Agilysys Inc AGYS
New risk paragraphs, branching off last year
Each branch is a paragraph in Risk Factors with nothing like it in the 10-K filed May 23, 2025. A box jumps to the paragraph as filed, below.
Paragraphs with nothing like them in last year's 10-K
14 paragraphs in Risk Factors of the 10-K filed May 21, 2026 with nothing like them in the 10-K filed May 23, 2025. Quoted exactly. Paragraphs that were only edited are under 10-K change.
We use AI in our platform and product offerings, and our success is dependent upon our ability to leverage data, develop competitive products, and manage related risks.
We utilize AI solutions both for internal productivity purposes and in products and services available to our customers. We also work with partners who have incorporated or may incorporate AI solutions in their products and services. While AI and machine learning present opportunities for enhanced productivity, AI also introduces cybersecurity, data privacy, IT, intellectual property, regulatory, legal, operational, competitive, reputational and other risks that could adversely impact our business. For example, if the content, recommendations or analyses that AI applications assist in producing are or are alleged to be deficient or inaccurate, or if the data used to train these applications are or are alleged to be deficient or
inaccurate or legally impermissible to train on, we could be subject to competitive risks, potential legal liability, and reputational harm. Furthermore, the integration of third-party AI models, including Large Language Models, within our products and services may rely, in part, on certain safeguards implemented by the third-party developers of the underlying AI models, including those related to the accuracy, bias, and other variables of the data, and these safeguards may be insufficient. In addition, AI may present evolving ethical issues. If our use of AI becomes controversial, we may experience reputational harm or legal liability.
Further, given the early stage of generative AI, the evolving regulatory landscape surrounding AI also poses a risk, as new laws and regulations could impose additional compliance burdens, resulting in increased operational costs to comply with U.S. and non-US laws concerning the use of AI. For example, the EU’s Artificial Intelligence Act, or AI Act, originally entered into force on August 1, 2024, and is expected to undergo amendments as introduced in the EU’s November 2025 Digital Omnibus. As enacted, the AI Act imposes significant obligations on providers and deployers of high-risk AI systems and encourages providers and deployers of AI systems to account for EU ethical principles in their development and use of these systems.
Likewise, within the United States, the regulatory landscape is rapidly evolving as evidenced by several U.S. states introducing or enacting legislation that is designed to govern the development or use of AI. Our ability to use and offer AI and machine learning solutions may be constrained by current or future laws, regulatory, or self-regulatory requirements. At the federal level, the Trump Administration has endorsed a federal moratorium on the enforcement of state AI laws, including through a December 11, 2025 executive order on “Ensuring a National Policy Framework for Artificial Intelligence.” Federal efforts have thus far have not been successful in curtailing state action on AI regulation, which has contributed to an increasingly complicated regulatory landscape. The rapid evolution of AI, including potential additional government regulation of AI and its various uses, may require significant resources for us to implement compliant and ethical AI practices. We may also be subject to significant enforcement actions or litigation in the event of any perceived or actual non-compliance.
The rapid adoption of AI technologies may create an “arms race” dynamic in our industry, in which competitors leverage AI to develop more efficient, automated, or lower-cost alternatives to our offerings or to disrupt traditional subscription-based software models. In addition, AI-enabled tools may bypass or reduce reliance on traditional application-layer workflows, which could diminish the value of our platform and adversely affect demand for our solutions.
We have integrated AI and machine learning features into our technology stack to improve guest analytics, demand forecasting, and automated workflows, among other things. These models are complex and their efficacy depends on the quality of the data used for training. If the underlying data is flawed, or if the AI models experience "hallucinations" or logical failures, our platforms may produce inaccurate, biased, or unreliable outputs. In a hospitality context, such errors—ranging from incorrect pricing recommendations to failures in guest service automation—could result in operational disruptions for our customers, contractual disputes, and harm to our brand reputation. The "black box" nature of certain AI models may also make it difficult for us to troubleshoot these errors or provide transparent explanations to our users and regulators.
Our solutions are primarily delivered through multi-tenant cloud architectures in which multiple customers are served through shared environments. Although we employ controls designed to logically segregate customer data, any failure, vulnerability, or misconfiguration could result in unauthorized access to or exposure of one customer’s data to another, which could lead to legal liability, regulatory scrutiny, and reputational harm.
We also rely on identity and access management controls to govern user access across our systems. Risks associated with overprivileged accounts or accounts not centrally managed through our single sign-on systems could create vulnerabilities. Any failure to promptly revoke or appropriately manage access credentials, including those of former personnel or compromised accounts, may result in unauthorized access or data exfiltration.
In addition, our services depend on integrations with third-party applications through APIs, OAuth tokens, and similar authentication mechanisms. Mismanagement, compromise, or misuse of such credentials could enable unauthorized access and allow threat actors to move laterally across interconnected systems, increasing the scope and impact of a cybersecurity incident.
The global regulatory landscape surrounding AI is rapidly evolving and remains highly uncertain. We are subject to emerging privacy and data protection mandates, which may impose stringent requirements on how we develop and deploy AI models. Compliance with these regulations requires substantial financial resources and management attention. Additionally, there is significant legal uncertainty regarding the use of proprietary or third-party data for AI training and the ownership of AI-generated content. We may face claims of intellectual property infringement or find that we are unable to protect our own AI-driven innovations. Furthermore, if our AI processes inadvertently process sensitive guest or employee data in a manner that violates privacy standards (such as GDPR or CCPA), we could be subject to fines, litigation, and regulatory enforcement actions.
AI introduces new and sophisticated cybersecurity vulnerabilities. Malicious actors are increasingly using AI to automate the discovery of software vulnerabilities and to launch highly targeted social engineering or "deepfake" attacks against our employees and customers. Additionally, the AI models we use may be subject to novel attack vectors, such as "prompt injection" or "data poisoning," where attackers attempt to manipulate the model's logic or access underlying training data. Any breach that compromises our AI infrastructure or the sensitive data utilized therein could result in the loss of intellectual property, material financial costs for remediation, and loss of customer trust in the security of our products.
Show the other 2
Because our cloud-based offerings depend on third-party hosting providers, we are also exposed to risks relating to service availability and performance. We may have contractual service level agreements (“SLAs”) requiring specified uptime and performance standards, and any failure to meet these standards, whether attributable to us or our third-party providers, could result in service credits, financial penalties, customer claims, or contract terminations.
While we do not believe that our products and services infringe any patents or other intellectual property rights, from time to time, we receive claims that we have infringed the intellectual property rights of others.
Paragraphs in the 10-Q filed Oct 26, 2026 that were not in the one before
New in the latest 10-Q
11 paragraphs in the 10-Q filed Oct 26, 2026 with nothing like them in the 10-Q filed Jul 27, 2026 (at least half the words must differ; new numbers alone do not count). Quoted exactly, grouped by the item heading above them. This is text that changed, not a judgement of how serious it is.
10-Q
of other grants Other common stock issuances, net 22 6 710 716 Net income 22,169 22,169 Unrealized translation adjustments (3,493) (3,493) Balance at September 30, 2026 33,342 $ 10,003 (5,127) $ (1,539) $ 143,883 $ 221,940 $ (16,554) $ 357,733 Six Months Ended September 30, 2025 Common Shares Issued In Treasury Capital in Accumulated
Read in the 10-Q
The following table summarizes the activity during the six months ended September 30, 2026 for SSARs awarded under our Equity Incentive Plans:
Read in the 10-Q
(In thousands, except share and per share data) Number of Weighted-Average Remaining Aggregate
Read in the 10-Q
On September 1, 2026 (the "Acquisition Date"), we acquired all the issued and outstanding shares of Gaming Hospitality Solutions, Inc. (“GHS”), a hospitality software company based in Denver, Colorado. GHS is now a wholly-owned subsidiary of Agilysys, Inc. The condensed consolidated financial statements include the results of GHS’s operations since the Acquisition Date. The acquisition expands the opportunity to increase our solutions-per-customer.
Read in the 10-Q
The purchase price consisted of $0.9 million of cash consideration paid at closing, funded with cash on hand, and $0.2 million in contingent consideration liabilities representing the fair value of estimated amounts payable to the seller. We allocated the purchase price for GHS to the intangible and certain tangible assets acquired and certain liabilities assumed based on their estimated fair values on the Acquisition Date, with the remaining unallocated purchase price recorded as goodwill.
Read in the 10-Q
Show the other 6
We determined the fair value assigned to the identifiable intangible asset acquired, customer relationships, by using the income approach, which discounts the expected future cash flows to present value using estimates and assumptions determined by management.
Read in the 10-Q
We assigned the acquired customer relationships an estimated useful life of 10 years. The identifiable intangible asset acquired amortizes on a straight-line basis, which we believe approximates the pattern in which the asset is utilized, over its estimated useful life.
Read in the 10-Q
The goodwill recognized in the GHS purchase price allocation is attributable to synergies in products and technologies to serve a broader customer base. The total goodwill recognized in the acquisition amounted to $0.4 million, which is not deductible for income tax purposes.
Read in the 10-Q
We have prepared the GHS purchase price allocation on a preliminary basis. Changes to the allocation may occur as additional information becomes available during the measurement period (up to one year from the Acquisition Date). The primary areas that remain preliminary include, but are not limited to, the intangible asset including the initial assumptions used in the estimate of fair value and the respective estimated useful life, income taxes, and residual goodwill.
Read in the 10-Q
Revenue and Net Income attributable to GHS included in our condensed consolidated statement of operations was not material.
Read in the 10-Q
Other information
Rule 10b5-1 Trading Plans On July 29, 2026, Ramesh Srinivasan, our President and Chief Executive Officer, terminated a written trading plan intended to satisfy the affirmative defense of Rule 10b5-1(c) previously adopted on August 28, 2025. Prior to termination, the trading arrangement was scheduled to expire on August 27, 2026, and covered the potential sale of up to 133,447 shares of the Company's common stock. Except as set forth above, during the fiscal quarter ended September 30, 2026, no director or officer (as defined in Rule 16a-1(f)) adopted or terminated (including any modification of) a Rule 10b5-1 trading arrangement or a non-Rule 10b5-1 trading arrangement.
Read in the 10-Q
The 10-Q's risk factors against the 10-K
Item 1A of the 10-Q filed Oct 26, 2026 against Item 1A of the 10-K filed May 21, 2026.
A 10-Q repeats only the risks that changed, so a 10-K paragraph that is missing from it is not counted as removed.
Added is a paragraph with nothing like it in the 10-K, Reworded a sentence that is in both with several words changed,
Quiet changes a sentence in both where one to four words changed and one is a hedge or a modal
(cut words struck through, new words highlighted).
Text on the page, not a score; each links to the filing at sec.gov.
What the 10-Q itself says about its risk factors:There have been no material changes in the risk factors included in our Annual Report for the fiscal year ended March 31, 2026 that may materially affect our business, results of operations, or financial condition.
Read in the 10-Q
1 paragraph in the 10-Q's Item 1A, and none of them adds, rewords or quietly edits a paragraph of the 10-K's.
Phrases used for the first time
Two-word phrases in Risk Factors of the 10-K filed May 21, 2026 that are in none of the 1 earlier 10-Ks stored, back to May 23, 2025
Where each one is used
- regulatory landscape: Further, given the early stage of generative AI, the evolving regulatory landscape surrounding AI also poses a risk, as new laws and regulations could impose additional compliance burdens, resulting in increased operational costs to comply with U.S. and non-US laws concerning the use of AI. For example, the EU’s...
- landscape surrounding: Further, given the early stage of generative AI, the evolving regulatory landscape surrounding AI also poses a risk, as new laws and regulations could impose additional compliance burdens, resulting in increased operational costs to comply with U.S. and non-US laws concerning the use of AI. For example, the EU’s...
- rapidly evolving: Likewise, within the United States, the regulatory landscape is rapidly evolving as evidenced by several U.S. states introducing or enacting legislation that is designed to govern the development or use of AI. Our ability to use and offer AI and machine learning solutions may be constrained by current or future...
- multiple customers: Despite testing prior to the release and throughout the lifecycle of a product or service, our on-premise and cloud-based solutions sometimes contain coding or configuration errors that can impact their function, performance and security, and result in other negative consequences. The detection and correction of...
- service level: Despite testing prior to the release and throughout the lifecycle of a product or service, our on-premise and cloud-based solutions sometimes contain coding or configuration errors that can impact their function, performance and security, and result in other negative consequences. The detection and correction of...
- level agreements: Despite testing prior to the release and throughout the lifecycle of a product or service, our on-premise and cloud-based solutions sometimes contain coding or configuration errors that can impact their function, performance and security, and result in other negative consequences. The detection and correction of...
Used at least twice, small words left out. Counted, not picked: a phrase can be new because the company renamed something.
What was taken out
2 paragraphs of last year's Risk Factors with nothing like them now. The longest are quoted.
Sources
Item 1A of each 10-K and 10-Q on EDGAR, quoted exactly. Each paragraph links to the filing at sec.gov.
Also
10-K change: what was reworded or cut, section by section.
Work with it
Excel workbook · CSV. Google Sheet that stays up to date: paste =IMPORTDATA("https://tickdelta.com/company/AGYS/export.csv") into a cell.