Using Voiceflow to Run More Efficient, Inclusive AI Design Reviews

Who: Conversation design team responsible for the mobile banking assistant at a Global Tier 1 Investment Bank

Before: Static tools and a siloed workflow slowed the review-and-approval process and required extra work and handholding from conversation designers.

After: The team uses Voiceflow’s project canvas and shareable prototypes to run design reviews that take less time, include more stakeholders, and produce better feedback.

Static Tools, Time-Consuming Process

For this conversation design team, design reviews are a valuable way to get everyone on the same page, tap into the collective expertise of colleagues, and ensure a high-quality customer experience.

But the team struggled with multiple challenges in the design review process, according to Gina Riley, VP, User Experience and Conversation Design Lead. As Gina recalls, “Everything was so manual and took a really, really long time.”

For one thing, the design team used to document their designs in multiple static tools—Figma, Gliffy, Visio, and more. This made it tough for designers to account for every customer scenario and harder still for reviewers to envision how the entire conversational experience played out.

As a result, design reviews required lots of handholding. Designers walked through static visual mockups in Design Forum meetings that stretched to 90 minutes or more. Faced with “millions of questions,” designers had to remember and explain their design choices, all while capturing reviewers’ comments.

What’s more, feedback came very late in the design phase. This led to a lot of tedious rework with updates to numerous files and screenshots. Schedules took a hit as designers returned to projects they thought were nearly complete. “It really messed up your timetable,” remembers Gina.

To top it off, “reviewers didn’t really see all the hard work that went into our designs,” adds Gina, which was discouraging for a team committed to delivering the best customer experiences possible.

Resetting the Design Review Process

As her team and its workload grew, Gina recognized that the old way of conducting design reviews wasn’t sustainable—especially when product leaders wanted to release new features faster.

Fortunately, the bank had adopted Voiceflow to expand the use of conversational AI in its digital assistant. Gina used Voiceflow to reimagine not only how her team designs conversational experiences, but also how they solicit input from product owners, developers, and other designers.

The team has now moved away from long, in-person meetings to asynchronous design reviews that save time, reduce frustration, and result in more meaningful feedback.

The Power of Prototypes

One of the most impactful changes to the team’s review process is the inclusion of prototypes.

When a design is ready for review, designers generate a high-fidelity prototype of the project—a step that takes just a single click and a few minutes. Then they share the prototype and Voiceflow project with reviewers, who have a week to provide feedback.

Live prototypes have brought clarity and efficiency to design reviews. Designers don’t have to walk through every step of a conversation flow because reviewers can simulate the customer experience, using the latest interaction model, right on their mobile phones.

Reviewers have fewer questions about how “things will work,” and designers have more time to create great conversational experiences.

Prototyping has also been a valuable tool for securing stakeholder buy-in. It’s fostered a deeper appreciation for the practice of conversation design, especially among senior product leaders. “Voiceflow has really elevated us as a team,” says Gina.

One Central Source of Truth

In addition to interacting with the live prototype, reviewers can look “under the hood” by accessing the project in Voiceflow.

Here they can view every aspect of the conversational experience—entities, intents, utterances, logic, etc.—all in one place. “The reviewer can actually see the work that went into the design, plus every single piece of the conversation and how it works,” notes Gina.

Reviewers use Voiceflow’s commenting capabilities to ask questions and leave feedback. Reviewers can comment anywhere on the canvas, and tagged teammates are notified by email. “It takes minutes because you just comment, and the designer will get back to you,” says Gina. “It’s pretty immediate.”

If a change is needed, designers quickly update the design by adding, moving, or deleting elements on Voiceflow’s drag-and-drop canvas.

Tighter Collaboration with Developers

Historically, conversation designers had little to no interaction with developers. After a design was approved, the designer uploaded files to JIRA in a classic over-the-fence handoff. The two functions were so siloed that designers didn’t see the results of their work until a new feature was in production.

Gina plans to use the new design review process to build a better working relationship with development. “With Voiceflow, we can bring developers in much earlier and actually collaborate with them,” she says.

Developers will be invited to interact with the prototype and project canvas so they can ask questions and assess the technical feasibility of designs. Identifying issues during design will help avoid unwelcome surprises and delays during the build phase.

Ultimately, Gina expects 25 to 30 people—including conversation designers, content managers, product owners, developers, data scientists, user researchers, and legal staff—to review and collaborate on a project in Voiceflow.

Best Practices from this Conversational AI Team

  • Leave behind the constraints of legacy tools and ways of working. Don’t continue to design conversations the way you did in Figma, Gliffy, Excel, or other static tools. Voiceflow is purpose-built for conversation design; explore and take advantage of all its capabilities.
  • Replace meetings with asynchronous design reviews. Share the project canvas and use Voiceflow as the hub for collaboration and commenting. Reviewers can provide more focused feedback when given the time to explore all aspects of the conversation design.
  • Share high-fidelity prototypes with reviewers. There’s no better way for reviewers to understand and evaluate your designs than by taking them for a test run. Prototyping can uncover holes in conversation flows while promoting stakeholder buy-in.
  • Integrate developers into the design review process. Everyone saves time when developers can ask questions and iron out technical issues before handoff. Plus, collaborating earlier in the process helps break down functional silos.
  • Establish standard operating procedures (SOPs) to drive consistency across Voiceflow projects. This will help align and improve communication across the broader conversational AI team. SOPs also make it easier for designers to explain their design choices.
RECOMMENDED
square-image

How Trilogy automated 60% of their customer support in 12 weeks

RECOMMENDED RESOURCES
No items found.