NiaKai

Zellaris was not meant to be another call-recording tool that simply generated transcripts and summaries.

The product had a much larger vision.

It needed to help sales representatives:

• Prepare before a meeting
• Understand prospects and accounts
• Receive assistance during live calls
• Analyse conversations afterwards
• Decide what to do next
• Practise and improve over time

The challenge was that all of these capabilities could quickly make the product feel overwhelming.

We needed to create a system that felt connected and easy to understand — not a collection of AI features placed inside the same dashboard.

Before designing individual screens, we organised the product around the complete sales conversation lifecycle:

Prepare → Engage → Review → Train

This became the foundation for the product.

The Prepare experience gives representatives everything they need before entering a conversation.

It includes campaigns, prospects, account intelligence, buying committees, deals and upcoming meetings.

Instead of displaying contacts as simple CRM records, account pages show the relationships between different stakeholders — such as champions, influencers, decision-makers and blockers.

The goal was to help salespeople understand the account, not just view its data.

The Engage module supports representatives before and during live meetings.

Before joining a call, they receive a brief containing relevant company information, previous interactions, possible objections, important talking points and recommended questions.

During the meeting, Zellaris provides real-time assistance alongside the transcript.

The main design challenge here was avoiding distraction. The product needed to remain useful without competing with the actual conversation.

We kept the live interface calm, readable and focused only on information that could help at that exact moment.

After a meeting, Zellaris analyses the conversation and helps the representative understand what happened.

Instead of showing unexplained AI scores, the review connects insights to real moments from the call.

Users can see:

• The outcome of the conversation
• Important questions and objections
• Customer interest or hesitation
• Commitments made during the call
• Recommended next actions
• Areas where the representative could improve

This information also contributes to wider business intelligence, helping leaders identify patterns across campaigns, personas and team performance.

The Train module allows salespeople to practise conversations with simulated prospects.

They can prepare for objections, different personalities and difficult sales situations before entering a real meeting.

This completed the product loop: users could prepare, have the conversation, review their performance and practise before the next one.

Zellaris contains a lot of information, so the biggest design decision was often what not to show.

Every screen was designed around a simple question:

What decision does the user need to make here?

This helped us prioritise the right information and avoid filling the interface with unnecessary AI-generated insights.

The visual system focused on:

• Clear hierarchy
• Calm and structured layouts
• Reusable components
• Consistent interaction patterns
• Simple explanations around complex insights
• Enough flexibility to support different workflows

We wanted the product to feel intelligent and enterprise-ready without relying on glowing gradients, excessive animations or the usual visual clichés associated with AI products.

As the engagement continued, my role expanded beyond the application.

I also worked on NiaKai’s website, pitch decks, event assets and overall product positioning.

The website was rebuilt around the conversation lifecycle rather than a generic feature grid.

We also created a clearer comparison between traditional conversation intelligence products and Zellaris.

Traditional tools usually begin after the call. Zellaris is designed to help before, during and after it.

That became the central story across the website, decks and marketing materials.

The project gave NiaKai:

• A clearer product architecture
• Connected workflows across the sales lifecycle
• A flexible design system
• A more distinct and credible visual identity
• Stronger product positioning
• A website aligned with the actual product
• Consistency across the product, decks and marketing assets

The most valuable result was not one individual screen.

It was creating a connected experience where the product, website and brand all communicated the same idea:

Zellaris does not simply tell salespeople what happened. It helps them decide what to do before, during and after every conversation.

NiaKai has been one of my longest and most collaborative projects.

It showed me that designing an AI product is rarely about displaying more intelligence.

It is about deciding when that intelligence should appear, how much context the user needs and how an insight can lead to a meaningful action.

What started as an MVP redesign eventually became an opportunity to help shape how the entire company looks, works and explains itself.