The whole team, one direction

When I was in GTM roles, I'd do a round of the team before any planning cycle.

Marketing. Sales. Customer support. I'd ask each of them the same question: what are your biggest customer asks right now? What keeps coming up? What are people asking for that we don't have?

And then I'd take all of it away and try to make sense of it.

The volume was significant. I could lose days just keeping up with the requests — synthesising across teams, reconciling different framings of the same problem, trying to find the signal underneath the noise. When I moved into product management, I did the same thing in reverse — pulling inputs from GTM teams before building anything, trying to construct a picture that represented the whole customer, not just the loudest channel.

The biggest risk was always the same: everyone turning up to planning with their own list.

Marketing with their list. Sales with theirs. Support with theirs. Product with another. Four teams, four versions of the truth, one meeting where whoever argued most confidently shaped what got built next.

The goal was always a combined list. One view that the whole team could look at together and say: this is where we are, this is what matters, this is the battle we're fighting. Small teams can move fast — but only if they're all moving in the same direction.

March was about building that.

---

Projects: one view and many

Projects launched in beta on March 17th. It's the feature that most changes what Circuit is.

Before Projects, Circuit was one workspace, one priority list, all feedback in a single queue. That works until it doesn't. Mobile app and web app. Growth team and platform team. Enterprise customers filing feature requests alongside free users reporting bugs. Mixed together, it's noise pretending to be signal.

Projects gives each team or product line its own feedback environment — its own surfaces, its own priority list, its own spec history. Slack channels map directly to projects. CSV and Sheets imports have a project picker so feedback lands in the right container from day one. A mobile app request doesn't pollute the web app's priority list. A billing request never competes with a dashboard feature.

But the part I care most about is the all-projects view.

The team still needs the 10,000-foot perspective. Where are we overall? What's rising across every stream? And when the same priority appears in multiple projects — when something is flagged as a challenge by the mobile team and the web team and the enterprise customers all at once — Circuit surfaces that. A consistent signal across projects isn't a coincidence. It's a directive.

That's what a combined list looks like. Not one person synthesising across four teams. The system doing it automatically, flagging the patterns that matter, giving the whole team the same starting point.

Ask Circuit

Ask Circuit shipped March 20th.

Type a question. Get an answer from your actual product data.

"What are customers saying about onboarding?" "Which priorities have the most enterprise backing?" "Are there patterns in what churned users asked for?" Circuit reads across your feedback, priorities, specs and shipping history and answers in plain language — with customer quotes attached.

This is different from search. Search finds a specific item you already know to look for. Ask Circuit synthesises across everything and surfaces what you didn't know to ask.

Someone walks into planning with that answer. The conversation is different. Not "I think customers want X" — "here's what 47 enterprise customers said about X in the last 30 days, and here's how it ranks against everything else." The debate about priorities becomes a debate about what to do about them, which is a much more useful conversation.

Smarter signal

The intelligence underneath Circuit got a significant upgrade in March.

Intent-anchored embeddings now produce tighter, more meaningful clusters. Circuit learns your product's specific terminology — "slow" means something different for a video platform than a B2B dashboard. Clustering auto-tunes per account as feedback density grows. By end of March, the model was upgraded to full resolution, capturing subtle differences in meaning that similar phrasing would previously merge.

Priority titles shifted to consequence-driven problem statements — "Slow search frustrates power users" rather than just "Search." Intents simplified to four: Bug (fix it), Feature (build it), Improvement (enhance it), Praise (talk about it). The signal gets cleaner the longer the system runs.

Sixteen languages. Feedback in French, Portuguese, Japanese, Arabic — all scored alongside English submissions. The original customer quote is preserved in the spec, in the language it arrived in. The early Circuit signups came from Norway, Thailand, Singapore, Venezuela. A product intelligence layer only works if it works for every market, not just the ones that write in English.

Multi-channel Slack

Slack got a meaningful upgrade in March. Connect multiple channels and map each one to the right project — so the sales Slack channel feeds the sales project, the support channel feeds the support project, automatically. Messages land with deduplication even for large channel histories.

Before this, Circuit could read one Slack channel. After this, it can read the whole organisation — each stream flowing into the right priority list without anyone manually routing it.

Customers

Every customer who has ever sent feedback now has a record.

The Customers page launched in beta on March 20th. Filter by revenue band. Search by name. See every piece of feedback a customer has ever submitted — with their submissions feeding directly into how priorities are scored. Not "someone asked for this" but "the person who asked is enterprise tier, has submitted six other requests and two of them have already shipped."

That changes how you read a priority. The weight behind a request becomes visible. The customer becomes a person, not a data point.

Park priorities

Not every priority is the right priority right now.

Parking moves a priority off the active list without losing it. Scores keep updating. Attached feedback stays intact. It comes back when you're ready. This is how teams handle the portion of the backlog that's real but not now — held without being lost to the queue forever.

The free plan and the navigation

March ended with two things that are really about the same thing: adoption.

The free plan — no credit card, no trial expiry, choose it upfront — removes the friction at the entry point. Product intelligence only works if the whole team feeds it signal. Paywalling the entry means fewer submissions, weaker signal, less accurate priorities. The free plan fixes that.

The navigation redesign — persistent left sidebar, Ask Circuit reachable from anywhere, settings as proper routed pages — makes Circuit feel like one product rather than a collection of screens. The full feedback-to-spec pipeline only works if the whole team actually uses it. Navigation that gets out of the way is part of adoption.

Sixteen languages. A free plan. A navigation that works. These aren't separate features. They're the same argument: product intelligence has to be accessible to everyone on the team, in every market, on day one.

---

February built the missing layer.

March made it work for teams — organised around projects, unified at the top, smart enough to surface the patterns no single person could see across all that volume.

The combined list exists now. The whole team can look at it together.

That's what I spent years trying to build manually.

---

Catherine Williams-Treloar is the founder of Circuit — the AI product system that turns customer feedback into scored priorities and build-ready specs for Cursor and Claude Code. She has 20+ years leading product, insights, strategy and GTM at scale-ups and enterprises. Circuit was founded in Sydney in November 2025 and launched in February 2026.