Reference

Glossary of SaaS, AI and delivery terms

Plain-English definitions of the terms you will meet when commissioning software in Asia-Pacific — written for decision-makers, not engineers.

Agentic AI

AI systems that plan and execute multi-step work toward a goal — searching, deciding, drafting and acting across tools — rather than answering a single prompt. Production agentic systems keep humans in the approval loop, as Ava does inside Clear Talent.

Generative AI

AI models that produce new content — text, summaries, code or images — from a prompt. In enterprise software it typically appears as embedded assistants, document drafting and analysis features powered by large language models.

LLM (Large Language Model)

A machine-learning model trained on large volumes of text that can read, summarise, classify and generate language. LLMs power modern AI assistants and are integrated into business platforms through APIs.

RAG (Retrieval-Augmented Generation)

A technique that grounds an LLM's answers in your own documents or data: relevant content is retrieved first and supplied to the model, which reduces fabrication and keeps answers specific to your business.

Multi-tenancy

A SaaS architecture in which one application securely serves many customer organisations (tenants), with each tenant's data isolated. It is what makes SaaS economics work — one codebase, many customers.

SaaS (Software as a Service)

Software delivered from the cloud on a subscription basis, rather than installed and licensed per machine. Customers access it through a browser or app while the provider operates the infrastructure.

MVP (Minimum Viable Product)

The smallest version of a product that can be put in front of real users to test demand. A well-built MVP sits on production-grade foundations so it can grow into the full product without a rebuild.

Dedicated development team

A stable group of engineers, designers and QA who work exclusively on your product under your direction, employed and managed by a partner such as Seditio Asia. Contrasts with project-based outsourcing, which ends at handover.

Offshore / nearshore development

Building software with a team in another country. 'Nearshore' implies adjacent time zones. For Asia-Pacific companies, a Philippine team is effectively nearshore: Cebu shares UTC+8 with Singapore and Hong Kong and overlaps the Australian business day.

Data residency

Where data is physically stored and which country's laws govern it. Several Asian markets have residency preferences or requirements — a key architectural decision for regional SaaS products.

Serverless / cloud-native

An architectural approach in which applications run on managed cloud services (such as Google Cloud Run or AWS Lambda) that scale automatically, rather than on fixed servers you maintain. Reduces operational overhead and improves resilience.

Legacy modernization

Incrementally replacing or re-platforming ageing software while the business keeps running — typically by stabilising the existing system, then migrating capabilities in phases rather than attempting a risky big-bang rewrite.

PWA (Progressive Web App)

A web application that behaves like a native mobile app — installable, fast, and capable of working offline — without requiring app-store distribution. Used, for example, in the ManagedPMO tenant experience.

Systems integration

Connecting software systems so data and processes flow between them — for example linking a booking platform to logistics operations, or a training website to SAP accounting, both of which Seditio has delivered.

UAT (User Acceptance Testing)

The pre-launch phase where the client's own team exercises the software against real-world scenarios and formally accepts it. In Seditio's process, UAT precedes every production go-live.

Pharmacovigilance

The science of monitoring medicines for safety signals after release. AI-driven pharmacovigilance platforms such as Biologit screen millions of medical publications to surface potential adverse events.