Talkpush

Defining, designing and developing a new and improved candidate experience for a recruitment automation company. Present in more than 10 countries around the world, each of the chatbots were designed with the company's brand and local behaviors in mind.

What I learned

Working in-house at Talkpush helped me develop as a conversation designer and then product manager. I’ve gained valuable experience working on various social media platforms with weekly release cycles learning the nuances of each under time pressure. Working at this pace helped me focus on creating clear and fully defined specifications, ensuring the design is fully considered before going to development teams. Getting into a good rhythm of user research was important in this role, finding quick and simple ways to engage with our audience and share insights with the wider team.

Conversation Design at Scale

Using Dialogflow, we added NLP into the bot conversations. I also created a standard design framework for the chatbot building process to be scalable and meet business needs. We implemented a standard user flow for the launch of the compan's branded bot "Stanley" based on conversation design best practices, and data collected from past conversations.

Persona design

I then developed different chatbot personas for Talkpush to offer its clients based on their brand profile, tone of voice, and company culture. With the rest of the product and engineering team, we looked at improving the product marketing on the flow as well as each of the steps to help conversion at each stage of the candidate journey.

Multichannel and Multilingual experiences

Once a standard framework was introduced, this was rolled out in various platforms (WhatsApp, SMS, websites, and Facebook Messenger), and in multiple languages to meet the needs of our clients and their candidates.

Depending on the platform used and client needs, we configured each bot to initiate every conversation in a custom way to increase conversion rates. Fixes usually included improvements in conversational pathways and error handling, prompt writing, training data, and content management. As Product manager, I developed a user-centric process with top-notch documentation: comprehensive flow diagrams, personality and tone style guides, prompts, training data, testing protocols, and user-testing reports.

Data-driven design