Research Intelligence

AURORA™

AI-Unified Research Optimization & Reasoning Architecture

Design Principle

A dataset is not a study. Research intelligence is what turns one into the other.

AURORA™ converts research data into structured study portfolios, assigns each study its correct analytical method, estimates sample size through SSE, and acts as the shared research intelligence layer behind AMI™, AIMA™, and PRN™.

Study Discovery Method Mapping Sample Size (SSE) Portfolio Output Ecosystem Intelligence Layer

See Capabilities → Ecosystem Map Request Demo

Discovered Studies

SSE — Sample Size

218

Required n (S1)

0.80

Target power

The Research Gap

Most research pipelines lose value between data collection and study design.

Datasets arrive rich in signal but leave without structure. Teams spend weeks debating what to study, which method fits, and whether they have enough data — AURORA™ answers all three, automatically.

Without AURORA™ — structural loss points

Dataset analysed for one study; 6 publishable questions left undiscovered
Method selection by default or committee — no statistical justification documented
Sample size calculated post-hoc, if at all — power assumptions informal
Each study designed independently — no portfolio view across the dataset
Analytical decisions locked to the analyst who made them
Integration between research and operational tools done manually

With AURORA™ — research portfolio intelligence

One dataset surfaces multiple defensible study pathways automatically
Method assignment uses the same rule-trace logic as MedStat Guide™
SSE connects study design intent to sample size and feasibility before analysis begins
Full portfolio dashboard: every proposed study with title, objective, and method plan
Study logic stored in structured, reproducible output — not in one analyst's memory
Native integration layer for AMI™, AIMA™, and PRN™ with no additional setup

What AURORA™ Can Do

Four modular intelligence blocks. One coherent research platform.

Each component can appear independently on the analytical surface, but all four work as a single reasoning chain — from raw input to portfolio-ready output.

Aurora Engine — Ingest, Frame & Scope

Reads available variables and domain context from a dataset or structured study description. Identifies the broad scientific or analytical scope and coordinates every downstream reasoning component.

Variable scanDomain framingScope mapping

Study Graph — Multi-Study Discovery

Transforms one dataset into multiple coherent proposed studies, each with a structured title, clear objective, outcome type, variable logic, and reporting direction. No study left undiscovered.

Study framingObjective logicOutcome typing

Method Mapper — Statistical Alignment

Assigns the right statistical model, assumption checks, and sensitivity plan to each proposed study. Analytical design becomes explicit, reproducible, and defensible — not dependent on the analyst in the room.

Model assignmentAssumption checksRule trace

AURORA™ SSE — Sample Size Estimator

Connects planned study design to power logic by inferring outcome structure, model type, and expected effect pattern — then calculates the required sample size pathway before a single line of analysis is run.

Power planningEffect inferenceFeasibility check

Portfolio Dashboard — Research-Ready Output

Presents all discovered studies in a structured, executive-readable dashboard. Supports planning, peer review, and downstream manuscript development. Handoff-ready without additional formatting.

Multi-study viewReview-readyExport logic

Ecosystem Layer — Reusable Intelligence Core

Operates as a plug-in reasoning layer behind AMI™, AIMA™, and PRN™. Converts operational and governance datasets into structured analytical study plans using the same four-engine pipeline.

AMI™ readyAIMA™ readyPRN™ ready

The 5-Step Reasoning Chain

From raw data to publication-ready portfolio in five structured stages.

AURORA™ enforces a decision-first sequence that makes every step traceable, every method choice documented, and every output reusable without the original investigator.

Ingest & Frame

Dataset or study description ingested. Variables scanned, domain context captured, analytical scope established.

Study Discovery

Study Graph surfaces all publishable research questions embedded in the data. Each gets a title, objective, and outcome type.

Method Assignment

Method Mapper assigns the right statistical model, assumption checks, and sensitivity plan to each proposed study.

Power Planning

SSE connects each study's design intent to sample size requirements and feasibility — before analysis begins.

Portfolio Output

All studies assembled into a structured dashboard: titles, objectives, method plans, SSE logic, and reporting direction.

Decision-first architecture

AURORA™ does not begin analysis. It designs the analysis. Every output is a structured research plan — not a result — so that when modelling begins in MedStat Guide™ or PRN™, the design is already defensible.

System Architecture

A modular, decision-first architecture — readable at every layer.

AURORA™ is built in four explicit reasoning layers. Each layer produces a documented output that can be reviewed, challenged, and reused independently of the layers below it.

Input

Dataset

Study Description

→

Variable scan

Domain framing

Scope identification

Discovery

Aurora Engine

→

Study Graph

→

Study titles

Objectives

Outcome types

— multiple studies per dataset

Design

Method Mapper

→

Model assignment

Assumption checks

Sensitivity plan

Rule trace

Power

AURORA™ SSE

→

Effect inference

Sample size pathway

Feasibility check

Power curve

Output

Portfolio Dashboard

→

Review-ready pack

AMI™ handoff

PRN™ study input

AIMA™ governance

Ecosystem Integration

AURORA™ powers AMI™, AIMA™, and PRN™ — the same research core, three contexts.

Rather than building separate intelligence layers into each product, AURORA™ functions as a reusable analytical design engine — converting domain-specific data into structured analytical studies and statistically grounded planning logic.

AMI™

Ayati Mix Intelligence

Monitoring upgraded to investigation

In AMI™, AURORA™ upgrades analytics maturity from monitoring to investigation. Instead of only reporting indicators, the platform can frame decision questions, assign methods, and generate structured study plans behind operational datasets.

Converts dashboards into research-grade diagnostic frameworks

AIMA™

Ayati AI Maturity Assessment

Governance through defensible study logic

In AIMA™, AURORA™ structures the analytical design needed to test AI performance, governance quality, and readiness. Moves organisations beyond checklist evaluation into evidence-grade assessment with full SSE backing.

Adds structured study rigour to every governance checkpoint

PRN™

Praxis Research Network

Design engine for clinical investigators

In PRN™, AURORA™ becomes the study design engine for clinicians and investigators. It identifies publishable studies from routine data, provides the method plan, and adds SSE-backed sample size reasoning to every research pathway.

Turns routine clinical capture into publication-ready design briefs

AURORA™ SSE

Sample size is a design question — not a post-hoc calculation.

The AURORA™ Sample Size Estimator integrates directly into the study discovery pipeline. By the time a study leaves AURORA™, its power logic is already documented and linked to the analytical plan.

Effect Inference

Infers effect pattern from study design

SSE reads the outcome structure and model type assigned by the Method Mapper and infers the expected effect pattern — without requiring the user to specify Cohen's d or equivalent inputs manually.

Model-Aware

Different logic for different models

Logistic regression, survival analysis, ANOVA, and linear models each receive model-appropriate sample size pathways. No single formula applied across incompatible designs.

Feasibility

Compares required n to available data

SSE checks the required sample size against the dataset size available in the current study context. Underpowered studies are flagged before analysis begins — not after results are in question.

Power Curve

Visual power-by-n reasoning

Each SSE output includes a power curve showing how power changes with sample size — giving investigators and supervisors a transparent basis for design decisions, not just a single required-n figure.

Portfolio Output Pack

Four deliverables. One structured research package.

Every AURORA™ run produces a research-ready package — structured so that a supervisor, peer reviewer, or funding committee can follow the design logic from input to intended output without asking the investigator to explain it.

Study plan

Titles & Objectives

Each discovered study receives a structured title, primary objective, secondary objectives, and a rationale drawn from the dataset's variable logic.

Method plan

Model & Assumption Map

Assigned model, assumption checks to run, sensitivity analysis plan, and the rule trace explaining why this method was selected for this study.

Power plan

SSE Summary & Curve

Required sample size, target power, effect assumption, model-specific formula, and power curve. Flagged if available data is insufficient.

Portfolio view

Dashboard & Handoff Pack

All studies in a structured executive-readable dashboard. Integrates directly with AMI™, AIMA™, and PRN™ — or exported for standalone use.

Deploy AURORA™

Standalone research engine. Or the intelligence core behind your entire Ayati stack.

Use AURORA™ directly to generate study portfolios from your datasets. Or deploy it as the reasoning layer behind AMI™, AIMA™, and PRN™ to convert raw data environments into disciplined, publication-ready analytical systems.

Request a Demo Narrative → View All Products

30-min demo agenda

→Dataset ingest and study discovery walkthrough

→Method Mapper rule trace — why this model, why now

→AURORA™ SSE: sample size and power curve output

→Portfolio dashboard and handoff pack review

→Live integration demo: AURORA™ inside AMI™ or PRN™

Recommended for principal investigators, research leads, and analytics directors.