Skip to main content
All work

Replacing The Globe and Mail’s keyword search with an AI assistant that subscribers can trust, from research to a live launch.

Timeline
Jan – Apr 2026
Role
Senior Product Designer
Platform
Web and app
Status
Live, subscriber beta

My role

I led design end to end: research, interaction design, prototype testing and handoff, working with data science, engineering, product and research. My first push was to get research done before the vendor deadline locked the scope.

The problem

Search that only matched words

The Globe and Mail’s search was a third-party keyword matcher. It couldn’t tell what a reader meant, and there was no way to refine a query. Readers had to phrase things the way the system wanted.

Data science had built a retrieval model (RAG) that could answer plain-language questions across articles, market data and author profiles. It needed an interface, and the vendor launch date was set before design joined.

Before: keyword search results. After: intent-based results grouped by what the reader asked.
Before and after: keyword matching vs. intent-based results.
How do you design one search box for a model that gives a different kind of answer depending on what you ask?

Research

What research overturned

I planned a three-phase research program. We ran the first two before any interaction design; the diary study is set for phase 2. The goal wasn’t to confirm what we believed. It was to find the beliefs worth challenging.

01

Survey

300 readers told us how they search for news. They wanted AI that makes search faster, not different. 69% wanted quick summaries, but 41% never use AI chatbots for news.

02

Concept testing

10 Globe and Mail subscribers used an interactive prototype that tested two search modes, the tab structure and grouped results.

03

Diary study Planned for phase 2

The launch timeline didn’t leave room for one. Phase 2 will track how readers use the assistant over several weeks, beyond a first session.

We assumed

We learned

01

Separate news and markets modes would help readers set context.

Picking a mode before typing slowed people down. One input that detects intent worked better.

02

A conversational interface would get people writing full questions.

People typed short keywords anyway. The look of a chat changed what they expected to see, not how they searched.

03

The hard part was making AI answers feel impressive.

The hard part was making them feel safe. Readers wanted reliable results and clear sources, not synthesis.

04

A ranked list is the clearest way to show results.

Grouping results under headings that match the question helped people judge coverage faster.

05

For market questions, a written AI answer would be most useful.

A live price, chart and key numbers felt more credible than generated text.

Fig. 1 · The first assumption research overturned.

Key insight

Readers weren’t asking for AI. They were asking for search that understood their question.

Research prototype

Exploration

Where a search starts

Before the prototype, I explored three empty states. Each one bet on a different way readers would get started.

Empty state with a personalized greeting and an example query

Personalized greeting

A time-of-day greeting and an example query for signed-in readers.

Empty state with prompt chips grouped by topic and a shuffle button

Prompts first

Prompt chips by topic, with shuffle for readers who don’t have a question yet.

Empty state with a news and markets mode toggle in the input

Two modes

A news / markets toggle inside the input.

Ruled out in research (see 01 above)

Interaction flows

I mapped every path the model could return for each of the three query types.

Scroll inside to see the whole flow
Flow for general news queries and their branches
Flow for company and ticker queries
Flow for author and person queries

General news: a topic or event query, with branches for clear, broad and no-result answers. Open full size

Key decisions

Five decisions that shaped the product

01 / 05

One input, no modes

News search and market search became a single input. The model works out the intent, so readers never have to choose before they type. A question about Shopify is about the stock and the news at the same time.

Tradeoff

One model is harder to tune at the edges than two specialist ones.

Desktop empty state with one search input
02 / 05

Two tabs, built for what comes next

Best matches shows AI-ranked results grouped by intent. All articles shows everything retrieved, with filters. The tabs leave room for AI summaries later without a redesign.

Tradeoff

Without summaries, the first release is less conversational than the vision.

Desktop Best matches tab with grouped results
03 / 05

An answer shaped like the question

Four result types, each matched to a kind of query: article cards for news, live price and chart for a company, author cards for people, and a carousel for comparisons. I worked with data science to define the output each one needs.

Tradeoff

Four result types meant a strict contract with the backend. Design helped define what the model returns.

Stock comparison carouselMultiple author cards
04 / 05

Trust on the screen, not in onboarding

A BETA badge, a link back to classic search, sources and a disclaimer on market data, and a feedback link. Each one sits where the doubt happens, not in a help page.

Tradeoff

More on every screen. Each element answers a specific doubt we saw in testing.

Fig. 2 · Trust signals where the doubt happens.
05 / 05

Design the vision, ship the MVP

Chat history, follow-up suggestions and AI summaries are fully designed and on hold. Summaries can get facts wrong, and in a newsroom that risk waits until the model has proven itself in real use.

Tradeoff

The first release feels less conversational, but it is reliable and ready to grow.

Fig. 3 · Phase two fills in without a redesign.

Fig. 2 · Trust signals where the doubt happens.
Fig. 3 · Phase two fills in without a redesign.
01 / 05Desktop empty state

Final design

Try it

The system

How a query becomes a result

The design work sat at the end of the pipeline: what the model returns, how it signals trust, and which component shows it.

Fig. 4 · Design shaped the last three stages.

Outcome

Live, on time, ready to grow

Shipped

Live for subscribers as a beta, on the vendor date, with research, design and engineering agreed on scope.

Trust

In beta testing, readers were more willing to try AI results when the trust signals were on screen.

Ready

Summaries, follow-ups and chat history can switch on without a redesign.

Reflection

The most useful trust feature was one line of text: Switch to classic search. People will try something new when they can see the way out. The hardest call was holding back summaries after we had designed and prototyped them. The model hadn’t yet earned that trust in a newsroom.

What I’d look at next

  • 01Where people retype a query, a sign the model missed.
  • 02How usage splits between Best matches and All articles.
  • 03When summaries can be turned on safely as the model improves.