Core runs the night. Insights, the $199 a month add-on, helps you understand the orders, waits, drink feedback and work behind it. Explore the decisions it supports, the value for each role, and how Lync helps you ask about your venue.
Core carries the order through the venue. Insights helps the manager understand the work behind it. Lync provides a conversational way to explore the context available to that role.
Manager: map floors, bars and pickup points; review menu availability and pricing; prepare staff access. A customer's bar choice should match the station receiving the order.
The value: a service plan the customer and team can follow, rather than a menu disconnected from the floor.
Plan the setupCustomer: choose the venue and bar, review drinks and modifiers, pay in the app, then follow the order toward pickup.
The value: the drink specification and destination travel with the ticket. The customer has a place to check progress without repeatedly interrupting the bartender.
Explore the customer appBartender: read quantities and modifiers, accept work, and mark it ready. The ticket holds the instructions; the bartender stays focused on the pour.
The value: a shared record of what was requested and when it moved. Shift evidence helps the manager distinguish a difficult station or drink mix from an individual performance issue.
Compare staff rolesGatekeeper: see ready orders, verify the pickup code, check the items, and confirm handoff. Their working screen stays focused on collection.
The value: ready and collected are separate events. The manager can investigate a pickup bottleneck without treating it as slow preparation.
Compare staff rolesCustomer: review the completed order, leave a rating and choose whether to tip. These are separate choices, not a condition of collecting a paid drink.
The value: feedback can be examined with the drink and order context. A few responses are a signal to investigate, not proof about every customer or staff member.
See how feedback becomes a next stepManager: review the night's orders, waits, drink mix and feedback. Use the relevant workspace to change a menu or plan coverage, then compare the next appropriate period.
The value: decisions have an observation behind them and a result to check afterward. Recommendations are not automatic changes or guaranteed revenue.
Open manager InsightsThese are example decision paths, not reports of a live venue's results. Start with the observed signal, check its scope, and review the outcome after acting.
Inspect: ready-to-pickup wait, outstanding ready orders, bar and time window. Keep preparation time separate.
Next action: review the handoff point and staffing coverage. Check signage and collection flow before assuming another bartender is needed.
Check afterward: compare pickup wait and completed handoffs over a comparable window. A shorter queue alone is not proof if demand also dropped.
Inspect: units, rating count, comments and station context. Check whether there are enough responses to support the pattern.
Next action: review the recipe, modifiers and availability in the menu; coach the preparation issue you can actually verify.
Check afterward: compare the same drink's later feedback with its response count. Do not turn a small sample into a staff verdict.
Inspect: compare the same elapsed hours and distinguish order count from revenue. A partial shift should not be judged against a completed night.
Next action: inspect demand by bar and time. Review availability and coverage before changing pricing or drafting a promotion.
Check afterward: use a comparable trading window. Forecasts and possible explanations remain estimates, not money already earned.
Inspect: ticket volume, drink complexity, modifiers, timestamps and attribution coverage. Different mixes of work need context.
Next action: review the station setup and the shift with the team. Use the underlying tickets to discuss a specific improvement.
Check afterward: compare comparable workloads; do not present missing or unattributed work as zero.
Lync can use role-specific context when the relevant records are available. These are useful questions to try, not prerecorded answers or a promise that every integration is active for every account.
How is tonight going, and what should I check first?
Available context can include venue orders, the trading window, feedback, pace and forecasts. Ask which period and evidence support the answer, then open the relevant operational screen.
What does my current workload look like?
Use the queue and available shift context to understand the work. Ticket details remain the reference for the actual order; a chat answer does not accept or finish a drink.
What is waiting on the ready wall?
Ready-order and pickup context can help explain the handoff workload. Use the ready wall and verification flow to collect an order; a conversational response is not pickup authorization.
What is popular here, and where is my order?
Menu, venue and order context can support the conversation. When popularity or timing data is unavailable, the answer should say so rather than inventing a bestseller or a pickup promise.
Lync suggestions need human judgment. An answer is not a completed operational action. Live permissions, data availability and account configuration determine what can be accessed; do not treat a conversation as proof that a price, payment, staffing plan or campaign changed.
Insights is not a PDF that lands on Monday. It is a set of screens inside the manager portal, each answering a different question at a different hour of the night. Open one, tap a number, land on the tickets behind it.
The one consolidated insights screen: generated insight cards, the insight deck, an end-of-shift recap, a Lync usage tile, and a feedback summary strip — total feedback, average overall rating, average drink rating, and how many came in at two stars or under. The old Copilot route is a permanent redirect here, so bookmarks and muscle memory keep working.
The live strip: sales, open orders, average wait, on shift, and needs attention, sitting over an order-by-order pipeline board. It computes a next best action rather than leaving you to infer one, and it states what an elevated or a high wait actually is in seconds instead of just colouring a tile.
Total orders, revenue, average order value and peak hour for the night just finished, with orders-by-night and revenue-by-night bars beside them so one Friday reads against the last several instead of standing alone.
The daily digest, previewed before it sends: email subject and body and the plain-text mirror, per audience and per period. Send one out of band, or read the history of what has already gone and the schedule that is stored.
A venue-scoped timeline of every Lync thread, tagged by who asked: manager, customer, bartender, gatekeeper, owner. Filter to the last hour, six hours, day or week. It exists for support, training and review, and it is the honest answer to “is anyone using this?”
Orders per shift, average pickup time, thumbs-up rate and total orders for every bartender, ranked by throughput and by pickup time, with station benchmarks beside the people so a slow rail is not blamed on whoever is standing at it.
Pickups, average ready-to-pickup wait, volume rank and trend for every gatekeeper, over seven, thirty or ninety days. Pickups that were never stamped with a person are surfaced as a coverage note rather than quietly dropped, so a total still reads as a total.
So they do not get one. Pick a role and read exactly what that person sees — and, for the gatekeeper, exactly what they deliberately do not.
Then the charts: drinks poured by day, tips by day, pour time against target, top drinks you made, which drinks eat your time, your rating over time, your hours.
No dashboard, no chart, no shift score. That is a decision, not a gap: the old gatekeeper dashboard and the old gatekeeper insights screen were both retired and now redirect to the ready wall.
Live tonight, then the recap in the morning: total orders, revenue, average order value, peak hour, and last night read against the nights before it.
These are not chat toys. Each one is a plain, testable function with a declared input, a declared output and a declared cadence, registered in one catalog. Two of the four are heuristics today, with the model-backed version as the explicit next step — we would rather say that than call a scoring function artificial intelligence.
Walks every venue's pulse snapshot, scores each one out of 100 and buckets it as excellent, healthy, warning or critical. Anything at or below the threshold becomes an alert with its reasons spelled out, so the alert arrives already explained.
Turns a customer's own signals into exactly three suggestions: one from what they already like, one deliberate step sideways, and one stretch from a different category. Built and tested, and not yet shown to customers: it switches on with customer drink suggestions.
Reads pending venue applications and sorts them into auto-approve, needs-review and reject, so a human reads the pile that actually needs a human.
Takes every open support case and stamps it with an urgency, a category and a suggested owner, so the queue arrives sorted instead of arriving.
This adds up prices, nothing else. There is no uplift assumption hidden in it, because we are not going to promise you money before it is money. Slide your stations and your own average drink price to see the bill and what the add-on weighs against your menu.
Core includes 1 station; each additional station is $299 a month. Insights is a $199 a month add-on you switch on and off from the manager portal. Setup is $500 one-time, whatever the number of stations.
The last line is arithmetic on the price you set above, not a forecast: $199 divided by your own menu price, rounded up. It is what the add-on weighs, not what it earns. One-time setup ($500) is not in the monthly figure. Card processing is separate and is not billed by us: your customers’ card payments land in your own Stripe account and Stripe charges you its standard card fees on them.
A site that promises a capability the product lacks is worse than a shorter site. So here is the boundary, in the same place as the sales copy.
The manager insights destination and its feedback summary. The action center with its live tiles, order pipeline and next best action. The nightly recap. The digest preview, send and history. The Lync activity log. Per-bartender and per-gatekeeper performance in the manager portal. The bartender's own drill-down shift screen. All four agents, implemented, registered and runnable.
An always-on scheduler driving the agents on their declared cadences, and the paging fan-out that sends a pulse alert to a manager’s phone. Model-backed versions of the drink concierge and the support classifier, both of which are heuristics today. Enforcement of the no-dairy and gluten-free flags the concierge already accepts.
If we can point at the screen that renders it, it is here. If we cannot, we cut it rather than soften it. Where a figure is a projection rather than a measurement, the product labels it as modelled on the screen itself, and so do we.
We set up your menu, stations and staff screens. You run a normal night. In the morning the recap, the digest and the per-person numbers are waiting, built from your own orders.
