Invention Disclosure · IDD Ref: DI-2026-NKK-0511-001
Agentic Artificial Intelligence System for Automated Software Requirements Analysis, Generation & Multi-Platform Export
Filed: 11 May 2026
·
Inventor: Nishanth Kumar Kishore
·
Digital Infotech
IDD Number
DI-2026-NKK-0511-001
Conception Date
11 May 2026
Disclosure Date
11 May 2026
Technology Class
G06F 40/30 · G06N 20/00
Status
● Disclosed — Pending Filing
Rights Owned By
Digital Infotech
Digital InfotechThis document constitutes an official Invention Disclosure establishing authorship and conception date of the described invention.
Nishanth Kumar Kishore
Inventor — Nishanth Kumar Kishore
01 — Abstract
Summary of the Invention
A multi-agent artificial intelligence system that transforms unstructured discovery-phase input — including stakeholder interviews, meeting transcripts, pain points, and business goals — into structured, platform-ready software requirements automatically and without human authoring.
The present invention discloses an Agentic AI Requirements Studio comprising a coordinated pipeline of six specialised artificial intelligence agents: a Discovery Parser, an Epic Generator, a Story Writer, an Acceptance Criteria Agent, an Estimation Agent, and a Platform Mapper. These agents operate in orchestrated sequence to produce complete software engineering artefacts — including epics, user stories, Given/When/Then acceptance criteria, Fibonacci story-point estimates, and priority classifications — formatted for direct import into Atlassian Jira and Rally (Broadcom).
03 — Patent Claims
Claims
1
Independent — SystemA computer-implemented agentic artificial intelligence system for automated software requirements generation, comprising: a discovery input module; an orchestration controller coordinating a sequential pipeline of specialised AI agents; and a platform export module emitting structured requirements data for import into at least one enterprise project management system.
2
Independent — MethodA computer-implemented method comprising: receiving unstructured discovery input; processing through a multi-agent AI pipeline; generating epics, user stories, acceptance criteria, and effort estimates; and outputting in Atlassian Jira or Broadcom Rally format.
3
Dependent — Discovery ParserThe system of claim 1, wherein the pipeline includes a Discovery Parser Agent that extracts themes, entities, pain points, business goals, and technical constraints from unstructured input using LLM inference.
4
Dependent — Epic GeneratorThe system of claim 1, wherein the pipeline includes an Epic Generator Agent producing epics with title, description, business value, priority {Critical, High, Medium, Low}, and component labels.
5
Dependent — Story WriterThe system of claim 1, wherein the pipeline includes a Story Writer Agent generating user stories in canonical format "As a [persona], I want [action] so that [benefit]" with domain-specific personas.
6
Dependent — Acceptance CriteriaThe system of claim 1, wherein the pipeline includes an Acceptance Criteria Agent producing Given/When/Then criteria per story, contextually specific to domain, persona, and action.
7
Dependent — EstimationThe system of claim 1, wherein the pipeline includes an Estimation Agent assigning Fibonacci story points {1,2,3,5,8,13} using domain-context-aware inference on complexity and technical risk.
8
Dependent — ExportThe system of claim 1, wherein the export module simultaneously emits Jira Cloud JSON (projects[].issues[]) and Rally Agile Central Feature/UserStory bulk upload JSON.
9
Dependent — OrchestrationThe method of claim 2, wherein a supervisor controller manages inter-agent handoffs, enforces schema validation, and provides real-time animated progress UI.
10
Dependent — TraceabilityThe method of claim 2, further comprising a bi-directional requirements traceability matrix linking each story to: (a) originating discovery input; (b) parent epic; (c) acceptance criteria.
05 — Novelty & Inventive Step
Prior Art Comparison
| Capability | This Invention | Jira / Rally | GPT / Copilot | Traditional BA |
|---|
| Multi-agent orchestrated pipeline | ✓ 6 specialised agents | ✗ | ✗ Single-shot | ✗ |
| Discovery input → structured backlog | ✓ Fully automated | ✗ Manual entry | ✗ Unstructured | ✗ Days of effort |
| Given/When/Then AC generation | ✓ Per-story, domain-specific | ✗ | ✗ Inconsistent | ✓ Manual |
| Fibonacci AI estimation | ✓ Context-aware | ✗ Manual | ✗ | ✗ Planning poker |
| Dual-platform Jira+Rally export | ✓ Simultaneous | ✗ Single platform | ✗ | ✗ |
| Real-time agent progress UI | ✓ Animated pipeline | ✗ | ✗ | ✗ |
IDD: DI-2026-NKK-0511-001
Inventor: Nishanth Kumar Kishore · Digital Infotech
Date: 2026-05-11T00:00:00Z
SHA-256: a4f8c2d1e9b3a752c8d2e1f4a9b6c3d7e2f5a8b1c4d7e0f3a6b9c2d5e8f1a4b7
Status: Invention Disclosure Filed — Provisional Patent Recommended
07 — Legal Notices
⚠ Not a Patent Application. This document does not grant patent rights. File a Provisional Patent Application (PPA) with USPTO or CIPO within 12 months to secure a priority date. The one-year statutory bar begins on the first public disclosure date.
Copyright: © 2026 Nishanth Kumar Kishore / Digital Infotech. All rights reserved.
Third-Party Marks: Jira and Confluence are registered trademarks of Atlassian Corporation. Rally is a trademark of Broadcom Inc. Claude is a trademark of Anthropic PBC. Referenced nominatively for descriptive purposes only.
Prior Art: The inventor represents the combination of features described constitutes novel subject matter not disclosed in any prior art as of the disclosure date.