Agilysys Quit Talking About the Scaffolding — and That's the Tell
Three guidance raises in a single fiscal year, 2,000+ Marriott properties live, and a keyword ledger that just swapped 'implementation efficiency' for 'growth path'.
AGYS · Earnings Call · 2026-10-26
Three raises and a change in vocabulary
Agilysys began fiscal 2027 pointing at $365–370M of revenue and at least 30% subscription growth. It has now raised both twice in six months, landing at $374–378M and at least 34% subscription growth, with adjusted EBITDA guidance nudging from 24% to 25% of revenue. Fiscal Q2 delivered a record $93.4M, up 17.7% year over year — the 19th consecutive record revenue quarter — on subscription revenue of $44.2M, up 32.6%, which for the first time crossed 70% of total recurring revenue. The financials behind it are less dramatic than the press release and more useful. The last filed quarter showed revenue of $88M, up 14% year over year and up 6% sequentially, with gross margin at 63.5% and operating margin of 12.3%. On this call, Dave Wood reported gross margin of 64.6% and operating income of $15.2M on $93.4M of revenue — a mid-teens operating margin, a visible step up from the filed quarter. Cash and marketable securities reached $137.6M with no debt. What's genuinely interesting is not the numbers; it's the vocabulary migration. For roughly two years, this company's own top keywords were the scaffolding of the story: the product ecosystem, implementation efficiency, and the large PMS rollout. Those three are now the hardest-falling keywords in Agilysys's own momentum history — each peaked and then decayed. In their place: growth path as the single freshest theme of the quarter, alongside shots on goal and big brand. That swap matters because it marks a company that no longer needs to explain why it should work. Ramesh Srinivasan can now say the domestic HRC hotels and resorts vertical had its best quarter ever by a fair distance, 23% above the prior peak, with a seven-figure-plus multi-property competitive replacement win spanning PMS, F&B inventory procurement and 15 additional modules. New customers averaged 5.6 products per deal; when PMS was in the suite, 9.3. There were 119 upsell instances to existing properties — the fourth straight quarter above 100 — averaging a record 2.8 products per agreement.Marriott: more mileage, less odometer
The Marriott project is where the disclosure has changed shape without getting clearer. Management now volunteers an absolute number where it previously declined to give one: “well more than 2,000 properties are live and the projects have gone well so far” — Ramesh Srinivasan, President and Chief Executive Officer · 2026-10-26. Select-service implementations are essentially complete; the full-service tranche is just beginning. But the percentage is the thing analysts want, and it is the thing they will not give.Nehal Chokshi of Northland tried the arithmetic directly — roughly 6,000 properties, ~2,000 done, so a third? — and was told to let the customer do that math. Note the framing shift against prior quarters. In July 2026 the answer was “we are still expecting the timeline to be somewhere in the next 18 to 24 months sort of completion time work is where we are working towards” — Ramesh Srinivasan, President and Chief Executive Officer · 2026-07-27; in May it was “It is going to be at least 2 years or possibly more than that” — Ramesh Srinivasan, Chief Executive Officer · 2026-05-18. The duration language has not tightened. Only the raw site count has. One disclosure detail deserves more attention than it got: essentially every sales metric on this call — ACV, backlog, ARR implemented — excludes the Marriott PMS subscription. Which means the record bookings and the near-record backlog are a non-Marriott number. That is a cleaner read on the underlying engine than the headline suggests, and it also means the Marriott revenue is being layered on top of a business already compounding at a high-20s/30s rate.In terms of percentage of completion... it varies, right, depending on whether you're looking at properties and/or rooms or revenue... we have not yet reached the halfway mark with any of them.
The monetization question, deferred again
The AI story is where Agilysys is most exposed to the gap between narrative and revenue. Thirty-plus features announced at the April user conference are partly in pilot, partly in production. The genuinely new operational fact this quarter is the AI Common Framework — a control plane for tokenization of personal data, LLM selection, usage monitoring and cost governance. That is an engineering answer to a business problem: AI features are expensive to run per-token, and nobody pays for them yet. Billy Fitzsimmons of Piper Sandler asked the right question about willingness to pay, and the answer was structurally evasive: “Some of these features will just be normal enhancements we do to our products... Some of the AI features will fall in the monetizable bucket” — Ramesh Srinivasan, President and Chief Executive Officer · 2026-10-26 — via SaaS fee uplift or transaction pricing. This is at least the third consecutive quarter of "too early." The CRS and Total Revenue Management System betas, now expected later this fiscal year and deliberately walled inside the ecosystem, are the first real test. The quarter's most informative surprise came from an unlikely line. “The hardware attach rate to POS sales continues to remain low” — Ramesh Srinivasan, President and Chief Executive Officer · 2026-10-26 — management opened the POS platform to Windows, iOS and Android, so customers can buy their own iPads. And yet product revenue hit roughly $12M, the highest in about three years, because POS volume swamped the mix. That's the difference between a mix problem and a demand problem, and this quarter it was demand.What the tape already knows
The stock has run 37.7% over the last 90 trading days into this print, yet it still sits 14.9% below its December 2024 high of $141.74. So this is a recovery, not a breakout — and a $3.5B market cap at 8.9x trailing revenue, down 20% year over year is a company whose multiple has de-rated even as the fundamentals improved. Guidance for a “38 to 39% growth rate for subscription” — Dave Wood, Senior Vice President and Chief Financial Officer · 2026-10-26 exit is aggressive against a $3.5B base and would ordinarily re-rate a software name. The offset is guidance shape: Q3 adjusted EBITDA is guided to only slightly north of 22% of revenue on unfavorable cost timing including professional fees, before snapping back to roughly 30% in Q4. Management's own history says the fiscal-year exit rate is a decent proxy for the following year — and in July they insisted. If that holds, the operating leverage argument is now a math problem rather than a faith problem. What to watch: whether the full-service Marriott tranche behaves like the select-service one, whether POS penetrates Marriott at all (management concedes it has not even scratched the surface), and whether the monetizable bucket gets a price before the CRS beta ships. Three guidance raises in one year buy patience for about two more quarters. The vocabulary shift says management knows it.no quantum increases in investments are needed anymore... we will just keep doing the incremental additions as we go along