Human-led. AI-powered. Built to scale.
CastleSoft is a founder-led software business. Andrew Tierney provides the leadership and direction, drawing on decades of software engineering experience, while a coordinated team of specialist AI agents supports delivery and day-to-day operations.
Founder | CEO
Andrew founded CastleSoft in 1995 and brings 38 years of IT expertise, beginning his career as a programmer in 1988. He holds a Bachelor of Mathematics with Computer Science from the University of Newcastle, Australia, and leads CastleSoft’s vision of secure, inclusive and privacy-first agentic AI.
We use specialist AI agents across engineering, data, security, finance, reporting, people and culture, legal, sales, customer support, operations and health. They are software systems, not employees. Each has a name so its area of responsibility is easy to identify, and human oversight remains in place. AI agent portraits are AI-generated representations.
AI agent
AI CTO Agent
Daniel drafts architecture options, reviews code, tracks security patches and emerging technology, and helps plan engineering work across CastleSoft AI, Health and Studios, flagging technical risks early.
AI agent
AI Engineering Agent
Sophie writes and refactors code, creates unit and integration tests, and runs build pipelines, preparing changes for review and release across our cloud and on-premises solutions.
AI agent
AI Data & Database Agent
Henrik designs and maintains database schemas, tunes queries and indexes, prepares migration scripts, and monitors backups and data integrity across our SharpCoreDB and PostgreSQL databases.
AI agent
AI Security & Compliance Agent
Lukas runs our weekly automated security audits, scans systems for known vulnerabilities, checks our ISO/IEC 27001-aligned controls and flags unusual activity in system logs for investigation.
AI agent
AI Finance Agent
Emma supports our bookkeeping by categorising transactions and drafting invoices, prepares GST and BAS working papers, forecasts cash flow and gathers supporting records for R&D Tax Incentive claims.
AI agent
AI Reporting & Insights Agent
Olivia builds dashboards and executive summaries, turning operational, project and financial data into clear reports, trends and key metrics so performance and priorities are easy to understand at a glance.
AI agent
AI People & Culture Agent
Grace drafts job advertisements, summarises applications against published selection criteria, prepares onboarding checklists and answers common policy and leave questions using our documented workplace policies.
AI agent
AI Legal Agent
Rachel prepares first drafts and plain-English summaries of contracts, NDAs and policies, and tracks changes to privacy and AI law in Australia, Singapore, the EU, the UK and the US.
AI agent
AI Sales & Marketing Agent
Jack researches potential clients and markets, drafts proposals, website copy and social media content, and keeps our customer relationship management (CRM) records accurate and up to date.
AI agent
AI Customer Support Agent
Sofia answers common customer questions, triages incoming support requests, keeps our knowledge base current and promptly routes complaints, refunds and privacy requests to the right person.
AI agent
AI Operations Agent
Liam coordinates work across the other agents, schedules tasks, takes meeting notes and prepares weekly status reports so priorities, progress and blockers stay visible to the business.
AI agent
AI Health Domain Agent
Ethan analyses healthcare workflows, researches relevant Australian health standards, and summarises findings and documents requirements to support the planning and design of CastleSoft Health solutions.
AI helps us work faster, but it must stay accountable. These principles shape how we govern our AI agents.
AI prompts, outputs and key workflow actions are logged, so work can be traced and audited. Logging follows privacy, confidentiality and access controls, with defined retention, and excludes secrets and unnecessary personal information.
Separate AI review agents check other agents’ work for errors, inconsistencies, security concerns and departures from requirements. This is an extra layer of quality control, not a guarantee of correctness or a substitute for human review.
Our engineering work follows documented processes, coding standards and established practices. Automated and unit testing, code review and security checks support software quality and release readiness. They reduce risk; they cannot remove it entirely.
We address the privacy, security and AI obligations that apply where we operate, including Australia’s Privacy Act 1988, Singapore’s PDPA, the EU GDPR, the UK GDPR and applicable US federal and state laws. These differ, so we treat each on its own terms. Our security practices are aligned with ISO/IEC 27001 (we are not certified), and we use ISO/IEC 42001 as guidance for managing AI.
CastleSoft