Desk Review: Lightspeed Restaurant, the Quiet Half of the Duopoly

There is a version of this Desk Review that I drafted two months ago and shelved. It was built around a Lightspeed AI labour-scheduling product that, on closer inspection, does not exist in the form I had described it. I pulled the piece, apologised to a couple of operators whose quotes I had asked them to verify against features they did not, in fact, have, and started over. The new version is below. It is not the punchy operator-panel write-up I had planned. It is a desk review of what Lightspeed Restaurant actually ships in 2026 — read against the company’s own press, its quarterly filings, and the trade-press coverage of its product launches.

The catalyst for redoing the piece is that in the last four months Lightspeed has made two strategic moves that change the shape of any honest desk review. In January 2026, the company launched its first cross-product AI layer, branded Lightspeed AI (Lightspeed, 9 Jan 2026). In April 2026, it sold the U.S. Upserve hospitality product line to Skyview Equity for up to $81M and re-anchored its restaurant strategy on Europe (Lightspeed, 29 Apr 2026). Read together, these two announcements tell most of what an operator needs to know about Lightspeed Restaurant in 2026. The rest of this Desk Review is the texture.

Methodology

This is a desk review, not the operator-panel “Vibe Check” format we usually run in this slot. The difference matters and I want to be explicit about it. Vibe Checks are sourced from real, named operators running specific features for at least 60 days and reporting back with metrics they can stand behind. Desk Reviews are sourced from public materials — company press, earnings filings, product pages, trade press — and represent my read on what a vendor actually ships, not how it performs in any particular dining room.

Every claim in this piece links to its source. Where Lightspeed has made a specific commercial claim (time savings, customer counts, financial guidance) I have cited the company’s own materials and noted them as vendor claims, not independent measurements. Where I render a judgment, it is mine, and it is not derived from running the product. If you want operator-panel reporting on Lightspeed Restaurant, I will run that piece once I have enough panelists on the platform whose features and data I can verify against. I am not there yet, and rather than ship a piece that overstates my visibility into the product, I am shipping this.

This methodology change is part of a broader correction inside TableTransfers about how aggressively we ought to use the Vibe Check format. The shorter answer is: only when the panel can actually do the work the format requires. Lightspeed Restaurant did not pass that bar this cycle. It may next cycle.

What Lightspeed Restaurant actually is, in 2026

Lightspeed Commerce (NYSE/TSX: LSPD) is a Canadian-headquartered commerce platform with two operating lines: Retail and Restaurant (which the company markets as “hospitality”). The most recent quarterly print, Q3 FY2026, reported total revenue of $312.3M, up 11% year-over-year, with gross margin of 43% and adjusted EBITDA of $20.2M (Lightspeed, 5 Feb 2026). Within that, the company explicitly identifies its two “growth engines” — retail in North America and hospitality in Europe — which together represent roughly two-thirds of revenue and grew 21% year-over-year in the quarter, with about 94,000 customer locations in those engines, up 9% year-over-year (Lightspeed Q3 FY2026 release).

That phrasing — “hospitality in Europe” — is doing real work. As of April 2026, U.S. hospitality is no longer a Lightspeed growth engine in any meaningful sense. The company sold its Upserve U.S. hospitality product line, covering approximately 3,200 customer locations and 70 employees, to Skyview Equity for up to $81M ($44M fixed plus up to $37M earnout) (Lightspeed, 29 Apr 2026). The deal removed roughly $14M of revenue and $26M of gross profit from Lightspeed’s fiscal 2026 base. CEO Dax Dasilva framed it as “another step towards focusing Lightspeed on the markets where we have the strongest right to win” (Lightspeed, 29 Apr 2026).

The relevant detail for U.S. operators reading this Desk Review: Upserve, the brand that historically served independent and small-chain U.S. restaurants, is now operated by a different company. The flagship Lightspeed Restaurant platform — the cloud product built from the consolidation of Kounta, Gastrofix, iKentoo and earlier Upserve features — remains a Lightspeed product, and Lightspeed has continued to sell it in North America, but the company’s sales investment and product centre of gravity is clearly Europe. That is the strategic ground on which any U.S. or Canadian operator should evaluate them.

The product surface, as of May 2026

The Lightspeed Restaurant product surface consists of, by the company’s own product page, a cloud POS, a Kitchen Display System, customer-facing display, tableside ordering, an Order Anywhere QR/online ordering module, integrated payments (Lightspeed Payments), accounting integrations, inventory at the ingredient level, a workforce management and payroll module, Advanced Insights reporting, a competitive analytics module called Benchmarks & Trends, the Lightspeed Pulse mobile app for managers, a service-pacing product called Tempo, an operational checklist product called Tasks, and the conversational Lightspeed AI layer (Lightspeed product page; Lightspeed AI product page).

Pricing is published, which is uncommon in this category. Starter is $69/mo, Essentials $189/mo, Premium $399/mo, and Enterprise is custom (POSUSA review; Lightspeed pricing). All tiers include the core POS, online ordering, contactless payments, and CRM/loyalty. Premium adds raw API access and multi-revenue-centre support (useful for hotel F&B and venues with multiple cost centres). Lightspeed does not publish an annual-billing discount on the restaurant SKUs, which is mildly unusual — most North American POS vendors have moved to an annual price.

That is the structural shape of the product. The interesting question for this Desk Review is what AI Lightspeed has actually shipped on top of it, and how that compares to the obvious peers.

What “Lightspeed AI” actually is

Lightspeed AI, the company’s January 9 2026 launch, is the umbrella brand for a conversational layer that sits across both the Retail and Restaurant products (Lightspeed, 9 Jan 2026). On the restaurant side it is, per Lightspeed’s own product page, an assistant that lets an operator ask plain-language questions about sales, menu performance, staffing, and order activity, and receive narrative answers in place of a dashboard (Lightspeed AI page).

The specific capabilities Lightspeed advertises on the restaurant side are: sales summaries with day-part and shift comparisons; menu-performance analysis that flags underperformers; a “staffing optimisation” view that combines sales, weather and event data into a recommendation about how to staff; an order-activity investigator for explaining unexpected demand swings; and an in-app technical assistant that walks operators through tasks inside the POS itself (Lightspeed AI page). Lightspeed’s own framing of where this goes next is, in the company’s words, that the assistant will “evolve from an assistant into a more autonomous agent” (Lightspeed, 9 Jan 2026).

Two things to flag honestly. First, staffing optimisation in Lightspeed AI is a recommendation surface, not a labour scheduler in the sense Toast and 7shifts mean. As of the January 2026 launch, Lightspeed has not announced a feature that generates a draft weekly staff schedule, learns from operator overrides, and pushes shifts to a staff app. Their own page describes the staffing capability as “smarter scheduling decisions” derived from sales, weather and events data (Lightspeed AI page) — coaching, not draft generation. Anyone (including me, two months ago) describing Lightspeed as having shipped an end-to-end AI scheduler in early 2026 is wrong. Second, Lightspeed AI launched without disclosed incremental pricing, which is the same playbook Toast used with the early versions of Toast IQ (Toast IQ release). The economic model will become legible later in the year, probably alongside the autonomous-agent expansion the company has flagged.

The other pieces of the Lightspeed AI stack worth naming individually:

Benchmarks & Trends is the AI-powered competitive analytics module Lightspeed launched in the U.S. in September 2024 and extended to the U.K. in May 2025. It compares a restaurant against anonymised peer data in its local market — pricing, sales mix, peak hours — and surfaces machine-learning-driven forecasts on up to four weeks of historical data (Lightspeed, May 2025; Lightspeed press). Inside Benchmarks lives a feature called Magic Menu Quadrant, which classifies each menu item against retention (share of customers who return after ordering it) and popularity (raw order count) into four buckets: Hidden Gems, Greatest Hits, Underperformers, One Hit Wonders (Lightspeed K-Series support). It is the most interesting operationally of Lightspeed’s AI features because it gives you a defensible reason to cut an item, which is usually a political conversation inside a kitchen.

Smart Items is the AI menu-builder. Lightspeed claims it can save “up to 30 minutes per 100-item menu, or even hours and days for those building content from the ground up” (Lightspeed, May 2025). Translation: it autogenerates product descriptions and item metadata when an operator is setting up or refreshing a menu in the POS. This is squarely in the category of AI features that quietly remove keystrokes — useful, not strategic.

Lightspeed Tempo is the December 2025 launch most likely to define how Lightspeed competes in 2026. Tempo is a service-pacing layer that uses the Kitchen Display System’s real-time data to track when items are ordered, prepped, ready and served, then prompts servers and managers when a table is drifting from optimal pacing (Restaurant Technology News, 12 Dec 2025). Lightspeed cites an internal Hospitality Report claiming nearly one-third of diners abandon restaurants before ordering due to wait times, and another third never return after a delay — the framing for Tempo is that turning service pacing into a measurable discipline recovers covers that intuition-led service is leaving on the table. The product is integrated, per Lightspeed, with the Kitchen Display System and Pulse analytics.

Lightspeed Tasks is the workflow-checklist product launched in the same Q3 FY2026 product cycle as Tempo — opening and closing checklists, cleaning protocols, completed in the Pulse mobile app with real-time completion logging visible to managers (Lightspeed Q3 FY2026 product release). This is not AI, but it is part of how Lightspeed is closing operational gaps in the platform.

That, by my read of the public surface, is the complete set of Lightspeed Restaurant’s AI and adjacent intelligence features in May 2026. There is no labour scheduler in the Toast/7shifts sense. There is no AI voice agent — Lightspeed has not announced one. There is no AI drive-thru, because Lightspeed does not meaningfully compete in drive-thru QSR. There is, on the other hand, a credible conversational analytics layer, a defensible competitive-benchmarking product with a usable menu-engineering module inside it, and a service-pacing tool that — if it works as described — addresses a real and underserved problem. The picture is narrower than Toast’s, but tighter and more honest about what it tries to do.

Where this sits against Toast and Square

The comparison everyone wants is Lightspeed vs Toast vs Square. The version that respects the readers’ time goes like this.

Toast (NYSE: TOST) has, by most independent estimates, the largest installed base in U.S. restaurant POS — somewhere in the 23-28% range depending on the dataset and how restaurant-specific the cut is (6sense, “Toast vs Lightspeed”). On the AI surface, Toast IQ — the company’s smart-assistant brand — has, since its expansion announced earlier this year, the most operationally invasive AI surface of the three: an assistant that can take in-app actions (edit menu items and modifiers, adjust shift auto-clockouts) and that understands shorthand like “86 all items with avocado” or “adjust a server’s clockout from Friday’s dinner shift” (QSR Web, Toast IQ expansion; Toast IQ release). Toast also launched Toast Drive-Thru in April 2026, an AI-enabled QSR platform with voice ordering integrations (Toast/Businesswire, 14 Apr 2026), and Toast IQ Grow in May 2026, the marketing-agent SKU at $499/mo. Toast is currently competing along more axes than Lightspeed.

Square’s restaurant offering (Square for Restaurants) is the volume-first comparison: cheapest entry point, broadest installed base across the SMB Square umbrella, but the thinnest AI surface relative to the two. Where Square wins is on payments-led pricing and on the fact that an operator can ramp up on it without ever talking to a salesperson. That’s not what Lightspeed Restaurant competes on.

The honest read on the three-way:

• Toast is the deepest, most aggressive AI surface, and the only one of the three with a meaningful U.S. enterprise drive-thru offering. It is also the most likely to keep adding modules through 2026 (Toast IQ Grow). We covered the Toast AI Suite as a Vibe Check earlier in this series — see Desk Reviews 7, 8, and 9 for the previous instalments in this format.
• Lightspeed Restaurant is the right answer for European hospitality and for operators with multi-revenue-centre venues (hotel F&B, multi-concept groups) that need raw API access on the Premium tier. Outside Europe, Lightspeed’s strategic focus is real but its sales investment is not what it was 18 months ago, and the Upserve divestiture changes who an operator should call when something breaks at 11pm on a Friday. That asymmetry deserves to be on the table.
• Square for Restaurants is the right answer if you are a single-unit independent or a small group that values self-serve onboarding above analytical depth, and is willing to live without a Toast-style assistant.

Marcus has written about how acquirers value a restaurant’s data layer in his Bottom Line series — see Posts 15-18 for the framing on how POS choice maps to enterprise value. The short version: the AI surface of your POS is not just an operating tool, it is increasingly a piece of the multiple a buyer will pay for the business. Lightspeed’s narrower AI surface, read through that lens, is a more meaningful gap than the feature-count comparison suggests.

Reach for Lightspeed if…

• You are a European hospitality operator. This is where Lightspeed is investing, where it is winning, and where its product team’s attention sits. Multiple geographies’ sales teams are being staffed up against European hospitality specifically, per the Q3 FY2026 earnings call (Lightspeed Q3 FY2026 release).
• You run a multi-revenue-centre venue — hotel F&B, a venue with a restaurant and a bar and a coffee programme on separate cost centres, a multi-concept group — and you need clean cost-centre separation and raw API access. Premium tier exposes both.
• You are an independent who actively wants conversational analytics but does not need an AI assistant that takes in-app actions. The Lightspeed AI surface is genuinely useful for asking questions about your sales and menu and getting narrative answers. It is not, today, a tool that does things on your behalf.
• You want the most thoughtful menu-engineering surface of any major POS. Magic Menu Quadrant is the best version of “which dish do I cut” I have seen built directly into a POS. It is the feature inside Lightspeed AI I would most actively reach for.

Hold off on Lightspeed if…

• You are a U.S. independent who valued Upserve as your account team. Upserve is now a Skyview Equity company. The Lightspeed Restaurant platform remains, but the dedicated U.S. independent-segment investment Upserve represented inside Lightspeed has been re-pointed at Europe. This is not a knock on the new Upserve owners; it is just a different vendor relationship than the one you had a year ago.
• You need a labour scheduler in the Toast/7shifts/HotSchedules sense as part of the POS. Lightspeed does not have one as of May 2026, and the “staffing optimisation” capability inside Lightspeed AI is a recommendation surface, not a schedule generator.
• You run drive-thru QSR. Lightspeed does not seriously compete here. Toast’s April 2026 enterprise drive-thru launch (Toast/Businesswire) is the obvious counter-position.
• You are an enterprise operator who wants an AI assistant that can edit menus and modifiers, retire items, and adjust shift clockouts via chat. Toast IQ does that today (Toast IQ release). Lightspeed AI, in the company’s own framing, is on the path to autonomous-agent capabilities but is not there yet.

What this Desk Review is saying, plainly

Lightspeed Restaurant in May 2026 is a credible European hospitality platform with a tightly-scoped, recently-launched AI surface and a clear strategic decision to stop competing for the U.S. mid-market independent. The conversational AI layer (Lightspeed AI), the competitive-benchmarking module (Benchmarks & Trends, with Magic Menu Quadrant inside it) and the service-pacing layer (Tempo) are the three intelligence products an operator on Lightspeed should be evaluating right now. The labour-scheduling story is, at best, premature; treat any framing that suggests otherwise — including some of the framing I shipped earlier this year — with the scepticism it deserves.

The deeper read, the one that touches the Four Margins frame Eitan laid out: the highest-leverage operating-margin tool a POS can give you is labour optimisation, and that is the one place Lightspeed does not yet have the strongest offering. The places it does have an edge — menu engineering and service pacing — are operating-margin levers as well, but they move smaller numbers than scheduling does. That is the strategic gap. Whether it closes in 2026 will depend on how aggressively Lightspeed builds out the autonomous-agent capabilities it has telegraphed publicly. The framing on what a true AI agent in this category will look like, and why scheduling is the wrong first product for one anyway, is the subject of the Voice Agent Curve.

A correction note: an earlier draft of this piece, briefly available in mid-March, described a “Lightspeed AI labour scheduler” launched in February with operator-panel results. That draft was wrong on the feature’s existence. It has been retracted. This Desk Review is the replacement and represents what I can stand behind on public sources. Operators who saw the earlier piece and built decisions on it should re-read this one and adjust accordingly. The Vibe Check format will return for Lightspeed Restaurant when there is a real, shipping, panel-testable AI feature on the platform that warrants it.

— Maya covers vendor reviews and restaurant-tech. Tips: [email protected].