A panel discussion on artificial intelligence and enterprise software featured executives from Rimini Street and Zeta Global (NYSE:ZETA) arguing that AI is more likely to strengthen established, mission-critical platforms than abruptly replace them. The moderator framed the conversation around what he described as “overanxiety” in the software industry—particularly the idea that generative AI could rapidly “vibe code” replacements for complex enterprise systems.
Rimini Street and Zeta outline their businesses
Seth Ravin, representing Rimini Street, said the company supports “thousands of customers around the world” across government, military, and private industry, primarily serving large enterprises. He said Rimini Street’s work focuses on reducing total cost to serve through labor, systems, and processes, and more recently through “agentic AI solutions.” Ravin added that the firm keeps large transactional systems running longer than original vendors typically want to support them, describing customers’ preference for stability over frequent upgrades.
On company scale, Ravin said Rimini Street has “over $400 million in annual recurring revenue,” operates in 28 countries, and employs more than 2,000 people. Zeta’s executive said that “in the middle of our range this year,” the company expects $1.755 billion in revenue and “almost $400 million in EBITDA,” with “60%+ of that” expected to drop to free cash flow. He also said Zeta is on its “fourth consecutive year” of 30%+ compounded top-line growth, with EBITDA growth greater than 50% and free cash flow growth greater than 75% over that four-year period.
Why participants said big platforms won’t be replaced overnight
Ravin rejected the notion that large ERP platforms could be quickly recreated by AI, describing it as unrealistic to replace “all 27,000 processes” and “millions and millions of lines of code” that are audited and subject to global standards. He said enterprise system replacement will be “a gradual process,” emphasizing that many systems can run for “20, 30 years” while new capabilities are built “over the top.” He also said Rimini Street believes “ERP software…is dead” over the long term, but not imminently.
Zeta’s executive drew parallels to prior technology cycles, including the internet, mobile, and cloud computing, arguing that companies that adopt new technology can become stronger. He said thin applications—particularly workflow tools without proprietary data or deep integration—face greater risk of disintermediation, including from small teams using “vibe coding” to replicate simpler products.
Moats: proprietary data, integration, and governance
Asked about competitive moats, Zeta’s executive pointed to its data cloud and enterprise relationships. He said Zeta has “552 million” active people in its data cloud and that its first-party tracking pixel sits on “trillions of pages of content.” He described high-frequency computation in real time to infer consumer intent, such as whether a person intends to buy a car or churn off a wireless plan.
He also stressed privacy controls, stating that Zeta “never sell[s] our data” and “never share[s] the personally identifiable information,” even with enterprise clients, instead masking identities using a Zeta ID number. He said Zeta does not feed its data into large language models and uses proprietary models, adding that clients have built “thousands of AI agents” using the company’s AI Agent Studio. He cited a net retention rate of 120% last year as evidence of stickiness.
Ravin said organizations risk “AI washing” products by simply adding “.AI” to releases without substantive change, contributing to confusion among CIOs and CFOs. He argued many companies could see major benefits just by improving processes and automation, and said Rimini Street provides checklists and evaluation processes to help customers determine where AI agents deliver value. He described the current environment as frenetic, with many failed AI projects due to unclear objectives and immature execution.
Partnerships with OpenAI and ServiceNow focus on “agentic” layers
Zeta’s executive discussed an OpenAI partnership and said the industry is shifting from foundational large language models toward “agentic AI,” where smaller models sit on top of larger ones. He said Zeta and OpenAI have a joint engineering team working on “Athena,” a voice-enabled “super agent” designed to help clients navigate Zeta’s user interface through voice commands. He emphasized that Zeta is not exposing its data to OpenAI and is not using OpenAI foundational models for intent scoring.
He contrasted Athena with a prior voice interface called ZOE, which he described as “clunky,” while stating that clients who adopted ZOE spent “250% more” on Zeta’s platform. He said Athena’s goal is to reduce friction in using complex software interfaces—allowing users to request outcomes (such as adding incremental customers at a lower acquisition cost) and have the system guide and execute actions.
Ravin described a partnership with ServiceNow, saying ServiceNow and other large vendors are building AI “toolkits,” but customers “don’t want to buy toolkits” and instead want solutions. He said the partnership is aimed at applying agentic AI to ERP and other transaction systems to reduce labor costs and improve flexibility. He also described deploying “Band-Aids” that sit over existing ERP systems to address specific problems in “weeks instead of months or even years” required for traditional changes and upgrades.
AI’s operational impact: productivity, headcount, and costs
On internal adoption, Zeta’s executive said the company has tracked employee usage of AI tools and focused on productivity, including a claimed 125% increase in engineering productivity over the last 12 months. He said Zeta expects to grow the business “30%+” this year and finish with fewer employees than it started with, and suggested headcount could remain flat for years outside of M&A while organic revenue growth continues at “greater than 20% a year” for many years.
Ravin said Rimini Street will keep “humans in the loop” given its mission-critical responsibilities, including work tied to military operations, but is using AI to improve response and resolution times. He said Rimini Street can respond “within 70 seconds” globally with an engineer available 24/7, and that AI tools have reduced issue resolution cycle time by 28%. He added that the company expects to lower its headcount at the end of 2025 compared with 2024, while continuing to add employees at a slower rate due to technology leverage.
Both speakers also discussed the complexity of AI pricing models and the challenges customers face in understanding agent costs relative to human labor costs. They predicted organizations will likely use multiple AI platforms and models, and both suggested commoditization could occur outside a small set of major large language model winners.
The moderator concluded by suggesting that while AI-driven uncertainty has weighed on software valuations, the technology could become one of the most beneficial developments for “good software companies” by expanding revenue opportunities and improving operating leverage for platforms that are complex and mission-critical.
About Zeta Global (NYSE:ZETA)
Zeta Global, founded in 2007 and headquartered in New York City, is a leading data-driven marketing technology company. The firm’s mission centers on helping brands acquire, grow and retain customers through a unified customer lifecycle management platform. Over the years, Zeta Global has built a reputation for leveraging big data and predictive analytics to power digital marketing programs across multiple channels.
At the core of Zeta’s offering is the Zeta Marketing Platform, which combines identity resolution, audience insights and real-time engagement capabilities.
