready solution?

Questions you need to answer are:

  • How to onboard customers to a new chatbot?
  • How to guide them through the process?
  • How to handle context switches during the conversation?
  • How to establish a long-lasting customer relationship through the chatbot?
  • Where do I get the data from?
  • How do I need to orchestrate my services and systems?
  • What is the roadmap?

Reply Robotics for Customers framework will help you answer all these questions by considering all the relevant aspects. It not only helps you to see it from the customer centric design perspective but also helps you to build a chatbot with personality by design.

Together with a recommendation system, chatbots can provide a tremendous improvement in enhancing existing touchpoints or even establish new ones. The Reply Robotics for Customers framework combines both to help you craft the best solution for you and your customers.

But the framework not only provides you with the right design principles and the best technology it also provides you with a development approach focusing on immediate results.

Starting with a design thinking workshop we help Reply customers to identify possible areas and scenarios for chatbots and conversational systems. This can be a chatbot in a fridge, a smart watch or an IT helpdesk bot accessible via your company’s Slack channels.

With an agile approach, we start building a minimum viable chatbot which allows our customers and the consumers to get their hands-on experience as soon as possible. Based on the customer’s existing IT landscape and background we carefully select the technology and tools to build the chatbot and the services needed to fulfill consumer requests. We assess the existing service and data and draw a plan how to step by step integrate the chatbot in the existing landscape.

Reply Robotics for Customers already demonstrated how conversational systems and recommendation systems can be introduced in various scenarios across different industries.

The important message here is: While everybody is talking about chatbots, it is still a new technology and is important for you as a company to learn about the possibilities and limitations together with your customers. You need to start and involve your consumers: Reply Robotics for Customers framework guides you along the way.