Dayos Selected for Singapore IMDA's Generative AI Agent Governance Framework Case, Demonstrating Architecture That Replaces ERP Operations Costing Over $1 Million Annually

Singapore — May 22, 2026 — Dayos Pte. Ltd., a Singapore-headquartered enterprise AI automation company, today announced that it has been selected as one of 14 industry case studies in the Info-communications Media Development Authority of Singapore's (IMDA) Model AI Governance Framework for Agentic AI (Version 1.5), released on May 20, 2026. The framework was launched at the ATxSummit with an opening address by Mrs Josephine Teo, Singapore's Minister for Digital Development and Information.

What it means for CFOs: A watershed moment for reassessing enterprise AI spending

For chief financial officers, this framework marks a turning point in how enterprise AI spending is evaluated. It is the first government-issued standard to clearly define what enterprise systems can truly be called 'agentic.' Systems that meet this standard have an economic logic fundamentally different from the 'workflow automation' category that has dominated enterprise back-office support for the past two decades.

The cost structure CFOs already know

A typical enterprise spends over $1 million annually to keep a single Oracle Cloud or Workday tenant running through traditional managed services. The root of this spending: underlying implementation costs range from $5 million to $40 million, and the support model maintaining these systems is built on senior consultants billing by the hour. Across an installed base of more than 16,000 Oracle, Workday, SAP, and ServiceNow customers, cumulative implementation sunk costs total approximately $320 billion. The workforce that originally built these implementations is still handling and billing for the resulting tickets today.

Based on publicly available local government Oracle Cloud contracts, the cost structure per ticket is clear: the average cost of an enterprise support ticket is approximately $3,233, with a low end of $1,744 and a high end of $4,722. About 30% of each ERP system's operational reliance depends on manual workarounds that cannot be self-serviced, and the labor pipeline supporting this work is shrinking. Globally, about 2 million computer science students graduate each year, but almost none of them are trained in Oracle, Workday, or SAP.

Dayos' agentic platform, Hero, is designed to replace both layers of this cost structure simultaneously. Hero Starter is priced at $60,000 per year (fixed fee), with AI-resolved tickets free of charge. The cost of one year of Hero Starter is lower than the cost of 20 enterprise support tickets at industry average pricing. Hero Enterprise fully replaces the original ERP implementation and comes with embedded forward-deployed engineers, priced at a one-time cost of $1.5 million to $3 million, delivering the same scope as projects quoted by traditional system integrators at $5 million to $40 million.

Dayos case study in the framework

The Dayos case study, included under the framework's 'Assess and bound the risks upfront' dimension, documents the company's retirement of its own ServiceNow ITSM instance within 45 days, reducing annual licensing costs by $121,000. The replacement is an AI ticket agent built on Hero, which employs three different reasoning strategies based on ticket risk:

• Tier 1 (60% of tickets):Low severity, fully reversible. Handled fully automatically by the agent. Reasoning chains and confidence scores are audited every two weeks.
• Tier 2 (30% of tickets):Medium severity, partially reversible. The agent handles diagnosis and proposes fixes, but requires sign-off from a qualified engineer before execution.
• Tier 3 (10% of tickets):High severity, limited reversibility. The agent does not touch these tickets.

This tiering mechanism aligns with the framework's recommendations on bounding agent autonomy based on severity, reversibility, and feasibility of human oversight.

A public procurement checklist

'The economics of enterprise back-office support have not been re-examined in thirty years,' said Brad McElhannon, founder and managing director of Dayos. 'System integrators charge $5 million to $40 million to build Oracle or Workday implementations. Managed service providers then charge $1 million annually, ticket by ticket, to keep them running. Hero replaces both layers with software economics. The IMDA framework gives financial leaders the technical standard to distinguish which AI vendors are truly capable of this and which are still selling human services.'

Dayos holds SOC 2 Type 2 and ISO 42001:2023 certifications and has been approved as a vendor under Singapore's IMDA 'GenAI x Digital Leaders' program. The program enables eligible Singaporean enterprise customers to receive government subsidies of 30% to 50% on AI projects.

About Dayos

Dayos is a Singapore-headquartered enterprise AI automation company with operations in the United States. The company builds Hero, an agentic platform deployable on Oracle, Workday, and other enterprise systems to automate workflows in IT management, accounting, human resources, and procurement. Founded in 2020 by Brad McElhannon, former head of financial engineering at Robinhood, Dayos serves customers in financial services, real estate, logistics, and retail. The company holds SOC 2 Type 2 and ISO 42001:2023 certifications.

For more information, visit dayos.ai.