Start with voice in 4 smooth steps

Create in a few steps a full-fledged voice assistant to represent your brand in this new world.


At Merkle Netherlands we are working on innovation. How can we stay innovative and anticipate on the latest technologies, so that we can advise and amaze our customers? By bringing together the various expertise’s within Merkle Netherlands, we can create the most interesting things for our customers.

Oxyma and aFrogleap, both part of Merkle, work together to expand their voice services. At the moment we are, in co-operation with various brands, working on innovation sprints and building voice interfaces. It’s time to share our knowledge about this! Innovation is an important aspect for almost any organization, but how can you stay ahead? And what can you do with voice as a brand? In this second article of our series on voice interfaces, we show you how you can start using this in just four smooth steps.

Start with voice? It’s quite easy to start, so they say. This is true if you want that the voice assistant to respond at programmed question. But a complete voice assistant who can represent your brand in this new world is a completely different league.

A big challenge for especially larger organizations is the speed by which they can respond to developments in the market. For as far as we can see now (based on the figures in the American and English markets), voice seems to be heading for a new adaptation record. So an early and good start is essential for success in the longer term.

voice adoptie curve

Voice Go-to-market framework

But how do you start with voice and what do you need for it? To answer this question, Merkle Netherlands has drawn up a Voice Go-To-Market framework that we will explain below.

voice-go-to-market framework

First of all, it is important to mention that working with voice is a repeating process that will continuously be improved by learning and managing, with no such thing as ‘first time right’. “No one knows where the f*ck this is going” said voice-guru Ben Sauer during his Keynote on Open Voice earlier this year. Flexibility is a must for everyone who starts.

 Step 1: Mindset, Knowledge & Focus

At the beginning of the process it is important to take, the team that starts working with voice, into this new world. Voice requires a completely different way of thinking. Consider, for example, the differences between a GUI (graphical user interface) and VUI (voice user interface), whereby old habits may have to be completely unlearned or adjusted. It is essential for the team to understand more about the backgrounds and the way of thinking around voice. Don’t start if the team doesn’t have the same level of knowledge, it will save you (and the team) frustration in the long term.

 Step 2: Ideas, Design and Validation

In the next step, usually within a 2-day workshop, the team starts to identify an idea for voice that has potential and adds value to the experience. There are a number of important variables that need to be considered. Keep especially the situation of the user in mind. The added value that voice offers and the reason that makes the user happy. There are a lot of ways to come up with ideas and to prioritize them. In the image below you can see an example of the post-it method we often use:

Post-it sessie

When the idea is set, the most important work begins, which is creating the design of the conversation. It is essential to start very small. Conversations or decision trees have the tendency to develop fast to a complex jungle of choices, answers and workarounds of which the user in particular will be the victim of. So try to undress the idea as far as possible and build up the conversation from that situation. You will see that during validation with real users (which you prefer to do continuously), even the most simple conversations can go very unexpectedly. “Fail gracefully” is the golden tip!

Step 3: Technical Design & MVP

Of course, during the process of designing the conversation, technical know-how is needed. Choose a combination of people who know what the existing landscape looks like and know how voice technology works. This is a perfect combination because voice will often reuse 80% of the existing infrastructure in order to achieve good experiences. You can think of the API landscape, CRM tooling, but also other existing tooling that can easily be connected. Keep in mind the necessary back-end work when the voice assistant expects to perform better in aspects of personalization and recognition.

The continuous collaboration between creative and technical people is important. Using the same words or answers could lose during the technical process, while one simple word can make a difference of day and night. This can ensure that the user experiences are no longer optimal because of technical choices. By constantly having a creative in the developers room, this problem will be avoided. This may sound inefficient but it is definitely a profitable tip.

Step 4: Go-live & Continue releases

Of course the assistant is tested extensively, after which you can work to the Go-live. Do not think too easily about this. Logically, for example Google has set up a complete acceptation process to ensure that the quality of the actions are high. In this context other things like, privacy statements, back-end calls, etc. must be completely in order. Keep this in mind, so that the team will not be surprised.

Before we described that the idea should be set up easily, after which the voice assistant can be expanded with an innumerable amount of functions. This also means that in the beginnning many good ideas will get lost because they might be to complex. Do not worry, make sure you keep a backlog and keep complementing it. Plan a release rhythm in which you want to answer new questions and add nice features.

In this way, the team has already gained more experience with the “relatively simple” idea and so the solvability has grown for the more complex components. In addition, the voice assistant remains interesting for the user as it becomes more complete with each release. Build a good content strategy around this and you will have a gold mine.

This article is written by Wouter Hosman (head of innovation, Oxyma – a Merkle company) and Sam Warnaars (head of innovation, aFrogleap – a Merkle company). This is the second article in a series about Voice interfaces. You can read the first article here. Sam Warnaars is also co-host of a voicecast, explore it on BNR.

Leave a Reply

Your email address will not be published. Required fields are marked *