Platform Roadmap · April 2026
ChainSys Smart
Data Platform
2026 Roadmap
& Q1 Achievements
Our 2026 platform strategy and Q1 delivery recap — covering planned releases, AI capabilities, data catalog, MDM intelligence, and analytical model creation.
Migration & Integration
Data Quality · EDM
MDM
Data Catalog
App Builder
AI Capabilities
3
Strategic Tracks
Q1 ✓
On Track
200+
Connectors
6.0.1
Platform Version
01
Planned Releases
High-level view of what is coming Q1 through Q4 across all solutions
02
AI · Data Exchange
AI features reducing time and cost in migration and integration
03
Data Catalog
Unstructured to structured conversion, lineage, PII, and governance
04
AI · MDM & Analytics
AI-driven master data management and analytical data model creation
AGENDA
What We'll Cover Today
Platform-wide 2026 strategy and Q1 delivery recap, structured around our core delivery pillars.
01
Strategic Overview
Three synchronized tracks — Solution Experience, AI Capabilities, and Platform Modernisation and how they compound value each quarter
02
Planned Releases — High-Level View
All solutions, all quarters at a glance — what is shipping and when across the 2026 roadmap
03
AI Capabilities · Data Exchange
AI features that directly reduce time in data migration and integration — with time and cost saving context
04
Data Catalog Capabilities
Unstructured to structured data conversion, automated cataloguing, lineage, PII governance, and business glossary
05
AI Capabilities · MDM
AI-driven master data management — survivorship, deduplication, golden records, and intelligent stewardship
06
AI · Analytical Data Model Creation
AI-driven dataset generation, OLAP model building, NLP-powered query, and autonomous dashboard creation
Agenda
STRATEGIC OVERVIEW
Three Synchronized Tracks
Our 2026 roadmap balances immediate customer value with long-term platform evolution through three parallel tracks.
1
Solution Experience
ENHANCEMENTS ACROSS YOUR SOLUTIONS
  • Migration & IntegrationAuto orchestration of migration flows, IntelliObject Creator, execution dashboards, smart process summaries, and reconciliation improvements — reducing setup time and increasing run visibility. Q3: Automated Data Replication adds schema-aware, near-zero-config sync as a third data movement method alongside migration and integration.
  • Enterprise Data ManagementAdvanced search, conditional filters, bulk governance improvements, dictionary matching, cleansing review, and request creation automation — accelerating EDM operations at scale.
  • Master Data ManagementSteward and business persona experiences, bulk deployment with versioning, AI-powered reconciliation insights, and cross-solution navigation to track data from source to golden record.
  • Unstructured to StructuredEnhanced OCR engine and form-based extraction with improved LLM orchestration — increasing accuracy of document-to-structured conversion. Bulk Data Object creation for profiling and improved business user involvement in reviewing extracted output.
  • Data Product-Centric UX Q3 · Platform-widePersona-aware views for Engineers, Stewards, and Business Users across Migration, MDM, Analytics, and App Builder — each role sees what is relevant without unnecessary complexity. In dataZense this becomes the Data Product Hub: a fundamentally new paradigm where data is published, owned, quality-scored, and discovered as a governed product.
  • Outcome VisibilityExecutive and audit views, solution health dashboards, and built-in adoption metrics — giving leadership a clear picture of value delivered across every solution.
2
AI Capabilities
INTELLIGENCE LAYER
  • Governed AI GatewayCentralised LLM control plane — all AI calls audited, rate-limited, and tenant-scoped across OpenAI, Anthropic, Azure, and self-hosted models.
  • Confidence-Scored RecommendationsEvery AI suggestion — mapping, tagging, survivorship — carries a confidence score so teams validate selectively, not exhaustively.
  • AI in Data ExchangeSmart Mapping, NL pipeline creation, intelligent load sequencing, anomaly detection, and error correction — directly reducing time and cost in migration and integration.
  • AI in MDM & Data QualitySurvivorship Assistant, AI Rule Builder, deduplication agents, and inbound mapping automation across the full master data lifecycle.
  • AI in Catalog & AnalyticsAuto-tagging, PII classification, NL metadata query, insight narration, and autonomous dashboard generation from catalogued data.
  • Learning SystemFeedback loops from every execution improve models over time — explainability workflows per decision available to stewards and auditors.
3
Platform Modernisation
ARCHITECTURE EVOLUTION
  • Smaller, Faster DeploymentsMigration to Quarkus delivers 60–70% smaller container images and millisecond startup times — reducing infrastructure footprint and deployment overhead across all services.
  • Auto-Scaling & Self-Managing InfrastructureKubernetes-orchestrated deployment with auto-scaling workers and auto scale-down of idle threads — the platform expands under load and contracts when quiet, without manual intervention.
  • Modern, Consistent UIProgressive migration to ReactJS across all new screens — faster rendering, a consistent component library, and a more responsive experience for every user.
  • Full Production VisibilityOpenTelemetry distributed tracing, structured logging, and metrics across all services — teams see exactly where issues originate before users are impacted.
  • Lower Cost at ScaleServerless execution for batch and non-critical workloads means pay-per-use scaling — infrastructure costs grow proportionally with usage, not ahead of it.
  • Continuous Security HardeningRegular dependency patching, vulnerability remediation, and security updates maintained through every quarter — no accumulated technical debt in the security posture.
All three tracks progress simultaneously every quarter — performance improvements from Track 3 amplify AI in Track 2, which surfaces insights through the improved experience in Track 1.
Strategic Overview · Three Tracks
01 · PLANNED RELEASES
2026 at a Glance — All Solutions
Progressive delivery across four quarters, building from a stable foundation to outcome-driven intelligence.
QuarterSolution ExperienceAI CapabilitiesPlatform Modernisation
Q1
Jan–Mar		 Smart Mapping Assistant · OAuth for Oracle BI · RBAC on Assessment Dashboard · AI Rule Builder · Kubernetes Enablement · ES Build Optimisation (57% faster) Unified AI framework · Metadata ingestion pipeline · AI-assisted mapping · Survivorship & Review Assistant Spring 6.2.14 patch · Node 22 · Angular 18 · Vulnerability remediation · OpenTelemetry pilot
Q2
Apr–Jun		 Smart Process Summary · Template Hub · IntelliObject Creator · ETL via natural language · Embedded Reconciliation insights · AutoFilter & Input Assistant · Unstructured: extended format support (PPTX+) · table extraction · confidence scoring · Bulk Data Object profiling Confidence-scored field mapping · Auto-tagging · Intelligent load sequencing · Anomaly detection · Chatbot-driven transforms Quarkus rollout (high-load services) · ReactJS for new screens
Q3
Jul–Sep		 Inline error insights · Steward & Business persona UX (all solutions) · Cross-solution navigation · Bulk Deployment with Versioning · AI Reconciliation Insights · Migration Flow Runbook · Automated Data Replication · Data Product Hub (dataZense) Error correction suggestions · Executive summaries per run · AI smart pilot with explainability · Inbound Mapping Automation (EDM) · Master Data AI search (MDM) · Milvus Vector Loader Iceberg + Presto standardisation · Auto-scaling workers · Serverless for non-critical workloads · Advanced analytics infrastructure · ReactJS rollout
Q4
Oct–Dec		 Executive & audit views · Solution health dashboards · Schema Drift Monitor · Reusable blueprints · Database & File Polling · Endpoints: Kinesis, Epicor, Zoho, Netezza Explainability workflows · Continuous model optimisation · Feedback loops · Overtime Job Predictor · Potential Error Detector · Request Creation Automation Architecture hardening complete · Serverless for critical workloads · Full OpenTelemetry rollout · Cost optimisation
All three tracks — Solution Experience, AI Capabilities, and Platform Modernisation — run in parallel every quarter. Each compounds the value of the others.
01 · Planned Releases
02 · AI CAPABILITIES
AI for Data Exchange — Faster Migration & Integration
AI capabilities that directly cut the time and cost of moving data — from field mapping to pipeline orchestration to error correction.
~70%
Mapping time reduction
AI chat-based field mapping replaces manual source-to-target mapping cycles
~60%
Error triage time saved
AI Load Error Correction Suggestor identifies and suggests fixes before manual intervention
~80%
Pipeline setup time
ETL pipeline creation via natural language removes manual configuration overhead
Real-time
Anomaly detection
Proactive flagging of data drift and anomalies before downstream consumption
Q1
Delivered — In Production
  • Smart Mapping AssistantAI chat-based field mapping — reduces complex migration mapping effort by up to 70%. Live.
  • OAuth in Oracle BI ConnectionSecure, token-based auth — eliminates credential sharing and manual auth setup overhead.
  • Databricks Service Principal AuthEnterprise-grade authentication to Databricks — no shared credentials, immediate connectivity.
  • Unified AI FrameworkSingle governed AI layer across all solutions. All LLM calls audited, rate-limited, and tenant-scoped.
Q2
Coming Next Quarter
  • ETL Pipeline Creation via Natural LanguageDescribe the pipeline in plain English — the platform creates the ETL objects automatically.
  • Confidence-Scored Field MappingEvery AI mapping recommendation comes with a confidence score — teams validate only low-confidence fields.
  • Intelligent Load SequencingAI determines optimal execution order for dependent load jobs — no manual dependency mapping.
  • AI Load Error Correction SuggestorAutomatically diagnoses load failures and suggests targeted fixes — cuts error resolution cycles.
  • Anomaly Detection in Data FlowsReal-time monitoring of data volumes, patterns, and quality thresholds during pipeline execution.
  • Chatbot-Driven Prevalidation & TransformsAuto-generate validation checks and transformation logic via conversational interface.
Q3
Planned
  • Error Correction Suggestions with ContextFull execution context surfaced alongside error — no log-diving required to diagnose pipeline failures.
  • Smart Process SummaryAI-generated plain-language summary of every ETL and migration run — immediate visibility for business stakeholders.
  • Execution Completion EstimationPredicted ETAs for Loader and Dataflow jobs — plan resource availability around job completion.
  • Automated Data ReplicationSchema-aware, continuous replication — database changes at the source are captured and applied at the target automatically, with near-zero ETL configuration. Complements extraction-based CDC for transformation-heavy pipelines; replication is the right choice when structure is preserved and speed of sync matters most.
Q4
Roadmap
  • Overtime Job PredictorPredicts jobs likely to exceed SLA windows based on historical patterns — act before not after.
  • Potential Error DetectorFlags likely failures before execution — uncompiled dependencies, broken connections, schema mismatches.
  • Database & File PollingEvent-driven ingestion — continuously monitors sources and triggers pipelines on new or changed data.
  • Amazon Kinesis EndpointStreaming data ingestion connector — real-time data lake population from event streams.
02 · AI Capabilities · Data Exchange
03 · DATA CATALOG & DATA PRODUCT HUB
Data Catalog & Data Product Hub
Catalog engine enhancements and Unstructured improvements ship Q2. Q3 introduces the Data Product Hub — a fundamental reimagining of dataZense where data is no longer catalogued as a technical asset, but published, owned, and consumed as a governed product.
Q2
Catalog Engine — Enhancements & Unstructured Improvements
Profiling Engine
  • Structured ProfilerMetadata engine + sampling engine for all RDBMS, cloud, and big data sources
  • Unstructured Profiler — OCR EngineReads text content, tables, form-based data, and images from documents — converts to structured output
  • Form-Based EngineIdentifies and extracts structured fields from semi-structured forms and templates
  • Extended Format Support Q2Extends extraction to additional formats beyond existing DOCX and XLSX support — adding PPTX, HTML, and further document types with no additional configuration required
  • Table Extraction with Structure Preservation Q2Tables within documents become governed rows and columns — not flat text — enabling direct downstream use
  • Extraction Confidence Scoring Q2Each extracted field carries an AI confidence score — low-confidence items automatically flagged for steward review
  • Bulk Data Object Profiling Q2Batch-profile multiple data objects in a single operation — significantly reduces cataloguing time at scale
Catalog Engine
  • Apache Solr Search EngineFull-text search across all catalogued metadata and indexed source data — unified discovery
  • Automated Tagging & Glossary GenerationAI auto-tags data objects and generates business glossary entries — no manual labelling
  • Data Lineage AutomationEnd-to-end lineage from source endpoint through transformation to target — forward and backward propagation
  • Natural Language Metadata Query"How many records loaded from Oracle in the last 3 days?" — answered instantly, no SQL required
Governance & Protection
  • PII Tag EngineAutomated identification, classification, and rectification of personally identifiable information
  • Data Protection EngineRow-level and column-level security — role-based access for data engineers, stewards, and analysts
  • Data CitizenshipBusiness users can view, comment, and approve catalogued data — governance with business ownership
  • Schema Drift Monitor Q4Detects and alerts on schema changes in source systems before pipelines are impacted
Q3
Data Product Hub — New Capability  ·  dataZense reimagined
dataZense is not being updated — it is being reframed. Data is no longer a catalogued asset. It becomes a governed product: published with an owner, a quality score, and an SLA — discoverable and consumable by any persona across the platform, and surfaced contextually in every ChainSys solution via Smartlets.
Publishing & Ownership
  • Data Product PublishingEngineers publish datasets as governed products — metadata, owner, SLA, quality score, and access policy defined at publish time
  • Ownership ModelEvery product has a named owner accountable for freshness, quality, and access — not just a catalog entry with no responsibility attached
  • Product VersioningProducts are versioned — consumers can pin to a version or auto-follow latest. Breaking changes are signalled, not silent
  • Automated Quality ScoreQuality score calculated on every pipeline run and published on the product — consumers always see current health at a glance
Persona-Aware Discovery
  • Role-Based Product ViewsEngineer sees schema + lineage · Steward sees quality + ownership · Business user sees description + usage — same product, right view per role
  • NL Product Discovery"Find approved financial datasets with PII removed and freshness within 24 hours" — natural language search across the product catalogue
  • Browsable Product CatalogueDedicated product catalogue with categories, tags, ownership, quality indicators, and usage metrics — purpose-built for consumers, not data engineers
  • Subscription & NotificationSubscribe to any product — notified on quality changes, schema updates, SLA breaches, or ownership changes
Smartlets & Cross-Solution
  • SmartletsContextual data product cards surfaced within Migration, MDM, Analytics, and App Builder — see the product linked to any job, hub, or dataset without leaving the solution
  • Cross-Solution Product LinkingMigration jobs, master data hubs, and analytical datasets are linked to their source products — full traceability from raw source to published product
  • Contextual Quality AlertsWhen a consumed product drops below its quality threshold, the alert surfaces inside every solution consuming it — not just in the catalog
  • Product Lineage ViewTrace from raw source data through all transformations and pipelines to the final published product — one view, end to end
AI throughout: Auto-tagging, PII classification, extraction confidence scoring, NL product discovery, and quality narration are all routed through the governed AI Gateway — audited, rate-limited, and tenant-scoped.
FLOW
From Raw Data to Governed Data Product
📥
Ingest
Documents, databases, and cloud sources ingested via 200+ connectors across all formats
🔍
Profile
OCR, form parser, structured profiler, and LLM extraction — with confidence scoring on every field
🏷
Catalogue
AI auto-tags, classifies PII, generates glossary, registers in Apache Solr — fully indexed and searchable
🏛
Govern
Lineage recorded, access controls applied, steward workflows triggered, audit trail maintained
📦
Publish Q3
Published as a governed data product — owned, versioned, quality-scored, and discoverable by every persona
03 · Data Catalog & Data Product Hub
04 · AI CAPABILITIES · MDM
AI-Driven Master Data Management
Intelligence applied across the full MDM lifecycle — from inbound data matching and survivorship to golden record management and outbound distribution.
Q1
Delivered — In Production
  • Survivorship AssistantAI guides survivorship decisions — recommends which version of a record becomes the golden record, with justification
  • Review AssistantAI-assisted record review for MDM stewards — surfaces the highest-risk records first, reduces manual review volume
  • AI Rule BuilderBulk governance rule creation via AI for EDM and Data Quality — rules generated from natural language descriptions
Q2
Coming Next Quarter
  • AI Deduplication AgentsML-powered entity matching across sources — identifies duplicates with confidence scores before merge
  • AI Enrichment AgentsAutomatically enriches master records with inferred attributes from related data — reduces data gaps
  • Anomaly Detection on Master DataFlags unexpected changes to master records — value drift, unusual merge patterns, source conflicts
Q3
Planned
  • Master Data View — AI-Powered SearchNatural language search across the master data hub — find any golden record using plain English queries
  • Inbound Mapping Automation (EDM)AI automatically maps inbound source fields to master hub attributes — removes manual schema mapping
  • AI-Powered Reconciliation InsightsIntelligent summary of reconciliation results — matches, mismatches, anomalies, and trends surfaced automatically
  • Policy Suggestion AgentsAI recommends governance policy updates based on data patterns and steward approval history
Q4
Roadmap
  • Request Creation Automation (EDM/DQ)AI generates data quality requests and EDM change requests automatically from observed data issues
  • Explainability WorkflowsEvery AI decision — merge, dedup, survivorship — has a full explainability trail available to stewards
  • Continuous Model OptimisationFeedback from steward approvals and corrections is fed back into the AI models — they improve with every cycle
The MDM intelligence layer sits on top of dataZen — which provides MDM, DQM, and Data Governance capabilities built on dataZap for all inbound and outbound connectivity. Golden records are stored in the Master Data Hub and distributed to ERP, CRM, analytics, and BI consumers.
04 · AI Capabilities · MDM
05 · AI · ANALYTICAL DATA MODEL CREATION
AI-Driven Analytical Data Model Creation
AI capabilities across the analytics layer — from dataset construction and OLAP model building to NLP-powered querying and autonomous dashboard generation.
Current Capabilities — dataZense Analytics
  • OLAP Cube EngineAggregation, slice-and-dice, and drill-down across multi-dimensional datasets
  • ML / Learning Engine (R + Python)Supervised learning (regression), unsupervised (clustering), reinforcement learning, and NLP — R and scikit-learn, TensorFlow
  • NLP EnginePython · NLTK · spaCy — natural language processing over data for text classification, entity extraction, and summarisation
  • AI Deterministic AgentsInsight narration · Anomaly explanation · Report summarisation · Forecast commentary — all AI-generated automatically per run
  • Dataset EngineQuery and cache layer for constructing governed analytical datasets from any connected source
2026 Roadmap Additions
  • Natural Language Metadata Query (Q2)Ask questions about data volumes, pipeline runs, and dataset attributes in plain English — no SQL or MDX required
  • Embedding Transformations (Q2)Introduces embedding, transpose, and HTML-to-text transformations in Dataflow — prepares data for ML pipelines
  • Milvus Vector Store Loader (Q3)Loads data into Milvus for vector storage and semantic similarity search — powers embedding-based analytical models
  • Executive Summaries per Solution Run (Q3)AI generates plain-language executive summaries after every analytics run — no manual report authoring
  • AI Smart Pilot with Explainability (Q3)AI recommendations come with reasoning — analysts can see why a model or insight was generated
  • Dashboard Generator — Autonomous (Q3)App Builder generates dashboards autonomously from data definitions — business users define outcomes, not layouts
  • Continuous Model Optimisation (Q4)Feedback loops from every analytical run improve model accuracy over time without manual retraining
All AI analytical capabilities are governed through the AI Gateway — model selection, prompt management, audit logging, and rate limits per tenant. Supports OpenAI, Anthropic, Azure OpenAI, and self-hosted models.
FLOW
From Raw Data to Analytical Model — AI Assisted
🗄
Source & Connect
Data ingested from 200+ endpoints — RDBMS, cloud apps, S3, Kafka, and more via dataZap
🧱
Dataset Construction
AI-assisted dataset engine builds governed analytical datasets with caching and access controls
🤖
Model & Analyse
OLAP cubes, ML models, NLP, and embedding pipelines run — agents narrate and explain outputs
📊
Visualise & Deliver
Autonomous dashboard generation, scheduled reports, real-time views — delivered to business users
05 · AI · Analytical Data Model Creation
06 · PLATFORM IMPACT
What This Roadmap Delivers
Across every capability area, the 2026 roadmap produces measurable reductions in time, effort, and risk — and measurable increases in data quality, visibility, and operational confidence.
↓
70%
Data mapping effort
AI Smart Mapping Assistant · Q1 live
↓
80%
Pipeline setup time
ETL via natural language · Q2 2026
↓
60%
Error triage cycles
AI Error Correction Suggestor · Q2–Q3
↑
95%+
Lineage automation
Catalog lineage engine · Q4 2026
↓
57%
Build time · in production
ES Build optimisation · Q1 delivered
30×
faster
Workflow approvals
10 min → 20 sec · Q1 delivered
CAPABILITY TIMELINE AT A GLANCE
Capability AreaQ2 2026Q4 2026
Data mapping and pipeline setup effort↓ 70% reduction via AI Smart Mapping↓ 80%+ with NL pipeline creation fully live
Unstructured document extraction effort↓ 50% via enhanced OCR + LLM profiling↓ 75%+ with full catalog integration
Pipeline anomaly detection↓ 80% faster — real-time flagging active↓ 95%+ — predictive detection before failure
Data lineage coverage↑ 80% automated across catalogued sources↑ 95%+ fully automated, end-to-end
Audit and compliance preparation↓ 70% time reduction↓ 90%+ — hours not weeks
Master data stewardship effort↓ 50% via AI survivorship + dedup agents↓ 75%+ with explainability workflows complete
06 · Platform Impact · 2026 Roadmap