With the advent of the moderns generative models, human-machine interaction has undergone a significant transformation thanks to the spread of so-called conversational interfaces. A platform like Voiceflow revolutionizes the way they can be designed and developed voice assistants and chatbots, allowing home users, professionals and businesses to create intuitive and personalized user experiences, without the need for advanced programming skills.
Voiceflow represents a real turning point in the field of design and construction conversational interfaces. The ultimate goal of the platform’s developers is to make the creation of voice assistants and chatbots accessible to everyone, thanks to the use of a visual environment drag-and-drop that allows users to design, test and implement complex conversation flows with ease.
Even in the completely free profile, Voiceflow offers interesting up to 100.000 token usable monthly with all major Large Language Models (LLMs).
What it is and how to create a chatbot or voice assistant with Voiceflow
You need one custom chatbot for your business? Voiceflow allows you to do it in minutes, with the possibility of obtaining the code to integrate it directly on sites and Web applications.
To get started, just click New agent, at the top right of the Voiceflow web interface, then choose whether you’re interested in developing a chatbot or voice assistant. Selecting one or the other item allows you to access very different functional blocks: it is therefore good to be clear from the beginning what you want to achieve.
Using a series of blocks and stepusers can define the dialogue between the user and the voice assistant. Voiceflow helps you manage information-based decisions user input or on other variables, integrate external API calls and obtain real-time data from any type of source.
Look at the example in the figure. The various blocks make up an intelligent tool that allows the user to receive practical information for planning a trip. This is the conversational travel assistant template adapted for use in the Europen language.
With Voiceflow, you can capture user input and pass it to the generative models (LLMs) of your choice. Simply drag from the main menu on the left a AI block then set specific preferences.
The advantage of a platform like Voiceflow is that every input or response can be stored within a variable. Its name can be chosen arbitrarily then retrieved later within any point of the workflow.
Company data and remote addresses to train the model, with one click
Interestingly, by choosing the option Knowledge Baseyou can switch toartificial intelligence of Voiceflow a set of URLs and business documents that will be used for model training activities. What you get is a chatbot or a voice assistant “tailored” based on the needs of each individual company.
By doing so, by building the workflow with the most appropriate blocks, it is possible to ensure that the web application generated with Voiceflow responds to specific questions of the activities carried out by the company, on its products, on their characteristics, on the salient aspects of each business reality, without inconveniencing generalist LLMs. Not very suitable in some application fields.
The option Knowledge Base effectively opens the doors to the RAG approach (Retrieval-Augmented Generation), a model that combines retrieval and generation techniques to produce content in line with company expectations.
Built-in test and debug modes
Before being implemented, each project can be tested directly within the platform. You can simulate conversations and evaluate the behavior of the voice assistant or chatbot in various scenarios, to identify and correct any errors or flow problems.
By clicking on the “Play” triangle on the left of the block Start green, you can perform the set tasks one after the other. However, each user is free to try the operation of individual blocks: each of them contains a small “Play” button.
The blocks must obviously be interconnected with each other to define the operations to be performed in succession.
How to use the functions
By clicking on Library, API Call Example, it is possible to have Voiceflow make GET or POST calls to third-party or own services, managed on dedicated servers and cloud systems. Using variables, you can define the data to transmit to the API and manage the output received, diversifying the behavior of the flow in case of success or failure of the operation.
The functions of Voiceflow are a truly precious gem: although the idea is to create conversational bots without having to do any coding, developers can insert routine articulated systems that receive input values through variables and return them as output.
Examples of integrations already usable with Voiceflow, they are available in the appropriate area of the site. Unfortunately, the solution that allows Voiceflow to communicate with i vocal modelsalso customized, available on Eleven Labs It doesn’t seem to be working at the moment. Not even by entering l’API key correct Eleven Labs, parameters such as the text to read and the speech model identifier to use.
It’s a shame because creating a bridge between Voiceflow and Eleven Labs would mean having the possibility of activating a motor text-to-speech (TTS), capable of reading any text generated by the most well-known LLMs in an effective and engaging way (also in Europen).
On the other hand, the function Voice Lab by Eleven Labs allows you to create a voice model from voice recording of any person (make sure you have collected their release to proceed in this direction). A few seconds of speech are enough to obtain a model that may be interoperable via API with Voiceflow.
Conclusions
Voiceflow represents a significant innovation in the field of computer design conversational interfacesallowing individuals and businesses to create rich user experiences without the need for advanced programming skills.
With its intuitive web interface, the ability to integrate external services and support for real-time collaboration (multiple users on the same team can contribute to developing a workflow), Voiceflow is positioned as an indispensable resource for anyone wishing to capitalize on potential of modern generative models.
Professionals can use Voiceflow for creation of voice assistants useful for managing customer requests and providing real-time support. But the platform can also be used to support student learning through voice interactions, implement voice assistants to guide users through the purchasing process on shopping sites e-commerce and to provide product information.
We also suggest not losing an eye on the Voiceflow changelog: it helps you discover all the features gradually added to the platform.