I design the interface, then automate what happens after someone clicks it.

Small and mid-sized business clients were losing revenue in two specific, unglamorous places: missed calls that never got a callback, and overdue accounts that nobody had time to chase down consistently.

Both problems were repetitive enough to automate, but sensitive enough that a bad bot experience would cost more in trust than it saved in labor.

The brief wasn't "add a chatbot" — it was to design something that could actually stand in for a receptionist or a collections rep without customers feeling handled by a machine.

• Call-routing and receptionist flows needed to handle real ambiguity — customers don't phrase requests the way a decision tree expects.
• Billing collection is a trust-sensitive conversation; getting the tone wrong risks the client relationship, not just the call.
• Every client had a different intake process, so the system had to be configurable, not one-size-fits-all.

I designed the conversation flows the way I'd design any interface — mapping the real paths a caller takes, not just the happy path, then working with engineering to wire that logic into the LLM and Twilio integration.

For the collections agent specifically, I designed the decisioning around intent and risk signals in the conversation itself, so the agent could propose a custom payment plan in real time rather than following a rigid script regardless of what the customer actually said.

Alongside the conversational product, I designed the visual and brand layer for the platform itself — the interface clients and their customers actually see and configure.

The collection agent's real-time risk-and-intent decisioning became the reference pattern for how later conversation flows on the platform were designed.

Seaport operates terminals and container yards across Toronto and Montreal, running hundreds of trucks and cranes and handling over 10,000 TEUs — the kind of operation whose scale and reliability needed to be immediately legible to a shipper evaluating them for the first time.

Logistics buyers make decisions on trust and capacity signals — safety record, fleet size, terminal capabilities — and needed to find them fast, alongside practical tools like getting a quote or tracking a shipment.

• The site had to serve very different visitors in one structure: prospective shippers evaluating capacity, existing customers tracking freight, and candidates checking careers.
• As a client-managed WordPress build, the design had to stay maintainable by a non-technical team after launch, not just look good at handoff.
• On-page SEO — headings, meta data, schema, clean URLs — had to be built in from the start, not bolted on after.

I structured the information architecture around the decision a shipper actually makes — company credibility first (safety, scale, locations), then services, then the practical tools (quote request, tracker) — rather than an org-chart version of the company's departments.

Trust signals that read as abstract claims elsewhere — "reliable," "safe" — were designed as concrete numbers instead: fleet counts, terminal capacity, kilometres travelled, elite safety status with their insurer. Specific numbers hold up better under a buyer's scrutiny than adjectives do.

Built on WordPress and Elementor so Seaport's own team could update service pages and locations without needing a developer for every change, with performance work — image compression, caching, lazy loading — layered in so the added visual weight of a trust-heavy site didn't cost load time.

Live at seaportgrp.com today, with the client team maintaining service and location pages independently since launch.

Travelpath runs two genuinely different businesses under one name: corporate travel programs for Canadian SMEs and enterprises, and personalized vacation planning for individuals, couples, and families.

A corporate travel manager evaluating policy-compliant booking tools and a couple planning a honeymoon are looking for almost nothing in common — different language, different proof points, different next step.

• Design the site to serve both without either audience feeling like an afterthought bolted onto the other's homepage.
• Keep a single, recognizable brand across both — this isn't two companies, it's one trusted name doing two things.
• Route each visitor to the right advisor conversation quickly, since both sides of the business ultimately convert through a human, not a self-serve checkout.

Rather than force one homepage to speak to both audiences, I designed the architecture to fork early — a clear split into corporate travel and vacation travel right from the entry point, each with its own language, proof points, and call to action, while keeping shared brand elements (identity, tone, advisor-led positioning) consistent across both.

Corporate pages lead with operational language — policy alignment, negotiated savings, real-time visibility — because that's what a travel manager is evaluating. Vacation pages lead with the advisor relationship and destination inspiration, because that's what moves a leisure traveler.

Both tracks converge on the same underlying goal: get the visitor talking to an advisor, so every path was designed to make that handoff feel like the natural next step rather than a form to fill out.

Live at travelpath.com, with the corporate and vacation tracks running as clearly separated journeys under a single, consistent brand.