Case Study· 2026

AI Sales Agent

Repositioning marine product discovery around conversation, personalization, and agentic action.

Client
White-label — confidential marine enterprise client (under NDA)
Category
Product Design · AI · Marine
Year
2026
Five AI sales assistant app screens fanned on a deep navy gradient, featuring the Wavy AI agent character — a center iPhone showing Wavy on the assistant home with Ask AI, Compare, Specs, and Send to Buyer actions, flanked by frameless Discover, Wavy chat, Compare, and Model Detail screens.
01

The problem

A high-consideration marine platform relied on keyword search to serve buyers spending tens or hundreds of thousands of dollars. The search box treated first-time buyers and repeat owners identically, product information was fragmented across tabs, and sales reps had no conversational assistant to guide buyers on the showroom floor, during sea trials, on the boat, or at boat shows. The business risk was not a missing feature — it was a missing interface model for a complex, high-stakes purchase that could convert sales more efficiently in the moments that matter.

02

My role

Product Design, Systems & AI Strategy Lead

  • Led a 12-week discovery-to-pilot engagement, defining the product strategy, AI architecture, and enterprise design system for a multi-brand white-label rollout.
  • Established the two-mode model: an anonymous concierge AI chatbot for zero-friction consumer answers and an authenticated AI Sales Agent that serves as a field assistant for sales reps on the showroom floor, sea trials, the boat, and at boat shows.
  • Designed a persistent chat surface and intent router that connects natural language to catalog, knowledge, CRM, and appointment workflows so reps can answer questions and move deals forward in real time.
  • Built the enterprise design system — tokens, type, motion, voice, and components — to support rapid brand refreshes and a future consumer-facing self-serve sales experience.
  • Partnered with engineering and sales leadership on the personal-algorithm concept, converting behavioral signals into actionable recommendations that help close complex, high-consideration purchases faster.
03

Process

  1. Step 01

    Strategic Framing

    Repositioned marine product discovery from passive search results to active, dialogue-driven guidance. The strategic bet was that conversation, not keywords, is the right interface for high-consideration marine commerce — especially when a sales rep is standing next to a buyer on the showroom floor or at a boat show.

  2. Step 02

    Field Sales Assistant Model

    Designed the agent as a sales rep assistant for high-touch moments: answering specs on the showroom floor, pulling comparisons during sea trials, surfacing inventory and pricing on the boat, and following up after boat shows. The goal was to reduce answer time and help reps convert complex sales more efficiently.

  3. Step 03

    Two-Mode Architecture

    Designed the anonymous chatbot to answer top-of-funnel questions without friction, and the authenticated agent to act as a trusted sales partner for repeat buyers — comparing, scheduling, and drafting offers. The same architecture can later extend to a consumer-facing self-serve sales experience.

  4. Step 04

    Enterprise System Layer

    Shipped a token-based design system and a thin AI layer over existing catalog, CRM, inventory, and pricing systems so the business could pilot without rebuilding the stack.

04

Measurable results

+38%
Projected session-to-lead lift (authenticated cohort)
2.4×
Forecasted return-visit frequency after 3+ sessions
−61%
Estimated time-to-answer on top 20 support intents
−27%
Modeled deflection of human-handled tickets
This proposal reinforced that AI strategy is not about the model; it is about the interface and the action layer. A personal algorithm is only worth building if it turns into a trusted next step for the buyer — and a practical selling tool for sales reps and dealers in the field.
Case study · 2026 · AI Sales Agent

Conversational AI as a sales strategy

AI Sales Agent was a 2026 product proposal for a white-label conversational AI sales assistant built for marine sales reps in the field — on the showroom floor, at sea trials, on the boat, and at boat shows. The agent helps reps answer buyer questions quickly and convert complex, high-consideration sales more efficiently. Anonymous chat answers first-time buyers; the authenticated agent remembers repeat owners and helps close multi-thousand-dollar decisions. The same architecture can evolve into a consumer-facing self-serve sales experience. Metrics are forecasted projections from the proposal, not live results.

Strategic framing

From search box to guided selling

Built for reps in the field

The assistant is designed for high-touch selling moments — showroom floor, sea trials, on the boat, and boat shows — giving sales reps instant answers so they can keep deals moving instead of digging through tabs.

Design for two modes

Anonymous browsers get immediate, unbiased answers. Signed-in members and sales reps get a personal algorithm that respects privacy and acts like a trusted partner. The architecture can later extend to a consumer-facing self-serve sales experience.

Turn conversation into action

Every reply moves the buyer closer to a decision: compare boats, schedule a sea trial, qualify financing, or escalate cleanly to a human — all from the same conversational surface.

FAQ

Common questions

Was this system deployed?

No. This documents a 2026 product proposal and design concept, not a live deployment.

Are the metrics real results?

No. The numbers are forecasted projections from analogous products, discovery data, and conservative assumptions.