It is not a coincidence if artificial intelligence rhymes with customer experience. High-speed internet and social networks have created new expectations in terms of daily interactions. The business world has followed the trend in order to answer the needs of a whole new generation of mobile-first consumers. Chatbots were born: the future of commerce would be conversational.

Nicolas Fayon is the founder and CEO of Heek, a website development tool using chatbot technology. It transforms the way SMBs build their presence and brand online with a genuine approach to the way websites should be created. As an entrepreneur and chatbot specialist, Nicolas is a natural fit to deliver key learnings on developing and making the most of AI and bots for your projects. Have you always wondered where to start to build your own chatbot? This article will definitely help you, with thoughtful insights and actionable tips. Here we go!

Interview with Nicolas Fayon, CEO at heel


The rise of the chatbot as a tool for support, redirection and even web development for individuals and professionals demonstrates the constant improvement in the quality of the interfaces available in terms of features, accuracy, and immediacy of execution.

The industry behemoths have clearly understood this, and are investing heavily in the design, sophistication, and training of their artificial intelligence (Alexa for Amazon, Siri for Apple for instance). Chatbots have become new standards in conversational platforms such as Messenger for BtoC and Slack for BtoB. The aim of all these tools? To facilitate tasks and shorten customer paths with one constant goal: optimizing added value.

By developing its own website designer and admin, Heek was able to capture one particular customer need: to have a presence on the internet without the usually required web development skills. The idea is to offer the simplest website builder in the world, empowered to offer the same creative service as an agency, but with a much greater capacity to manage requests, as well as uninterrupted responsiveness.

→ When building a chatbot, having a clear value proposition is a must-have to stand from the crowd.


When too many bots promise the impossible — and therefore, ultimately fail —, Nicolas stresses the need to fundamentally think the chatbot as a solution for well-defined needs, and not for anything and everything. For instance, Heek has been thought as a true web designer, autonomous when ordered, with a range of features adapted to users who are allergic to code, and foreign to CMS interfaces. Its interface is highly understandable and based on the most widespread means of communication in the world: the SMS.

As Nicolas explains, the use of a chatbot depends above all on its target, more precisely on their requirements and their relationship to the tool. He deplores the proliferation of badly positioned bots, which are detrimental to the average quality of the interfaces available to users, and consequently to the image of the chatbot. According to him, the spectrum of the bot should be as thin as possible, and its utility clear and proven. That is the best way to avoid extraordinary expectations and focus on appropriation, to allow users to extract the maximum added value by themselves. Creating a powerful and meaningful chatbot is all about precision, data, and strategy. Still, building a comprehensive tool remains key to its own improvement in time, as explained below.

→ To sum it up, there are three fundamental steps to take into consideration when drafting your chatbot strategy:

• Target: who will use it?
• Goal: what will they use it for?
• Mission: how will it help users?


The bot must not only be functional and adapted to its target, but also designed for the improvement of its own performance over requests and intentions. « Every startup evolves with its customers, and must develop a product that builds up at the same time as the company and its project. » As an illustration, Heek enables the refined design of original websites, resulting from a textual dialogue with the user, which is unique in essence. The tool fortifies its artificial intelligence with its own experience as a web designer and progressively absorbs additional resources (templates, images, etc.).

Enabled by the enrichment of data, this growth allows the product to be refined according to more than 1,500 professions, grouped into 40 industries with varied characteristics. Among these work fields, the chosen examples of yoga and fitness meet precisely the expectations of the Heek bot: a charted universe (showcase websites with basic features), with a rich imagery and message in terms of atmosphere, wording and visuals.

→ Product vision is essential when building your chatbot. The latter is likely to evolve (maybe event pivot), so you need to think ahead of the next steps while iterating quickly.

Preview of the chatbot website editor on Heek


By designing precisely targeted chatbots, the different industries will know how to make the most of always more effective and quickly actionable tools in replacement of a durable obstacle. Nicolas details the essential steps for the creation of a bot in the form of strategic guidelines.

• First, the concept must justify the very existence of the tool: is the chatbot the right solution to the existing issue? A support chatbot is always welcome, but it must above all help to apprehend the product, beyond the current hype of this kind of support on the internet.

• The second focus is the question of positioning and therefore of the audience, which will define the appropriate platform (rather Slack for B2B, Facebook for B2C for example), depending on the expected use and level of precision.

• Finally, the third step is more technical: it is a question of indexing the intentions, namely what will be datamined by the tool, the degree of customization and diversity of the suggestions which result from the richness of user profiles.

→ Problem-solving abilities, platform positioning, and customization features must be considered before starting your chatbot project.


The chatbot needs to be configured as a product, based on existing data access and expectations on the other side of the screen. Its mission is to cut the dispensable steps and avoid unnecessary questioning. Its implementation, in turn, depends on its good architecture, supposedly aimed at enhancing its training. Once the technical framework is finished and the service ready for implementation, the bot must be trained to improve the user experience over the intentions, and this depending on the language used, preferred choices, etc. It is this training strategy that will determine the future potential of the bot with respect to the requirements that are attached to it.

Eager to create your own chatbot? The selected tool will depend primarily on your ability to code. Still, many platforms allow the creation of chatbots without any development, such as DialogFlow for bots on Facebook, Telegram, Slack, or Pandorabots for a conversation bot built in a few minutes. Statbot is also a great tool, which gives numerous statistics in a few queries and emits great synthesis graphs.

In parallel, it is also possible to create a chatbot by coding it integrally, which would make you master of its evolution, and guardian of its possible flaws. Plus, new tools are coming in every day, and the chatbot platform industry is following the same dynamic as the website creation platforms; towards more autonomy, for extremely different expectations. Nicolas created by himself the Heek chatbot without using any platform, which would have required a constant interface with the publisher and content. Entirely developed in Javascript with Angular2, the Heek product is a proprietary bot built from scratch.


Basically, the chatbot is a fascinating chance to allow many more users to produce much more content, much faster, and to remain owners of their achievements. The commercial and functional potential of the tool for the B2C market is colossal, with the ability for users to take control of any maneuver, to submit any type of request, and to process complex data mining, on all media; an approach which would have been unimaginable before, and which is now completely revolutionary.

Once again, it is essential to underline the importance of setting each bot according to its functionality and target (mass/specific), to ensure the production of quality tools. The chatbot is above all a universal tool which have the potential to bring immediate added value with an outstanding reach, due to the platforms it relies on. Just think of the 1.3 billion users on Messenger.

Heek logo adhesive hooked up to their office's wall


→ In recent years, more and more companies have developed their own chatbots, willing to benefit from what can be a fantastic tool for their customers and workmates, with minimum costs of development and implementation.

→ This trend led to the mushrooming of thousands of chatbots on conversational platforms (Slack, Messenger, Whatsapp, etc.) and websites, although at uneven levels of quality and accuracy.

→ Nicolas from Heek reminds us how essential the reflection on the real mission of chatbots is. Businesses have a responsibility to create smart tools in order to revolutionize customer habits.

→ By building chatbots able to learn from the requests and intentions they receive and proceed, chatbots developers are working on transforming the way we communicate online.