Overview
ParentSense is a digital parenting platform designed to help parents navigate pregnancy, baby care, early childhood development, feeding, sleep, health, and everyday parenting routines through trusted digital guidance. The platform provides personalised support, developmental insights, expert-backed content, and AI-assisted experiences aimed at making reliable parenting information more accessible at scale.
Founded by parenting expert and occupational therapist Meg Faure, ParentSense focuses on helping parents cut through overwhelming parenting information by delivering practical, science-backed guidance tailored to each child’s developmental stage and the parent’s evolving journey.
As the platform continued to expand its digital ecosystem and product vision, ParentSense required a modern cloud-native architecture capable of supporting intelligent AI-driven experiences, secure user management, scalable orchestration, governed operational workflows, and future AI innovation. The business needed an architecture that could support rapid growth while improving scalability, observability, maintainability, and long-term platform flexibility.
The Challenge
As ParentSense evolved, the platform needed to move beyond traditional application infrastructure and establish a scalable foundation capable of supporting intelligent parenting assistance and AI-powered user experiences.
The existing environment relied heavily on tightly coupled systems, shared database access patterns, and serverless dependencies that limited flexibility, scalability, portability, and long-term maintainability. At the same time, the business wanted to introduce more advanced conversational AI capabilities powered by retrieval-augmented generation (RAG), contextual AI workflows, and governed knowledge management.
This created several key architectural challenges:
Supporting intelligent conversational AI experiences at scale
Introducing structured knowledge ingestion and retrieval workflows
Modernising authentication and user security
Improving orchestration and service communication
Enabling AI governance and evaluation workflows
Maintaining operational visibility and monitoring
Avoiding long-term infrastructure lock-in
ParentSense therefore required a secure, scalable, and modular platform architecture capable of supporting rapid innovation while remaining flexible enough to evolve alongside future AI capabilities and growing user demand.
The Solution
To support ParentSense’s long-term vision, Tati Software designed and architected a scalable AWS-based microservices platform built around containerised services, governed APIs, and intelligent AI orchestration.
The solution introduced a layered architecture that separated the platform into dedicated domains for:
Authentication and identity management
Business orchestration and routing
AI processing and retrieval
Knowledge ingestion and governance
Monitoring and operational observability
This separation created stronger service boundaries, improved maintainability, and established a more secure and extensible foundation for future platform growth.
Modern Identity & Security Architecture
One of the platform’s most important transformations was the introduction of a dedicated Account Service responsible for the full user identity lifecycle.
Tati Software implemented a modern authentication and access-control architecture that centralised:
User authentication
Session management
Role-based access control
Credential management
Administrative identity operations
To strengthen data protection, Tati Software also implemented field-level encryption (FLE) for sensitive user information such as emails, phone numbers, and personally identifiable data. This ensured sensitive information remained protected both at infrastructure and application level.
The platform also introduced a staged migration strategy that allowed legacy and modernised authentication systems to coexist during rollout, ensuring existing users could transition safely into the new architecture with minimal disruption.
Intelligent AI & Retrieval Architecture
To support intelligent parenting assistance, Tati Software developed a dedicated AI Service responsible for:
Retrieval-Augmented Generation (RAG)
AI orchestration
Contextual response generation
Knowledge retrieval
AI evaluation workflows
Prompt orchestration and governance
Using LangChain and LangGraph, Tati Software designed controlled AI workflows capable of delivering grounded and context-aware parenting guidance rather than generic chatbot interactions.
Instead of sending prompts directly to language models, the platform implemented a structured retrieval pipeline that:
Determined user intent
Identified relevant knowledge collections
Retrieved relevant information from vector-enabled databases
Ranked and filtered contextual information
Constructed grounded prompts
Applied safety and response controls before delivery
This allowed ParentSense to deliver more accurate, explainable, and personalised AI-generated parenting guidance.
The AI platform also supported contextual enrichment using parenting-specific signals such as developmental stages, previous interactions, behavioural state, and user-specific context supplied through governed APIs. This enabled more adaptive and personalised parenting experiences across the platform.
Knowledge Management & AI Governance
A major focus of the solution was establishing governed AI knowledge workflows rather than unmanaged vector storage.
Tati Software designed and implemented a structured knowledge ingestion and management framework that allowed ParentSense administrators to:
Upload and categorise knowledge content
Manage document ingestion workflows
Reprocess and re-index knowledge assets
Manage embedding pipelines
Maintain version control and archival workflows
Govern AI retrieval quality and indexing visibility
The platform also introduced formal AI governance and evaluation tooling that enabled administrators to manage:
Prompt versions
Agent configurations
Model selection
Retrieval parameters
Evaluation datasets
Benchmark testing workflows
Fallback logic and permissions
This governance layer ensured AI quality could be continuously monitored, evaluated, and improved before and after production deployment.
Technical Architecture
The solution was deployed using a secure AWS cloud-native environment built around ECS container workloads operating within a Virtual Private Cloud (VPC).
The architecture included:
Amazon ECS
PostgreSQL databases
Redis caching
Amazon SQS
Application Load Balancers
AWS WAF
Amazon Route 53
Amazon S3
IAM & Secrets Manager
pgvector-enabled vector databases for AI retrieval workloads
Tati Software also designed the platform with future scalability and operational resilience in mind, including support for:
Autoscaling
Multi-AZ deployment
Disaster recovery planning
Distributed tracing
Operational observability
Structured monitoring and alerting
Business Impact
The ParentSense platform modernisation established a scalable and future-ready foundation for intelligent parenting experiences and long-term AI-driven product innovation.
Through this solution, Tati Software enabled ParentSense to:
Deliver more personalised AI-assisted parenting experiences
Improve platform scalability and operational flexibility
Strengthen security and governance
Introduce governed AI experimentation workflows
Improve maintainability and reduce platform coupling
Expand AI capabilities without major architectural redesigns
Establish stronger operational visibility and monitoring
By combining modern cloud-native architecture, intelligent AI orchestration, secure identity management, and governed AI knowledge workflows, Tati Software helped ParentSense build a scalable AI platform capable of supporting continued product growth and future innovation.
Metrics for Success
KPI 1: Faster Release of Platform Improvements
Baseline:
Before the mordenisation, ParentSense had tightly coupled workflows and shared service dependencies. Routine platform improvements often required broad regression review across unrelated areas, slowing release preparation.
Target:
Improve release readiness by separating key service domains and introducing more controlled deployment and validation practices.
Measured Result:
Average release preparation time for routine platform improvements reduced from approximately 4–5 working days to 2 working days after the platform was separated into clearer service domains for identity, orchestration, AI processing, knowledge management, and observability.
Measurement Method:
Comparison of pre- and post-modernisation release notes, regression checklists, deployment records, and change-impact assessments.
Business Impact:
ParentSense could deliver product improvements and AI-assisted parenting features faster, while reducing the operational risk associated with production releases.
KPI 2: Faster AI Parenting Content Updates
Baseline:
Before implementation, updates to AI parenting guidance content required more technical intervention because knowledge ingestion, retrieval controls, prompt alignment, and evaluation workflows were not fully governed.
Target:
Improve business agility by introducing governed AI knowledge operations for content ingestion, indexing, retrieval review, prompt control, and evaluation.
Measured Result:
Standard AI parenting content update cycles reduced from approximately 3 working days to 1 working day for routine updates after governed knowledge ingestion and retrieval workflows were introduced.
Measurement Method:
Comparison of pre-modernisation content update handling against post-modernisation knowledge ingestion records, retrieval configuration, prompt/version controls, and evaluation workflow evidence.
Business Impact:
ParentSense could update parenting guidance content faster and more safely, improving its ability to keep AI-assisted advice relevant, accurate, and aligned with business and expert-review requirements.
Outcomes
The ParentSense modernisation created a more secure, scalable, and governable AWS operating model for AI-assisted parenting experiences. The architecture reduced coupling, improved operational visibility, strengthened user-data protection, and enabled future AI expansion without requiring the platform to remain tied to tightly coupled legacy workflows.
Lessons Learned
Key lessons included the importance of separating identity, orchestration, AI processing, knowledge management, and observability early in the architecture. The engagement also reinforced the value of staged migration, field-level encryption, governed AI knowledge workflows, structured monitoring, and controlled release practices for customer-facing AI platforms.
Tati Software will continue applying these lessons to future customer environments by standardising service-boundary design, strengthening observability from the outset, defining governance controls early, and ensuring production workloads have clear operational ownership before scale increases. (edited)