Generation of Requirements (Use Cases, User Stories, and Epics)
This activity involves capturing and documenting functional and non-functional requirements for an IT system. It includes gathering input from stakeholders, understanding business goals, and translating these into structured artefacts such as Use Cases, User Stories, and Epics.
Maturity Levels
|
Level |
Name |
Description |
Technology |
Example tools |
| 0 | Non-existent | No AI support. User stories, epics, and use cases are created entirely manually. | None | Traditional documentation tools (Word, Excel, Confluence) |
| 1 | One-Off Assist | Analysts occasionally use general-purpose LLMs to draft user stories or epics, but outputs are not standardized or integrated into project workflows. | Off-the-shelf LLMs, prompt engineering | chatGPT, Claude, Gemini, Mistral AI |
| 2 | Integrated Assist | AI capabilities are integrated into project management tools to suggest, version, and align user stories and epics with project objectives. | LLMOps, AI Agents, NLP | Gem (Gemini), Artifacts (Claude), Jira/Confluence Cloud with Rovo, MS365 Copilot (Word, SharePoint) |
| 3 | AI-Human Collaboration | AI collaborates dynamically with analysts to refine and contextualize requirements based on project data and feedback loops. | Agentic frameworks, Orchestration, Vector DBs | Langgraph, Retrieval frameworks integrated with project platforms |
| 4 | Full Autonomy | AI generates, maintains, and evolves user stories and epics from live project data and stakeholder inputs, ensuring continuous alignment with business goals. | Autonomous agents, Causal-inference models, Continual learning | End-to-end project orchestrators, autonomous PM platforms |
| AI Maturity Level: Indicates the level the technology vendors claim to have reached in deploying AI solutions that actually work in real-world applications |
Quality Check of Requirements
Quality checking ensures that requirements are correct, complete, unambiguous, consistent, and testable. The process verifies alignment with business objectives, validates terminology, reviews dependencies, and identifies gaps or contradictions within or between artefacts.
Maturity Levels
|
Level |
Name |
Description |
Technology |
Example tools |
| 0 | Non-existent | No AI support. Requirements are reviewed manually without automated validation. | None | Manual review in Confluence, Word, or Excel |
| 1 | One-Off Assist |
Analysts occasionally use general-purpose LLMs to improve or critique requirements. Results are not standardized or tracked. |
Off-the-shelf LLMs, prompt engineering | chatGPT, Claude, Gemini, Mistral AI |
| 2 | Integrated Assist | AI capabilities are embedded into project management tools to validate requirements against defined quality standards and maintain version control. | AI Agents, LLMOps, NLP | Custom GPTs (chatGPT), Gem (Gemini), Artifacts (Claude), ScopeMaster, Jira/Confluence Cloud with Rovo, MS365 Copilot (Word, SharePoint) |
| 3 | AI-Human Collaboration | AI collaborates with analysts to validate and refine requirements using contextual project data, dependencies, and business rules. | Agentic frameworks, Orchestration, Vector DBs | Langgraph, integrated validation frameworks |
| 4 | Full Autonomy | AI autonomously reviews, updates, and corrects requirements based on live project data and evolving business logic. | Autonomous agents, Causal-inference models, Continual learning | End-to-end project orchestrators, autonomous quality assurance systems |
| AI Maturity Level: Indicates the level the technology vendors claim to have reached in deploying AI solutions that actually work in real-world applications |
Assistance in Structuring Analysis Documents and Documentation
This activity focuses on organizing and structuring deliverables such as reports, specifications, and technical documentation throughout the Project and Service Lifecycle. It involves determining the appropriate document layout, defining sections, ensuring logical flow, and maintaining coherence across multiple documents.
Maturity Levels
|
Level |
Name |
Description |
Technology |
Example tools |
| 0 | Non-existent | No AI support. Structuring and formatting of analysis or technical documentation are done manually and vary by analyst. | None | Manual editing in Word, Confluence, or Markdown editors |
| 1 | One-Off Assist | Analysts occasionally use general-purpose LLMs to get suggestions for document structure or content summaries. Outputs are ad hoc and not standardized. | Off-the-shelf LLMs, Prompt engineering | chatGPT, Claude, Gemini, Mistral AI |
| 2 | Integrated Assist | AI is embedded in documentation and project tools to suggest structure, formatting, and initial drafts using project artefacts (requirements, design docs, code annotations). | AI Agents, LLMOps, NLP | Custom GPTs (chatGPT), Gem (Gemini), Artifacts (Claude), Jira/Confluence Cloud with Rovo, Notion AI, MS365 Copilot |
| 3 | AI-Human Collaboration | AI collaborates with analysts to organize, merge, and contextualize documents from multiple sources (repositories, analysis, code) into cohesive deliverables. | Agentic frameworks, Orchestration, Vector DBs | Langgraph, Retrieval frameworks integrated with Confluence, GitHub Copilot for Docs |
| 4 | Full Autonomy | AI autonomously structures, maintains, and updates documentation sets (user manuals, developer guides, API docs, runbooks) using live data from requirements, designs, and code repositories. | Autonomous agents, Causal-inference models, Continual learning | End-to-end documentation orchestrators, Autonomous doc generation systems |
| AI Maturity Level: Indicates the level the technology vendors claim to have reached in deploying AI solutions that actually work in real-world applications |
Generation of Analysis Models (Business Processes, Sequence Diagrams, Data Models)
This activity covers the creation of visual models that describe system behavior, data flows, interactions, and structures. Business requirements are translated into graphical representations such as BPMN diagrams, UML sequence diagrams, and conceptual data models.
Maturity Levels
|
Level |
Name |
Description |
Technology |
Example tools |
| 0 | Non-existent | No AI support. Analysis models (BPMN, UML, ER diagrams) are created manually from text inputs or business notes. | None | Draw.io, Lucidchart, Visio, Enterprise Architect |
| 1 | One-Off Assist | Analysts occasionally use LLMs to generate draft diagrams or model descriptions from text. Outputs are not standardized or reusable. | Off-the-shelf LLMs, Prompt engineering | chatGPT, Claude, Gemini, Mistral AI, ScopeMaster, Mermaid (manual integration) |
| 2 | Integrated Assist | AI is integrated into modeling tools to transform structured text inputs or requirements into consistent models and diagrams. | AI Agents, LLMOps, NLP, Code-to-diagram converters | Custom GPTs (chatGPT), Gem (Gemini), Artifacts (Claude), Mermaid |
| 3 | AI-Human Collaboration | AI collaborates with analysts to iteratively refine, validate, and synchronize models with evolving requirements and architecture context. | Agentic frameworks, Vector DBs | Langgraph, Model Orchestrator APIs |
| 4 | Full Autonomy | AI autonomously generates, maintains, and updates business, sequence, and data models from live documentation, system behavior, and code repositories. | Autonomous agents, Causal-inference models, Continual learning, Model synthesis engines | End-to-end modeling orchestrators |
| AI Maturity Level: Indicates the level the technology vendors claim to have reached in deploying AI solutions that actually work in real-world applications |