Sue's already on. You decide how far she goes.
Every StayUpfront workspace comes with Sue — the AI teammate who triages your support the moment a ticket lands. She's on by default, works inside the limits you set, and you're always one click from full manual.
One thing she does runs today; more arrives for launch. Every bit of it sits behind controls you own — and they're on this page.
Every tool I evaluate right now wants to put AI in front of my customers. Auto-written replies. Generated status updates. An "AI agent" that resolves the ticket so I don't have to read it. I get the appeal. I also know what it costs.
When something breaks, my customer doesn't want a fast answer from a machine. They want to know it's not just them, and that a real person is on it. An AI reply in that moment doesn't reassure them — it tells them I've automated the one part of this that was supposed to be human.
So we pointed the AI the other way. Inward, at the person running support — sorting the queue, surfacing the urgent one, doing the quiet work so the human has more time for the parts that need a human. It's there from day one — you don't switch it on, you dial it back if you want to. We didn't make "AI" the headline, because it isn't the point. And we won't put it in your customers' inboxes unless you've turned that on yourself, on purpose.
— Rob Gough, Founder
Meet Sue, the AI inside StayUpfront.
Sue is short for Stay Upfront Expert — the AI layer inside the product. She comes switched on for every workspace, doing the kind of work that used to need a spare pair of hands you didn't have. You don't set her up; you decide how far she goes.
Today, that's triage. The moment a ticket lands, Sue does the sorting herself — she sets the priority, files it under the right category, tags it against the right components, and links it to a related incident if there's one open, so the queue is already organised before anyone opens it. Always pointed at making your team's time go further, never at replacing the human your customer actually wanted.
And she does it your way. You can give her your own instructions — what she should know about your product, and what "urgent" actually means for you — and she takes all of it into account when she triages. She works to your judgement, not a generic playbook. (More on the controls below.)
On by default. Yours to control.
Sue ships switched on, because a capable teammate shouldn't need assembling. The trust is in what surrounds her: clear limits, a full record of every call she makes, and an off switch that's always one click away.
- It works inside the limits you set
- Sue never overrides a human. If you've set something by hand, hers defers to yours — even when she's sure. She acts inside the boundaries you define, and stays there.
- Every decision is logged, with the reasoning
- Open any ticket and you can see exactly what Sue did and why — the priority she set, the category and components she chose, the incident she linked, and the reasoning behind each. Nothing she does is a black box, and nothing's lost when you override her.
- She works to your instructions
- Give Sue your own instructions — what she should know about your product, what counts as urgent, the calls you'd make differently — and she takes them into account every time she triages. The judgement is yours to shape; she just applies it consistently.
- Granular control — turn off anything you don't want
- Every one of Sue's capabilities is on by default, and each has its own switch. And the calls you'd rather make yourself, you just make — set a ticket's priority by hand and Sue won't touch it. The granularity is there to dial her back, capability by capability — never to make you assemble her piece by piece.
- Turn it all off, any time
- Prefer to run things by hand? One toggle stands Sue down completely — you're always one click from full manual, no support request, no waiting.
What Sue does today. Sue
Running today: Sue triages your support tickets. The moment a new ticket lands she reads it and sorts it herself — sets the priority so the urgent one doesn't wait behind the routine ones, files it under the right category, tags it against the right components, and writes a short summary so you can see what it's about before you open it — so the queue is organised before anyone opens it.
Sorted in seconds — or it skips straight to you. Triage happens in seconds, so the queue's already organised by the time you look. And there's a hard fail-safe: if Sue can't finish in time, the ticket goes straight to the inbox, untouched — nothing ever sits waiting on AI. What you feel is one clean hand-off instead of a burst of half-finished notifications.
She works from context you give her. You can tell her what your product is, who your important customers are, what "urgent" means for your team, and the vocabulary she wouldn't otherwise know — and you can change any of it whenever you like. It takes effect on her next pass, no waiting on us.
She also connects the dots — when a ticket looks like it's caused by an incident you've already got open, she links them, so whoever picks it up sees it straight away. And anything you've set by hand, she leaves alone.
What's coming after that — honestly
Triage today — that's where we are, not where we'll stay. The platform underneath Sue is built to grow: every new AI feature plugs into the same gate and inherits the same rails. Triage, and the first acknowledgement Sue sends your customers, are switched on and working today. More will follow — built on the same foundation.
So we're not going to promise you a roadmap of AI features with dates on it. What we'll promise is the shape of how we add them: pointed at making your team's work lighter first, proven on real tickets before they ship, and never quietly turned on without you noticing.
Here's the one she already does for your customers: when a ticket comes in, Sue sends the first acknowledgement straight back, so no one's left wondering whether it landed. She's working up to drafting the first version of an incident update, too — for you to review, edit, and publish. When Sue is the one replying to a customer, she's labelled — her name, the bot icon:
Sue sends that first acknowledgement so your customer isn't left hanging — then a person picks up the conversation from there. That's the shape of all of it: Sue does the groundwork, your team does the part that needs a human.
Sue's already part of the team. Get in early.
Private beta is a few weeks out, and I'm letting people in a small group at a time. Sue's on from day one — the earliest customers shape what she does next, what should stay off until you ask for it, and where the line is on anything that touches your customers. If you've got an opinion about how AI should and shouldn't show up in customer ops, I want it.
Drop your emailDirect email from Rob when your slot's ready. No drip sequence.