What does a conversational strategist actually do?

What do you actually do?

When I first started at VERSA as a conversational strategist, the voice design industry did not exist. Neither Google or Amazon had any smart speaker presence within Australia, and had not revealed any immediate plans to. I had to learn from overseas specialists, and build a role in anticipation of a new industry that may or may not come to the country.

With the industry evolving so rapidly, the day to day duties of a conversational strategist are constantly changing. One day could be spent educating clients and leading a workshop, another spent knee deep in flowchart design trying to find the best way of constructing a conversation.

The developing smart speaker industry has created a need for experts to work with clients on how to apply the technology to their markets

On yet another, I could be wading through code, trying to debug it and figure out why the voice experience (VX) isn’t working as intended. The other side to my job is sitting down to brainstorm the different ways a user can book a car via voice (currently 80,000+ users for GoGet). The role is ever-changing and new tools are constantly being added, which means always having to adopt and adapt.

The difference between a conversational strategist and a conversation designer is that you also need to have a solid grasp on technology and development. A good conversational strategist must design around the platform’s limitations and quirks and answer the pointy questions that come with conversation experience, such as: What region are we creating the experience(s) for? What platform(s) are we utilising? What natural language processing tool is best? Are there any new features recently added that will benefit the customer? What current restrictions do we have and how do we get around them?

I strive to craft the most engaging conversation for humans. However, I must also be able to break this conversation down into its component parts and look at it like a machine. Most of our everyday conversations can be broken down like this; everything we say and reply to is based on conditions from the person we’re speaking to and our environment.

Most of my work is based on assumptions. How do I think a user will respond to this? Why do I think they will respond that way? Of course, these assumptions then need to be proved with user testing, which can either lead to a celebratory drink after work, or going back to the drawing board!

What does a good working day look like?

The best days are when I’m right about those assumptions; the phrases I think users will use to invoke the experience are the ones they actually use, the user can complete tasks without any friction, and the voice assistant understands 100% of the time. The best parts are when users try to say something outlandish, then look at me, surprised, when they realise I’ve actually constructed a response for it.

There are also days where my team and I are the first to accomplish something in the space, such as ordering the country’s first pizza via conversational interface.

What’s the most difficult part of your job?

The flipside of these successes is that conversation design feels like a job that can never be fully completed. No matter how hard I work, I will always be constricted by reality. If I’m creating an experience to book a holiday, users shouldn’t expect to be able to ask for a pizza, yet they probably will. You have to draw a line somewhere, and that decision is difficult to make.

What are your KPIs and how do you ensure you meet them?

KPIs in conversational strategy are non-traditional. Instead, I benchmark my success on the ability to create engaging, efficient and innovate conversation. As long as we are making our designs better than our last ones, and trying every bit of technology we can to make it a better experience for the customer, I think I’m doing my job right.

One of the best metrics we use to gauge this is repeat usage. A large part, however, still comes down to creating an engaging conversation while meeting budgets and deadlines. A good conversation can still be simple and lean, and a conversation that may only be three steps long can be more difficult to write than one that is longer.

How does your role keep you on your toes?

When dealing with any new technology, a lot of pressure can come with dealing with the unknown. When faced with a new problem, there’s a high chance that I’m one of the few people that have encountered this before. While the problem-solving aspect of this is great, it can be difficult having to be the first person to make the call on a lot of solutions. We’re constantly finding limitations to the platforms and language models in general, and having to design clever ways around them.

The most stressful part is definitely determining if it’s your design or implementation that is failing, or the voice assistant simply not understanding you correctly. Many hours have been spent debugging non-existent issues, only to discover that it was something as simple as testing while the device was muted, or not set to the correct language.

Although the job can be pretty broad at the moment, once the world of voice experience and conversation design starts to mature, I think a lot of these tasks will be turned into highly specialised roles. Being one of the first conversational strategists is amazing, and I can’t wait to see how this role develops.

Tyler Hamilton is a conversational strategist at VERSA.