Building Core Infrastructure for a VC-Funded AI Email Marketing Platform

Saasquatch implemented a custom Shopify app for a VC funded AI email marketing platform operating at the intersection of ecommerce, customer intelligence, and automation. The platform was designed to centralise customer data from multiple sources and apply machine learning driven segmentation to inform how brands engage with their audiences across email and related channels. From the outset, the work focused on building a technically sound foundation that could support complex data models, high volume processing, and future expansion beyond a single commerce platform.

The product was backed by venture capital investment and supported by New Zealand machine learning and artificial intelligence grant funding, reinforcing a strong emphasis on research led development and long term scalability. This funding context shaped early architectural decisions, particularly around data ownership, privacy compliance, and extensibility, ensuring the platform could evolve into a broader analytics and activation layer as new integrations and capabilities were introduced.

Business & Technical Challenges

Building an early stage platform alongside an evolving product vision introduced uncertainty at both business and technical levels. Requirements shifted as funding milestones, research inputs, and stakeholder feedback came together. The platform needed to demonstrate real analytical capability early on, while leaving room for significant expansion. Decisions made at this stage carried long term implications for performance, data integrity, and maintainability.

A key challenge involved working within Shopify as an initial anchor platform while avoiding architectural lock in. The system had to extract meaningful customer and order data without assuming Shopify would remain the sole source. This required careful separation between integration logic and core data processing, ensuring future connections to other commerce and marketing systems could be added without reworking foundational components.

Handling customer data at scale raised early questions around storage strategy, transformation pipelines, and query performance. Full historical records needed to be retained to support changing analytical models over time. At the same time, the platform had to remain responsive during initial syncs and ongoing updates, even as large volumes of records were processed and cached concurrently.

Privacy, consent, and data removal requirements added further complexity. The platform needed to comply with Shopify policies, GDPR expectations, and internal data governance standards from the outset. This affected how customer data was stored, how deletion requests were handled, and how integrations could be safely removed without leaving residual data behind in the system.

From a product perspective, the platform aimed to surface insights that felt meaningful without overwhelming early users. Many analytical features depended on future data sources, automations, or machine learning models that were not yet in place. The challenge lay in presenting credible proof of concept functionality while designing interfaces and data structures that would support far more advanced behaviour later.

Integrations introduced their own operational risks. Syncing data on a scheduled basis rather than relying solely on webhooks was necessary to ensure completeness, but increased the likelihood of partial failures, rate limits, and inconsistent states. The system needed to remain usable while background processes ran, retries occurred, and data volumes fluctuated across different customer accounts.

User management and billing presented additional constraints uncommon in simpler Shopify apps. The platform supported multiple organisations, multiple users per organisation, and fine grained access controls. These requirements influenced authentication design and permissions handling, while Stripe based billing needed to coexist with Shopify app review rules and external charge approvals.

Finally, the platform was developed in a funding environment that placed strong emphasis on technical credibility. Venture backing and New Zealand AI grant support brought expectations around robustness, auditability, and future research integration. This required disciplined engineering practices early on, including testing, deployment automation, and documentation, even while operating within tight timelines and an incomplete long term product picture.

Why Saasquatch Was the Right Fit

Saasquatch was selected due to its experience delivering complex Shopify applications that extend well beyond standard ecommerce workflows. The team brought a strong understanding of data intensive systems, serverless architecture, and long lived platform development, which aligned closely with the client’s technical and research driven direction. Prior work with analytics heavy products, third party integrations, and privacy sensitive data informed early architectural decisions. Saasquatch also operated comfortably within an environment shaped by venture funding and public innovation grants, where documentation, robustness, and forward planning carried equal weight alongside delivery pace.

Solutions We Delivered

Creating an Early Proof of Concept for Venture Capital Funding

Stage one focused on delivering a working proof of concept that could clearly demonstrate the platform’s underlying approach to customer data and insight generation. Saasquatch built a custom Shopify app that grouped customers using existing tag structures, then processed order data to surface aggregated sales metrics. Summary tables and visualisations allowed stakeholders to explore performance by segment across selectable date ranges, supported by test data where live data was limited.

The application was intentionally lightweight on presentation while being rigorous in how data was retrieved, processed, and stored. A serverless backend handled authentication, Shopify API access, and webhook processing, with early consideration given to caching and scalability. The app was structured to reflect future expansion rather than a disposable prototype.

This early implementation gave investors and grant reviewers a tangible view of how the platform would operate in practice. It demonstrated technical feasibility, analytical depth, and a clear path toward broader integrations, supporting successful venture capital fundraising and the award of New Zealand research and AI development grants.

Creating Scalable Infrastructure and Data Ingestion with AWS

The next phase centred on building infrastructure that could scale alongside both data volume and product scope. Saasquatch implemented a serverless AWS architecture to manage authentication, data ingestion, processing, and analytics workloads. The system was designed to remain independent of any single commerce platform, allowing multiple integrations to feed into a shared data model without introducing tight coupling or operational risk.

Data ingestion focused on capturing complete historical records and maintaining an accurate, queryable dataset as new information was added. Raw data was stored alongside processed outputs, supporting changes to analytical logic as the platform matured. This approach reduced the need for reprocessing and supported more complex insights over time.

Performance and reliability were key considerations throughout. The infrastructure was built to handle large sync operations while keeping the application usable, with attention given to caching strategies, concurrency, and fault tolerance. This foundation supported ongoing development of analytics, segmentation, and automation features.

Machine Learning for Customer Segmentation with Typology

This stage of the platform also introduced machine learning driven customer segmentation through a typology model designed to classify customers based on behavioural and contextual signals. Saasquatch integrated an external psychographics and typology API, allowing customer attributes such as vocation and engagement patterns to be translated into consistent segment identifiers. These identifiers were stored directly against customer records and became a core input for analytics, insights, and downstream marketing activity.

The segmentation model was designed to operate across multiple data sources and remain stable as integrations were added or removed. Customer profiles combined transactional history, survey responses, and behavioural events, creating a unified view that could support increasingly sophisticated analysis. Early implementations focused on establishing reliable data flows and repeatable classification rather than fully automated optimisation.

By embedding typology into the platform’s data model from the outset, segmentation became a foundational layer rather than a reporting feature. This enabled later work across insights, recommendations, and campaign targeting to build on a consistent understanding of customer behaviour and intent.

Designing and Developing a Comprehensive Email Marketing Platform and Insights Dashboard

Saasquatch designed and developed a multi section dashboard that surfaced customer insights, revenue trends, campaign performance, and behavioural activity in a single interface. The structure supported organisations with multiple users and data sources, while remaining approachable for teams reviewing performance day to day.

Insights were generated from a combination of transactional data, engagement events, and typology based segmentation. Users could explore revenue attribution, customer activity, email performance, and survey responses, with the ability to filter and export data as needed. Reports were designed to remain accurate as attribution models and underlying logic evolved.

Alongside analytics, the platform supported campaign creation, automation workflows, and email template management. These features were closely tied to the insights layer, allowing segmentation and behavioural data to inform how campaigns were structured and evaluated. The result was a unified system that connected data, decision making, and execution.

Integrating with Shopify and Future Integrations

Shopify was implemented as the initial integration, providing access to customer, order, and behavioural data that formed the foundation of the platform. Saasquatch built the integration to align with Shopify’s app and privacy requirements while maintaining clear boundaries between integration logic and core platform services. This ensured Shopify data could be ingested, transformed, and removed without affecting unrelated parts of the system.

From the outset, the integration model was designed to support additional platforms. Data sources were assigned unique identifiers, allowing multiple systems to coexist within a single organisation account. This made it possible to aggregate activity across channels while preserving traceability back to each source.

The same approach allowed for planned integrations with point of sale, marketing, and payment platforms. By avoiding assumptions tied to any one provider, the platform could evolve toward a broader ecosystem of data sources as product scope and customer needs expanded.

The Results

The work delivered a functioning platform that moved beyond concept documents and into a working product. Early stakeholders were able to interact with real data, explore customer segmentation, and assess how insights would be generated and acted on within a live environment. This shifted conversations from theoretical capability to practical implementation.

The stage one proof of concept supported successful venture capital fundraising and unlocked further investment through New Zealand research and artificial intelligence grants. With core infrastructure, data models, and integration patterns in place, the platform was positioned for continued development across analytics, automation, and multi-channel marketing without revisiting foundational decisions.