Core
Experience

Core
Experience

Role & Responsibility

Role & Responsibility

PRODUCT DESIGN

PRODUCT DESIGN

Product direction, hands-on contribution & team management

Product direction, hands-on contribution & team management

PRODUCT RESEARCH

PRODUCT RESEARCH

E2E research collaboration - planning, execution & insight reporting

E2E research collaboration - planning, execution & insight reporting

PRODUCT STRATEGY

PRODUCT STRATEGY

Vision definition, strategy roadmap & stakeholder alignment

Vision definition, strategy roadmap & stakeholder alignment

Collaborators

Collaborators

Product Design (1 lead, 2 designers) · Product Lead · CEO · Engineering

Product Design (1 lead, 2 designers) · Product Lead · CEO · Engineering

Project Duration

Project Duration

12 months

12 months

As part-time Head of Design, I led a strategy and experience reset for an AI-powered B2B CRM. The redesign drastically cut user abandonment rate and significantly increase engagement rate, and helped win two new clients, both citing the redesigned experience as a deciding factor

As part-time Head of Design, I led a strategy and experience reset for an AI-powered B2B CRM. The redesign drastically cut user abandonment rate and significantly increase engagement rate, and helped win two new clients, both citing the redesigned experience as a deciding factor

Outcome

Outcome

61%

61%

Decrease in
abandonment

Decrease in abandonment

88%

88%

Core task
completion rate

Core task completion rate

37%

37%

Increase in product
usage frequency

Increase in product usage frequency

The situation

The situation

Kaboom is a B2B CRM that uses AI to aggregate internal and external data across prospects, accounts, and contacts, then turns that data into clear next-step recommendations. I joined two years after the company started.

Despite early clients, there was no clear product vision. Roadmap decisions were driven by client requests and founder intuition, which produced a fragmented product with too many loosely connected features. Adoption was low, abandonment was high, and the product was expensive to maintain and hard to understand.

Rather than keep patching features, I argued Kaboom needed a full reset of both strategy and experience, and I took ownership of guiding the team through it.

Kaboom is a B2B CRM that uses AI to aggregate internal and external data across prospects, accounts, and contacts, then turns that data into clear next-step recommendations. I joined two years after the company started.

Despite early clients, there was no clear product vision. Roadmap decisions were driven by client requests and founder intuition, which produced a fragmented product with too many loosely connected features. Adoption was low, abandonment was high, and the product was expensive to maintain and hard to understand.

Rather than keep patching features, I argued Kaboom needed a full reset of both strategy and experience, and I took ownership of guiding the team through it.

Kaboom's interface, before I joined.

Kaboom's interface, before I joined.

Finding the core

Finding the core

I started with client interviews. One pattern was consistent: users valued the core meeting experience (AI-assisted prep, note-taking, summaries, and follow-ups), but everything around it felt disconnected. AI was underused outside of meetings, which capped its perceived value, and users wanted reliable access across devices. I kept clients in the loop through the rest of the project so the direction stayed tied to real needs.

Working with stakeholders, I set three goals: prune the feature set down to what users actually rely on, define a clear core value proposition, and rebuild on a foundation that could scale.

I started with client interviews. One pattern was consistent: users valued the core meeting experience (AI-assisted prep, note-taking, summaries, and follow-ups), but everything around it felt disconnected. AI was underused outside of meetings, which capped its perceived value, and users wanted reliable access across devices. I kept clients in the loop through the rest of the project so the direction stayed tied to real needs.

Working with stakeholders, I set three goals: prune the feature set down to what users actually rely on, define a clear core value proposition, and rebuild on a foundation that could scale.

Excerpt from client interview summary.

Excerpt from client interview summary.

Resetting the strategy

Resetting the strategy

I facilitated vision-setting workshops to define the core value proposition and long-term direction, then audited the full feature set. Non-core, low-adoption features came out to cut complexity, with some parked for later. I also introduced an end-to-end architecture diagram that became the shared reference for product, design, and engineering, guiding both MVP scope and future expansion.

I facilitated vision-setting workshops to define the core value proposition and long-term direction, then audited the full feature set. Non-core, low-adoption features came out to cut complexity, with some parked for later. I also introduced an end-to-end architecture diagram that became the shared reference for product, design, and engineering, guiding both MVP scope and future expansion.

Kaboom's feature sets transformation.

Kaboom's feature sets transformation.

Architectural diagram for Kaboom's new core experience.

Architectural diagram for Kaboom's new core experience.

Leadership was hesitant about a design system, worried about scope and inclined to reuse the existing UI. I pushed for a systemized approach, showing how the existing inconsistencies would compound as the product scaled and drag down both usability and development speed. We aligned on a lean system of essential components with room to grow, and I refreshed typography and color to give it a coherent foundation.

Leadership was hesitant about a design system, worried about scope and inclined to reuse the existing UI. I pushed for a systemized approach, showing how the existing inconsistencies would compound as the product scaled and drag down both usability and development speed. We aligned on a lean system of essential components with room to grow, and I refreshed typography and color to give it a coherent foundation.

New design system examples, highlighting typography and colors.

New design system examples, highlighting typography and colors.

Designing the experience

Designing the experience

The hardest design question was what information mattered most, especially on the dashboard. That meant balancing hierarchy and clarity with product judgment about which insights, alerts, and actions actually drove value. I pushed the team past surface-level AI features and introduced a generative AI assistant as a central layer connecting workflows.

I also audited alerts and notifications with leadership and users, cut redundant signals, and consolidated everything into a smaller set of meaningful categories to reduce noise.

The hardest design question was what information mattered most, especially on the dashboard. That meant balancing hierarchy and clarity with product judgment about which insights, alerts, and actions actually drove value. I pushed the team past surface-level AI features and introduced a generative AI assistant as a central layer connecting workflows.

I also audited alerts and notifications with leadership and users, cut redundant signals, and consolidated everything into a smaller set of meaningful categories to reduce noise.

Kaboom's new homepage.

Kaboom's new homepage.

The redesigned product is organized around one core concept: Nudges, AI-generated alerts that surface prioritized insights for a user's day, each with a recommended next step they can act on directly. This creates a tight loop between insight and action, and the AI assistant lets users retrieve information, make updates, and run tasks without breaking their flow.

The redesigned product is organized around one core concept: Nudges, AI-generated alerts that surface prioritized insights for a user's day, each with a recommended next step they can act on directly. This creates a tight loop between insight and action, and the AI assistant lets users retrieve information, make updates, and run tasks without breaking their flow.

Core experience flow prototype.

Core experience flow prototype.

A few smaller calls: I used a distinct empty-state treatment with subtle animation to encourage adoption of the assistant. I tested requiring users to select an account before issuing commands, but it created friction (people forgot to switch context and made errors), so I dropped it. Prompt suggestions tested well but were deferred for scope. For viewing nudge details, I explored modals and drawers, initially leaning away from overlays, but testing showed overlays were the most intuitive and least disruptive option, so that is where we landed.

A few smaller calls: I used a distinct empty-state treatment with subtle animation to encourage adoption of the assistant. I tested requiring users to select an account before issuing commands, but it created friction (people forgot to switch context and made errors), so I dropped it. Prompt suggestions tested well but were deferred for scope. For viewing nudge details, I explored modals and drawers, initially leaning away from overlays, but testing showed overlays were the most intuitive and least disruptive option, so that is where we landed.

AI assistant iterations.

AI assistant iterations.

Drawer explorations.

Drawer explorations.

I also advocated for responsive design. The existing mobile experience was a basic engineering adaptation, and while mobile usage was low, users still expected reliable access across devices. I got it into scope and secured the resources to deliver it.

I also advocated for responsive design. The existing mobile experience was a basic engineering adaptation, and while mobile usage was low, users still expected reliable access across devices. I got it into scope and secured the resources to deliver it.

Mobile screen examples.

Mobile screen examples.

What's next

What's next

Expanding AI capabilities (prompt suggestions, saved prompts), deepening integrations with email and chat, and reconsidering previously removed features with a more disciplined approach to validation.

Expanding AI capabilities (prompt suggestions, saved prompts), deepening integrations with email and chat, and reconsidering previously removed features with a more disciplined approach to validation.

Website by David · davidryudesign@gmail.com · LinkedIn