TRAMONTINA

AWARD-WINNING PROJECT IN THE INNOVATION SECTOR

Tramontina's voice assistant aimed to use AI to help children with the task of choosing a gift for Father's Day integrated into Tramontina e-commerce

20 Million

Campaign reach

+25%

Increased sales in the Cutlery | Barbecue sector

+35%

Increase in interactions on social media

THE CHALLENGE

How we could develop a tool a practical, with a simple user flow and that could optimize the end-to-end user experience

Every year is a great challenge for children to get the right present, which is why Tramontina decided to lend a hand and create Tramontina's Father's Day Assistant. A Google voice assistant, integrated into Tramontina's e-commerce.

How we could develop a practical tool, with a simple flow and that could optimize the end-to-end user experience, from identifying the parent's profile, through choosing the product and ending in an uncomplicated checkout.

DISCOVERY

Optimize time in search of content already produced by Tramontina

Due to the tight schedule to develop all the stages, we did some internal research to understand how we could be guided by content already produced by tramontina, and thus gain “some time”, and collect inputs to develop a user´s flow first idea .

MOODBOARD TIPS MAKE TRAMONTINA BEAUTIFUL

On Tramontina blog, where was possible to find a series of tips "to make beautiful", then we came across a variety of content related to do it yourself, gastronomy, decoration, barbecue, among others.

IDEATION

During the co-creation workshop, the team and stakeholders had the insight to create the father's favorite spot at home

In an inspiring co-creation workshop, the team and stakeholders came together to brainstorm and envision an exciting idea: the ultimate favorite spot in the home for dads. It was a moment of collaboration that sparked creativity and passion, leading to a wonderful concept that celebrates a father’s haven!

HOUSE SPOT AS SEGMENTATION

DECISION TREE

After the dynamics, we gathered insights and were able to understand how to evolve to a more assertive interface prototype within the process

In conversational design, we aim to develop a humanized language, based on conversations and human experiences, our goal was to make more natural and fluid interactions.

In this type of design, the development of the decision tree is a fundamental part of the process, through it the stakeholders and the UX Writing team remained connected, developing together the entire conversational structure of the solution.

WIREFRAME DECISION TREE

UX WRITING AND CO-CREATION

Co-creating the decision tree and collecting feedback quickly, thus optimizing the UX Writing process

INTERACTIVE TEXT FLOW BOARD

We chose to use figma jam, as previously mentioned. With this tool it was possible to keep the UX writing team and stakeholders in direct contact. We shared the figma board, so the client could make comments about the conversational flow, and adjustments were quickly made by the team, saving time and without the need for meetings related to this type of decision.

NLP Technical Specificities

Intent Recognition: The assistant used Natural Language Processing (NLP) to interpret human language and understand the specific "intent" behind a user's search for a gift

Conversational Flow: It was designed with context-awareness, acting as a conversational guide rather than just a search bar, which is essential for voice-first interfaces

E-commerce Integration: The AI was directly connected to the e-commerce engine to provide real-time product recommendations and purchase paths based on the conversation.

Context-Awareness: Ensuring the AI maintained the thread of the conversation by remembering previous user inputs throughout the session.

Error Recovery: Designing the system to handle misunderstandings gracefully, preventing users from getting stuck or leaving the purchase funnel.

USER TESTING

After the process of building and approving the conversational flows, it's time to test whether the interactions are simple and natural.

After the of construction and approval phasesof the conversational flows, we started testing the text and intonation of words. We opted for the - Google cloud - text to speech.

These tests helped us understand how the pronunciation was in the voice of the Google assistant, and in this way, adjust words and terms. Using good conversational design practices we created more humanized interactions, which was the focus at this stage of the process.

Testing the flows with users helped us fine-tune the final phase of the project.