Product Narrative

Upgrade Intelligence
for Sonata

AI-powered decision support that transforms complex upgrade planning into clear, visual, real-time insight — built natively into the Sonata product experience.

2views
Functional & Technical
1click
To capability drill-down
Real-time
Heatmap rendering
MCP-connected
Live Jira & Confluence data

Executive Summary

Upgrading the Sonata platform is one of the most operationally significant decisions a wealth management or superannuation firm can undertake. With hundreds of functional capabilities, a complex integration surface across Sonata Business Services (SBS), and release cycles spanning multiple minor versions, upgrade projects carry real risk — not just technically, but commercially.

The Upgrade Analyser is an AI-powered decision-support tool built natively into the Sonata product experience. By connecting Jira change data and Confluence release documentation through Model Context Protocol (MCP), the tool uses a large language model (LLM) to synthesise thousands of individual change records into interactive, real-time heatmap views — giving delivery, architecture, and business teams a shared picture of where the real upgrade impact lies.

Users select a 'from' and 'to' version directly within the application. The heatmaps update instantly, and any individual capability can be clicked to reveal the specific list of changes driving its concentration score.

Core Value Proposition

Select a 'from' version and a 'to' version. The Upgrade Analyser instantly surfaces where change is concentrated — by functional capability and by technical integration — with full drill-down to the individual changes behind every heatmap cell.

Upgrade Complexity at Scale

Sonata platform upgrades span multiple release versions, each containing a dense mix of functional enhancements, regulatory changes, SBS schema updates, and breaking technical changes. Organisations face three compounding challenges when planning an upgrade:

Volume of change data
Hundreds of Jira release notes and Confluence articles span every upgrade, making manual triage impractical and error-prone.
Cross-version accumulation
Clients upgrading across multiple minor versions must consider the compounded impact of all intervening changes simultaneously.
Dual impact surface
Functional changes affect business operations and user journeys; SBS schema changes affect integration consumers — both must be assessed independently and together.

A Native, Interactive Upgrade Intelligence Experience

The Upgrade Analyser is delivered as a native capability within the Sonata product — not a separate tool or offline report. Users interact with it directly within the application, with all heatmap views rendered in real time via a rich HTML interface.

1
Data Ingestion via MCP
Jira release tickets and Confluence release notes are continuously loaded into the LLM context through Model Context Protocol (MCP). No manual data preparation required.
2
Version Selection
The user selects a 'from' version and a 'to' version within the application. All available Sonata release versions are presented as selectable options.
3
Real-Time AI Analysis
The LLM analyses the full change set in range — aggregating scores by capability, applying weighted impact scoring (Low = 1, Medium = 5, High = 10), and classifying by type and integration risk.
4
Live Heatmap Rendering
Both heatmap views update instantly in the browser. No page reload or export step required. The selected version range is reflected immediately.
5
Capability Drill-Down
Clicking any capability cell opens a detailed change list — every Jira item, impact classification, change level, and release note summary — connecting the heatmap directly to the change evidence.

Three Core Interaction Moments

The experience takes a user from 'which versions am I upgrading?' to 'exactly which changes do I need to address?' in three connected steps.

MOMENT 01
Version Selection
Two version pickers — 'Upgrade From' and 'Upgrade To' — populated from the product release history. On selection, both heatmap views populate immediately. No further input required.
MOMENT 02
Heatmap Exploration
Functional and Technical heatmaps render side-by-side or as tabs. Cell colour intensity is proportional to accumulated change score. Capabilities with no change are visually distinct.
MOMENT 03
Capability Drill-Down
Click any cell to open a change detail panel — Jira keys, release versions, level badges, classification tags, LLM-generated summaries, and for SBS: schema change descriptions with request/response impact.
Design Principle

The version selection is the only input required. Everything else — data retrieval, AI analysis, scoring, and rendering — happens automatically and in real time. The user arrives at insight without configuration or preparation overhead.

Two Views, One Upgrade Picture

The Upgrade Analyser presents change concentration across two complementary views, each designed for a different audience and decision type.

1
Functional Change Heatmap (SCM)

Presents change concentration across the full Sonata capability model, organised by business domain and stakeholder type. Gives any member of the delivery or business team an immediate read of functional risk across the upgrade scope.

Digital — Adviser, Customer, Employer, Employee Non-Digital — Paper, Cashflow, Asset Register, Revenue Adviser Back Office — Pension, ISA, Drawdown, Reporting
Scoring Model

Each change is scored: Low = 1 pt, Medium = 5 pts, High = 10 pts. Scores accumulate per capability across all versions in range. A capability with ten Medium changes scores higher than one with a single High — preventing outliers from distorting the picture.

2
Technical Change Heatmap (SBS)

Focused on the Sonata Business Services (SBS) integration layer — the API surface connecting Sonata to third-party systems, client portals, and middleware. The primary decision-support view for integration and architecture teams. Every cell displays the most severe breaking-change classification present for that operation.

Code Classification Meaning for integration consumers
AB Always Breaking Mandatory schema changes that break all existing consumers. Immediate remediation required before go-live.
SBC Sometimes Breaking for Client Changes to optional elements that may break consumers currently using them. Assessment required.
SBL Sometimes Breaking Logically Response enhancements with new optional elements. May affect parsing logic or downstream consumers. Review recommended.
NBT Non-Breaking Technically Additive-only changes (new optional request elements) that do not affect existing consumers.

What the Upgrade Analyser Delivers

Native application experience
Real-time heatmap rendering
Capability drill-down
Version-to-version scoping
Dual heatmap views
Weighted impact scoring
Breaking change classification
Live data via MCP
Change type classification
LLM-generated summaries
Request / response impact detail
Cross-version accumulation

Who It Is For

Delivery Managers & Programme Leads
Scope upgrade workstreams visually in real time, identify high-effort capability areas, and share live heatmap views with stakeholders directly from the application.
Solution Architects
Assess functional architecture impact, identify which business domains require design review, and use drill-down to review the specific changes behind each capability score.
Integration Teams
Use the Tech Heatmap and drill-down to identify every breaking SBS change in the upgrade window, understand the exact schema change description, and determine test and remediation scope.
Business Analysts
Explore what product capabilities are changing by domain and stakeholder type, then drill into individual capabilities to understand specific operational impact.
Bravura Consulting / PS
Deliver fast, evidence-based upgrade impact assessments directly within the product during client engagements — without offline preparation.
Client Operations Teams
Self-serve an at-a-glance view of whether an upgrade cycle affects the capabilities their organisation actively uses, and click through to understand the detail.

The Difference It Makes

Without Upgrade Analyser
With Upgrade Analyser
Manual triage of hundreds of release notes across multiple versions
Real-time, AI-synthesised heatmap view across all versions in scope — generated in seconds
Risk of missing compounded cross-version changes
Accumulative scoring across every version in the upgrade window, visualised as a single heatmap
No clear view of where upgrade effort should concentrate
Colour-coded heatmap gives an immediate visual signal of the highest-concentration capabilities
Evidence for change impact buried in Jira and Confluence
One click on any capability cell surfaces the full list of impacting changes with context
Days of analyst effort per upgrade assessment
Self-service upgrade impact view generated in minutes from live MCP-connected data
Difficult to align functional and integration teams on scope
Shared dual heatmap — functional and technical — accessible to all teams within the same native application screen

Future Direction

The initial release delivers the core version-scoped heatmap intelligence with real-time rendering and capability drill-down. The following enhancements are on the product roadmap:

🎯
Client-aware filtering
Dynamically filter the heatmap to show only capabilities and SBS operations relevant to a specific client's configured implementation.
💬
Natural language query
Ask questions such as 'what changes affect drawdown in this upgrade?' and receive AI-generated narrative summaries alongside the visual heatmap.
📄
Automated upgrade readiness report
Generate a structured, exportable upgrade assessment document directly from the heatmap and drill-down data.
🧪
Test scope generation
Automatically propose a regression test scope mapped to the highest-concentration capabilities and breaking SBS changes.
📈
Change trend analytics
Compare heatmaps across multiple upgrade windows to identify capabilities undergoing sustained high-change cycles.
🔗
Jira integration for delivery planning
Push heatmap outputs into Jira as pre-populated epics or stories, with capability context drawn from the AI analysis.