Where do you stand regarding voice technology? Connected speakers, VoiceBot or voice assistants are becoming more and more present into the economic landscape, becoming almost tried and tested products. Voice has many advantages for both the individual and the company, which we have discussed at length in our blog. However, one topic is not covered: how to manage voice technology integration ? Imagine that in only 3 steps, it is very easy to cover this process, we explain everything.
Getting started by identifying needs and opportunities
As in any project, the first step is to carry out a diagnosis of the situation. In this sense we intend to identify within the company which activity and more particularly which type of process is conducive to voice. Indeed, it is essential to use voice to create meaningful interactions that improve, not interrupt, the lives of people, customers or employees.
Despite the ideas that have been developed, voice is not necessarily suitable for all uses: sometimes it will simply replace a manual action with no performance benefit… On the other hand, in a majority of cases, well thought out, voice combined with intelligent integrations is a generator of value. For example, a very simple illustration, between switching on the light by voice and via a switch, the value is low, but programming a complete scenario for the luminaires allows the use to be moved to another stage.
Thus, it is necessary to identify the different flows and the elements that are close to them in order to have a global vision of the situation and to estimate the benefits of the project. This first step is therefore the foundation of the project, and especially the structure necessary to establish the specifications that we will see in the next step.
Structuring the voice project by setting up a set of specifications
Once the environment has been studied and targeted, it is now necessary to review the various desired functionalities, as well as to estimate their feasibility. To do this, it is very simple, the creation of a specification (common to many “project” operations) is highly recommended. On the one hand you are assured of the content and structure of the project, on the other hand it allows you to have a reference document with your technical partners (in case you use external services, which is usually the case if you do not have the resources).
The elements that may be taken into account in the specifications may be the following:
- The specifications concerning the material and/or the equipment foreseen in the use of the vocal tool (resistance, lifespan, adaptation, nature, memory etc…)
- Parameters for storing and sending data (planned connectivity, offline operation, synchronization, planned servers, communication etc.)
- Necessary data collected (if required) during use (customer habits, product list, references, machine orders, etc.).
- How to test the solution with the criteria to be validated.
- The expectations concerning the user interface (UX/UI) and also the Human-Machine Interaction.
- A description of the different functionalities, sometimes grouped in modules, which condition the project’s progress.
- Etc…
Of course this list is not exhaustive and it is on the last point that we would like to draw your attention. Indeed, drawing up the list of functionalities is not always easy because it is necessary to have a transversal strategic vision in order to fulfill all the expectations of the project.
The functionalities, often referred to as “use case“, fall under the notions of design and VUX (Voice User Experience). The aim is to combine ease of use, satisfactory results in relation to the expected benefits and continuity of the Human-Computer interaction (using voice from start to finish, without having to vary the types of interactions).
Once the specifications have been drawn up and the project management methodology structured, development can take place, preceding the test phase before complete voice technology integration.
Securing a result with test and integration
Here we are, your voice technology integration and projects are done in their first version, but there’s no guarantee that everything will work perfectly the first time. Then comes the test phase for several reasons: to make sure the system works and to identify possible areas for improvement. To do this, test protocols have to be put in place to ensure the rigour of the tests.
Here again, several elements must be taken into account:
- Who is testing? Indeed, it is important to set up a defined population of testers with precise criteria, so as not to bias the results. It can be customers, employees outside the project, etc…
- How long does the test take? This duration may vary according to the quality requirements but also the relevance of the tests according to their complexity.
- What test criteria should be established? It is necessary here to group the different tests by theme and to associate, for each unit test, a technical or experiential validation.
Again, the list here does not pretend to be complete because each project is by nature different from another. However, it does give you again an idea and leads to steer yours in the right direction.
Once the tests are completed, the solution is finally operational and ready to be deployed. To do this, and depending on the type of technology (Cloud or embedded) the procedures will differ. If it is a Cloud solution, the deployment will be facilitated because it is dematerialized. On the other hand, for an embedded technology, especially in hardware, the installation and logistics measures will have to be taken into account more widely.
There you are, your voice AI solution is normally ready to use. Your users, whether they are your customers or your employees, are able to use it. But your work is not finished yet. As with any innovation process, it is essential to set up user feedback systems in order to set in motion the continuous improvement of the solution.
