Spiral Scout

Found required a mobile application to support rapid user growth and deliver personalized weight management experiences. The system needed to integrate health data, support continuous experimentation, and be ready for production within a two-month timeframe.

A cross-platform mobile application was built using React Native, allowing a single codebase to run on both iOS and Android. The application integrates with Apple Health and Google Fit, enabling ingestion of user health data to support personalized insights and tracking.

The development process was structured around short iteration cycles, enabling rapid feedback and continuous refinement. A/B testing capabilities were built into the application, allowing features and user flows to be evaluated in real time based on user behavior. Analytics instrumentation supports measurement of engagement and conversion across different variants.

CI/CD pipelines were implemented to automate builds, testing, and deployment. This allows frequent updates to be shipped without disrupting users and supports ongoing iteration after the initial release.

The architecture supports scalability and maintainability, ensuring the application can evolve alongside the product. Integration with third-party services such as analytics and messaging platforms enables communication and engagement workflows within the app.

What now exists is a production mobile application that supports cross-platform usage, continuous experimentation, and integration with health data sources. The system runs in production, supports ongoing feature iteration, and is owned by the client team, allowing independent development and scaling without reliance on external engineering support.

Appspace required a content delivery system capable of handling high request volumes while maintaining low latency and consistent performance. The existing system struggled with scaling, manual content workflows, and inefficient data access patterns.

A modular content delivery platform was built using Golang as the core execution layer. The system processes and distributes content through an event-driven architecture designed to handle high concurrency and complex query patterns. Components are decoupled, allowing independent scaling and controlled resource allocation.

Temporal was introduced to manage workflow automation for content scheduling and updates. Content delivery processes run as workflows, reducing manual intervention and ensuring consistent execution across distributed components. This enables reliable handling of asynchronous operations and real-time updates.

Kubernetes-based orchestration manages containerized services, allowing the platform to scale dynamically based on traffic and workload. This ensures stable performance under high load while maintaining operational flexibility.

Data access was optimized through indexing and query improvements, reducing latency and increasing throughput. The system supports high-volume filtering and retrieval operations with efficient resource usage.

What now exists is a production content delivery system that processes high traffic with predictable performance. The platform handles thousands of concurrent connections and high-frequency requests while maintaining low resource consumption. Content workflows execute automatically through Temporal, and the system is owned and operated by the client, with the ability to extend and scale without reliance on external engineering support.

Inturn operated a PHP monolithic application that became difficult to scale as the platform grew. Tight coupling between components slowed feature development, limited independent scaling, and made it difficult to manage complex, long-running processes.

The system was re-architected into a microservices-based platform using Golang. Core functionality was decomposed into independent services that can be deployed and scaled separately, allowing the platform to handle increased traffic and evolving product requirements without impacting the entire system.

Temporal was introduced as the orchestration layer to manage distributed workflows and stateful execution across services. Workflows coordinate service interactions, handle retries, and maintain consistency for long-running processes. This removed the need for custom state management and reduced failure points in inter-service communication.

The architecture supports asynchronous execution and fault tolerance. Services communicate through defined workflows rather than tightly coupled integrations, ensuring that failures can be retried or recovered without breaking the overall process.

A phased migration strategy was defined to transition from the monolith without disrupting existing operations. New services were introduced incrementally while maintaining system continuity.

A testing framework was established to validate microservice behavior, performance, and reliability across the distributed system. This ensured that the new architecture could operate under production load.

What now exists is a production microservices platform built on Golang with Temporal-managed workflows. The system supports independent service scaling, faster feature delivery, and reliable execution of distributed processes, with full ownership and ongoing development handled by the client team.

Wippy AI is a multi-agent automation framework designed to replace rigid, rule-based systems with adaptive, coordinated workflows. Traditional automation approaches relied on predefined scripts that failed under changing conditions, leading to bottlenecks, manual intervention, and limited scalability.

A distributed agent architecture was built where each agent is responsible for a specific function such as task execution, validation, orchestration, or error handling. Agents operate as part of a coordinated system, sharing context and dynamically redistributing work based on real-time conditions. This allows workflows to continue operating even when inputs, priorities, or system states change.

The system includes an internal communication layer that enables agents to exchange data, request assistance, and coordinate multi-step tasks. This replaces isolated automation processes with a connected execution model where tasks are decomposed, routed, and completed collaboratively.

An observability and workload management layer monitors execution, tracks performance, and adjusts task prioritization. Agents can reassign workloads and rebalance resources to maintain throughput under high demand. The architecture supports parallel execution across distributed infrastructure, enabling high-volume task processing without degradation.

Error handling and recovery are built into the execution model. Agents detect failures, retry operations, or reroute tasks without requiring manual intervention, ensuring continuity in long-running workflows.

What now exists is a production automation framework that executes tasks through adaptive, multi-agent coordination. The system handles high volumes of work with consistent accuracy, reduces manual intervention, and allows organizations to extend automation logic without rebuilding core infrastructure.

Sourcing Value processes supplier data across a distributed supply chain, where monthly Excel files arrive in inconsistent formats. Manual normalization, mapping, and validation created delays, errors, and heavy operational overhead before data could be used.

A production data processing system was built using a multi-agent architecture on Wippy. The system ingests raw Excel files, identifies structure and schema variations, and transforms inputs into normalized, structured datasets. Agents specialize in validation, mapping, and formatting, coordinating through defined workflows to produce consistent outputs.

The pipeline includes automated data validation and correction, with a human review step to verify results when needed. This maintains accuracy while removing the need for manual preprocessing. Data quality is enforced through system-level rules rather than ad hoc cleanup.

Custom API and EDI connectors integrate the pipeline directly into existing enterprise systems. This allows processed data to flow into downstream tools without additional handling, establishing a continuous and repeatable workflow.

The architecture supports scaling across increasing data volume and supplier variability. Each stage of processing is modular, allowing new data formats or validation rules to be introduced without reworking the system.

What now exists is a production system that converts inconsistent supplier inputs into structured, reliable datasets through agent-driven workflows. Processing that previously required over a month of manual effort now completes in minutes, with significantly reduced error rates. The system is owned by Sourcing Value and can be extended without dependency on external engineering support.

Project Fortress is a legal transaction management platform built on Salesforce, designed to handle complex deal workflows across M&A, real estate, and joint ventures. The system needed to process large volumes of unstructured legal documents while maintaining strict data integrity and compliance requirements.

A multi-agent architecture was built using Wippy and Temporal to automate document processing and workflow execution. Agents ingest contracts, emails, and supporting files, extract structured data, and map outputs directly into Salesforce. This creates a consistent, queryable data layer while preserving Salesforce as the system of record.

Workflows run as long-lived, auditable processes with permissioning, failure recovery, and human oversight. Agents coordinate tasks such as checklist generation, issue tracking, and status updates, reducing manual coordination across legal teams. Human-in-the-loop controls ensure that critical decisions remain reviewable and governed.

A secure integration layer enforces role-based access control and real-time validation. Sensitive data is handled through controlled interfaces, ensuring compliance with legal and client requirements. The system supports multiple AI models and uses semantic retrieval to operate on verified document content.

The platform includes a modular registry system with version-controlled components, enabling extension across different transaction types without modifying core workflows. Multi-tenant capabilities support multiple clients while maintaining strict data isolation.

What now exists is a production system where legal deal workflows are executed through agent-driven processes integrated directly into Salesforce. Document analysis, task orchestration, and data synchronization run continuously with auditability and control. The platform is owned by the client and can be extended without dependency on external engineering support.

Gorgias operates a customer support platform used by over 15,000 e-commerce brands. As their AI capabilities expanded, the system needed to support thousands of concurrent agents handling both reactive and proactive workflows such as support requests, order changes, and upsell scenarios.

A workflow system was built on Temporal Cloud to orchestrate multi-agent execution across long-running, asynchronous processes. Each agent runs as an independent workflow with durable state, enabling coordination between support, sales, and routing logic while maintaining full traceability.

The architecture supports distributed transactions across external APIs with retry logic and failure recovery. Saga patterns and asynchronous step handling ensure agents can pause, resume, and continue interactions across channels such as email, SMS, and chat without losing context.

A polyglot workflow model was implemented to integrate Python-based AI models with TypeScript orchestration logic. Temporal’s SDKs allow both environments to operate within a single workflow system, aligning machine learning execution with application-level coordination.

Temporal Cloud provides managed infrastructure, reducing DevOps overhead while supporting multi-tenant, parallel workflows at scale. Permissioning and auditability are enforced at the workflow level, ensuring reliable execution in production environments.

What now exists is a production system where AI agents operate as durable workflows capable of handling complex, multi-step customer interactions. The system runs in production, processes large volumes of requests concurrently, and is owned and extended by the Gorgias engineering team without dependency on external support.

At the Snapdragon 2023 event, Qualcomm required an interactive system to demonstrate on-device AI capabilities through real-time user interaction. The goal was to show how voice input, language models, and image generation can operate together on mobile hardware.

A voice-driven application was built where users provide partial prompts that are transcribed and completed in real time. Speech input is processed using Whisper, producing structured text that feeds into a generative pipeline. The system interprets user input and generates corresponding outputs, including dynamically created text and images.

Image generation runs on-device using Fast Stable Diffusion on Snapdragon hardware, enabling visual outputs to render in under four seconds. This allows the system to respond immediately to user interaction without relying on external processing.

A backend service coordinates communication between components, including voice processing, prompt handling, and image generation. The system maintains synchronized state across the mobile device and external displays, ensuring that generated outputs are rendered consistently during live interaction.

The application follows a constrained interaction model where user inputs are structured into predefined patterns, allowing predictable execution while still demonstrating generative capabilities. This ensures reliability in a live event environment.

What now exists is a production-ready interactive system that processes voice input, generates structured prompts, and renders images in real time on mobile hardware. The system was deployed at a live event and operates as a working demonstration of on-device AI workflows without reliance on external infrastructure.

Temporal is an open-source orchestration platform for building reliable, long-running workflows. To extend its adoption into the PHP ecosystem, an official PHP SDK was developed to enable native workflow execution for PHP developers.

The system required full compatibility with Temporal’s core architecture, including workflow and activity execution, while maintaining parity with existing SDKs written in Go, Java, and TypeScript. There was no prior PHP implementation, so the SDK was built from first principles based on Temporal’s internal design and APIs.

The SDK integrates with Temporal Core written in Golang, enabling communication with the Temporal server while preserving a PHP-native developer experience. Workflow definitions, activity handling, and state management follow Temporal’s execution model, allowing PHP applications to run durable workflows with retry logic, fault tolerance, and deterministic replay.

Development was carried out in close collaboration with Temporal’s engineering team to meet strict performance and reliability requirements. The SDK was designed to align with PHP best practices and provide a consistent interface for developers adopting Temporal without switching languages.

Documentation and integration guides were authored alongside the SDK to support adoption and ensure developers can implement workflows without relying on external support.

What now exists is an officially supported PHP SDK that expands Temporal’s language ecosystem and enables PHP developers to build production-grade workflow systems. The SDK is part of Temporal’s open-source infrastructure and continues to evolve through ongoing contributions, with full ownership residing in the open-source community.

Alaska Rubber Group required a CPQ system that could handle complex product configurations and pricing logic in a consistent and reliable way. Their existing quoting process relied on spreadsheets, email threads, and senior sales engineers, creating bottlenecks, inconsistencies, and limited scalability.

A production CPQ workflow system was built around explicit decision stages and system-enforced rules. Quoting logic was decomposed into structured workflows where configuration, validation, and approval steps are handled deterministically. A rules engine encodes product dependencies and pricing constraints, ensuring that invalid configurations are blocked before a quote is generated.

Temporal workflows orchestrate the quoting process, automating routing, enforcing execution order, and reducing the need for manual intervention. Human involvement is limited to defined exception points, while standard cases execute end-to-end within the system. This replaces ad hoc coordination with traceable, repeatable workflows.

A unified data model was introduced as a single source of truth for product and pricing data. This eliminates discrepancies caused by fragmented information across spreadsheets and internal communications, ensuring consistent outputs across all quotes.

The system shifts quoting from expert-driven execution to a controlled workflow that any sales team member can operate. Decision logic is transparent and auditable, making outcomes predictable and easy to review.

What now exists is a production CPQ system that handles complex configurations through rules-driven workflows. Quote generation that previously required extended back-and-forth now completes in minutes, with consistent logic applied across all transactions. The system is owned by Alaska Rubber Group and can be extended without reliance on external engineering support.

A beauty and wellness SaaS platform with an AI receptionist needed to support reliable call handling on mobile devices. Their browser-based calling system failed under real-world conditions where devices are locked, apps are backgrounded, and operating systems restrict execution. Missed calls translated directly into lost bookings.

The work focused on defining a production-ready mobile VoIP architecture that behaves like a real phone without requiring a full backend rewrite. A native execution model was designed where devices register with the system and can be awakened at the operating system level to receive inbound calls. The backend manages call state, holds incoming requests, and routes them once a device becomes available.

Routing logic was redefined from static phone numbers to device-aware resolution. The system determines in real time which device should receive a call based on activity and availability, supporting scenarios where staff use multiple or shared devices. This removes reliance on fixed endpoints and allows routing behavior to evolve without changes to core orchestration.

Backend extensions were introduced to support device state tracking and routing decisions while preserving the existing system structure. Event-driven components handle call delivery, queuing, and fallback behavior.

A hardware-based validation model was established to test reliability under real operating conditions, including locked screens, background execution limits, and concurrent call handling. Device and OS coverage requirements were defined as part of the system design.

What now exists is a validated architectural foundation for a mobile VoIP system that can operate reliably in production environments. The client owns a clear implementation path, including routing logic, backend extensions, and QA requirements, without dependency on external engineering support.

The client operates a vertical CRM built on top of Salesforce for distribution and manufacturing companies. The system contained critical sales data across custom objects such as accounts, quotes, invoices, and playbooks, but adoption was low due to the complexity of navigating multiple interconnected screens.

A conversational runtime was built to replace manual CRM navigation with natural-language interaction. The system runs on a multi-agent architecture using Wippy and Temporal, where agents query live Salesforce data, interpret relationships across objects, and return structured responses in real time. Sales reps can access account context, log activities, and receive recommendations without navigating the underlying UI.

The architecture integrates structured Salesforce data with proprietary sales playbooks. Initial validation was achieved by embedding playbook logic directly into agent instructions, with a defined path toward a retrieval-based knowledge system for scalable document ingestion.

All interactions execute through API and WebSocket endpoints, preserving the client’s org-level authentication model and avoiding changes to their existing security layer. The conversational interface is embedded directly into the client’s React application.

To support scale, a centralized multi-tenant registry was introduced. Agents and workflows are version-controlled and deployed across 30+ isolated Salesforce orgs without duplication. Feature flags and Git-based configuration allow controlled rollout and updates across tenants.

What now exists is a production system where CRM interaction is handled through agent-driven workflows instead of manual navigation. The client’s team can extend the system by composing new agents and workflows independently, without relying on external engineering support.

An early-stage product team set out to improve how executives prepare for earnings calls. The existing workflow relied on manual aggregation of transcripts, competitive analysis, sentiment interpretation, and briefing assembly under tight deadlines. Critical steps were undocumented and depended on individual judgment, creating inconsistency and operational risk.

The work focused on system-level discovery. The earnings call preparation process was reconstructed end-to-end, including informal steps, stakeholder handoffs, and data dependencies. A workflow decomposition map was built to define each stage, from data ingestion and transcript extraction to summarization, sentiment tagging, script drafting, and Q&A preparation. Time costs and failure points were identified across the pipeline.

A system architecture was defined around a knowledge-first foundation. Verified transcripts, extracted Q&A segments, and sentiment signals are structured into a central knowledge base. AI operates on top of this layer, generating summaries and draft outputs only from traceable source material.

Clear operational boundaries were established for AI. Tasks such as ingestion, extraction, and tagging are defined as automation candidates with auditability. Higher-risk steps such as script approval, tone interpretation, and final messaging remain under human control. Agent roles were specified with strict interfaces and handoff points to prevent overlap and error propagation.

The competitive landscape was analyzed through workflow ownership rather than feature comparison. This identified a gap in pre-call preparation, where no existing system provides structured, end-to-end support.

What now exists is a defined system architecture and Phase One blueprint. The client owns a clear market position, a decomposed workflow model, and an extensible design for introducing AI agents without compromising trust, compliance, or executive control.

Danfoss manufactures safety-critical hose assemblies sold through a distributed network of independent distributors. Product configuration requires aligning multiple technical variables, historically handled by experienced staff with years of domain knowledge. This created a bottleneck that limited scale and introduced risk in manual configuration.

A multi-tenant product intelligence platform was built to encode this expertise into a system-enforced rules engine. The system evaluates compatibility at each step of the configuration process, narrowing valid options dynamically and blocking invalid combinations before a quote is generated. This replaces manual validation with deterministic logic that runs consistently across all distributor locations.

The platform runs as a tenant-isolated architecture powered by Wippy. Each distributor operates within its own co-branded environment while Danfoss maintains centralized control over catalog data and safety rules. Tenant isolation ensures that local customization does not compromise global governance.

To remove integration friction, the system supports self-service deployment through tokenized iframe embeds and provisioned subdomains. Distributors can activate independently of their existing CMS or ecommerce stack, eliminating custom integration work and reducing onboarding time.

A license activation flow with built-in EULA acceptance tracking provides auditable onboarding and automated provisioning. The system is hosted independently on AWS ECS, decoupled from manufacturer infrastructure, allowing direct control over deployment and operations.

What now exists is a production system that converts expert-driven configuration into a rules-driven workflow. Distributors can generate valid assembly configurations and quotes in minutes, while Danfoss retains control over product logic and network-wide visibility. The platform is owned and operated independently without reliance on external engineering support.

CarAdvise operates a transaction-heavy automotive marketplace where pricing, reconciliation, and payouts depend on consistent data across multiple systems. As volume increased, small data defects led to operational escalations, delayed payouts, and repeated manual reconciliation based on implicit knowledge.

The system lacked canonical definitions for when data, orders, and payments could be considered trustworthy. This created ambiguity around financial readiness and introduced risk when attempting to automate decisions.

The work focused on defining a control layer where AI agents operate within strict constraints. A read-only Data Health Agent architecture was specified to continuously evaluate vehicles and orders through event-driven triggers and scheduled scans. The agent detects anomalies, classifies business risk, and routes structured context for early intervention, without modifying source data.

A reconciliation model was introduced with agent-defined states derived from three-way matching between internal orders, payment records, and vendor statements. These states exist outside core schemas and separate “completed” from “payout-ready,” removing ambiguity in financial workflows.

An exception classification framework was defined to structure how reconciliation issues are categorized and resolved. This allows repeated issues to be captured systematically and reused in future workflows.

All agent interactions are constrained by design. Agents validate, classify, and recommend actions but do not update financial data or approve payouts. Authority remains with human operators, ensuring auditability and control.

What now exists is a defined agentic control system that makes financial workflows observable and deterministic. The architecture, states, and agent patterns are owned by CarAdvise and can be extended without vendor dependency.

Staq is a B2B embedded finance platform operating across Saudi Arabia, Iraq, and the UAE. The company set out to build a B2C autonomous banking application where AI agents manage financial actions such as subscription cancellation, overdraft protection, and cash sweeps with human approval that can transition to full automation.

The system needed to support real money movement in regulated markets. The core challenge was not model capability but how to run AI agents inside deterministic, auditable workflows with strict isolation and compliance guarantees.

A Temporal-centric orchestration model was built to handle agent-generated workflows. Each workflow maps to Temporal primitives with typed inputs and outputs, retries, and idempotency. This ensures that agent-driven processes remain deterministic on replay and can be audited end to end.

Agent lifecycle behavior was defined at the system level. The runtime supports versioning, rollback, compensation logic, and human-in-the-loop approval flows. These controls allow agents to operate within bounded execution paths while still supporting dynamic decision-making.

To meet regulatory requirements, a reference layer was introduced that prevents personally identifiable information from entering workflow state or LLM contexts. Agents operate only on opaque identifiers. All sensitive data is stored separately, encrypted at rest, and accessed through permission-controlled interfaces.

The runtime enforces zero-trust isolation. Agents execute in sandboxed environments with no direct database access. All actions are gated through policy enforcement points using OPA/Rego. Access is scoped through issuer-specific tokens managed by Keycloak. This ensures that agents cannot access unauthorized resources or cross tenant boundaries.

Agent execution loops can generate large volumes of workflow steps. To prevent performance degradation, execution guards were implemented to enforce step and cost limits. Long-running workflows are segmented using continuation patterns, keeping execution logs bounded and replay times predictable. Telemetry systems track latency, cost, and error rates in production.

The architecture separates deterministic workflows from interactive agent behavior. Predefined agents run as standard Temporal workflows, while interactive chat-based agents are isolated behind a separate orchestration layer built on Wippy. This prevents unpredictable agent behavior from affecting core financial workflows.

The system runs on an open-core model using Wippy. The runtime core can be modified and deployed independently, while enterprise features such as audit exports and telemetry can be added as needed. This structure avoids vendor lock-in while reducing the need to build infrastructure such as artifact storage, message versioning, and isolation layers from scratch.

A key requirement was that Staq owns the system and can extend it without external dependency. The architecture, patterns, and workflows were transferred directly to the internal engineering team. The system is designed so new agents and workflows can be composed independently.

What exists now is a production-ready architectural foundation for an autonomous banking platform. It defines how AI agents, durable workflows, and compliance controls operate together in regulated environments. The system handles multi-agent execution, enforces isolation and policy constraints, and supports auditability across jurisdictions.

Saysh is a community-oriented brand, founded by an Olympic gold medalist, that creates products for and by women. We helped them to build a number of landing pages with beautiful forms for mobile and tablet, which were designed to collect emails in their brick and mortar store and sync with their online CRM system. We then launched an NFT focused landing page with a number of NFT offerings on it. We integrated Klaviyo as their main marketing tool, and wrote a number of custom Shopify solutions from scratch for their specific business needs. 

Creating an engaging modern blog that users wanted to read. Zoom uses their blog to spotlight stories of exceptional people using their video communications application in creative ways. They were in dire need of a design refresh to give the blog a more modern look and feel with faster load times, more relevant articles for readers, and better search and subscriber signup capabilities.

Spiral Scout dedicated a six-person team to the Zoom blog redesign project, including a project manager, multiple front and back end developers, and a lead designer. After reviewing the existing blog, brand book, style guide, and 20 blogs of similar corporate IT companies, our team assembled a complete UI Kit for the new Zoom blog. The kit included typography, color palettes, and all element states based on the clients’ requirements.

During development, our team prioritized international localization to ensure the new Zoom blog was accessible and translatable for readers across the globe.

TECHNOLOGIES:

WordPress, Polylang, Lingotek, Pardot, HotJar, TrustArc, YoastSEO, Advanced Custom Fields PRO

Learn more on our website

Empowering students and landing multi-million dollar deals. Benchmark Education is transforming education for the digital age through custom eLearning development in the classroom. As a premier educational publisher, Benchmark Education is on a mission to support all students with the tools they need to succeed including struggling readers, gifted and on-level students, students with learning disabilities, and those still mastering English as a second language.

Benchmark was looking to scale up their eLearning curriculum development and knew they wanted an outsourced technology partner who could support their internal technical team. They hoped that additional engineering resources would help them get more software products out to market within a short period of time and allow them to surpass their competition.

The Spiral Scout team began by quickly working to gain a deeper understanding of what Benchmark did as a company and what their business goals were. We assembled a top-tier team of eLearning app developers who could lend their specialized expertise and experience in building educational tools. Our team worked closely with the internal Benchmark Education software development team as well as Benchmark product owners who helped provide direction for their digital dream. 

While the majority of our work was done remotely from our Minsk offices, we regularly visited the Benchmark team at their offices in New Rochelle, New York. We worked seamlessly together to build out desktop and iPad versions as well as put together documentation to share with other Benchmark departments. This helped with the organization and getting sign-off from major stakeholders.

The software we helped to write led to Benchmark winning a multi-million dollar contract with the LA Unified school district and put it in the hands of millions of students in California. 

Learn more on our website

Ecommerce Payment App and Inventory Management System. Chequeout provides a novel way for shoppers to purchase merchandise during physical pop-up events, virtual events, or when paying directly on a website checkout. 

Their mobile payment app and inventory management system facilitates quick and easy payments through scan-and-go QR codes - bringing contactless shopping and pickup to in-person purchases, while also adding a new dimension of functionality to virtual events hosted on Zoom that seek to replicate intimate purchasing experiences like QVC or Tupperware Clubs.

By combining flexible selling tools, a robust inventory management system, and helpful reporting, Chequeout wanted their platform to be the payment tech solution that removes purchase barriers and unleashes business potential. The checkout process had to be simplified down to a single QR code scan, and streamlined platforms had to be built for inventory management, order processing, shipping platforms, and a mobile application.

After reviewing the scope of work with the Chequeout team, we mapped the key project goals and built a plan to include specific features on the Chequeout app and CMS.

The original plan was to introduce Chequeout as a physical, in-person scan-and-go app with a backend inventory piece, with the virtual event capabilities coming later in the roadmap. However, the app happened to be released just before the pandemic hit in March 2020.

With brick and mortar stores suddenly closing and in-person pop-up stores being delayed, we had to quickly pivot and transform the product into a fully online solution, converting our beta product into a live product. From the first day, there was a huge demand and the platform had hundreds of merchants and shoppers using Chequeout over Zoom to facilitate ordering and payments using QR codes. 

TECHNOLOGIES:

Spiral Framework, React.JS, React.Native, TypeScript, Kubernetes, PostgresQL, AWS, AWS SQS, AWS SNS, Stripe SDK.

Learn more on our website

Supercharging sales and raising $8M in seed investment. Made In Cookware is one of the fastest-growing premium cookware brands in the U.S. They are on a mission to make home cooking fun, accessible, and affordable with professional-quality cooking essentials all made in America. As a web-only retailer, their direct-to-consumer approach is revolutionizing the kitchen goods market like never before.

We know that the best eCommerce website development practices are centered around the online customer - from the way they view and interact with products to the ease of checkout and even how they engage with the brand. We started this project by giving Made In Cookware several different options for setting up a user-focused and mobile-friendly website with an in-depth list of the pro's and con's associated with each one. With a very tight budget and a one-month launch deadline, we knew we needed to hit the ground running. Once Jake and Chip had made their choice, we worked with them to quickly scope out the project and finalize the UI/UX designs and their Shopify needs.

Made In Cookware continues to grow like crazy with sales through the roof, and they have raised over $8M in seed money as of 2019 to help expand their team and product line.

Spiral Scout provides ongoing outsourced web development and website maintenance support for Made In Cookware. Since the initial website was built, our team has created HTML email templates for their business and integrated them into their email service provider, Mailchimp. We also launched a survey tool for them to gauge customer feedback and interest in new products so they could make more informed business decisions around their offerings.

TECHNOLOGIES:

Shopify, JavaScript, HTML5, CSS3, Vue.js, Facebook API, Mailchimp API, Segment.io, Google Optimize, PPC campaign, SEO optimization.

Learn more on our website

Flow Reading Fluency is an online learning platform that enables students aged 7-17 to increase their reading fluency and vocabulary with the guidance and monitoring of their teachers. They asked us to build their platform from the ground up in 2020 on Moodle, embedding video, audio, quizzes and additional educational content in a streamlined and customized way. 
 

Using a combination of Moodle’s existing tech framework, and our own custom-built plugins, we successfully created Flow Reading Fluency’s 100% digital platform in a few months. The platform now helps thousands of children improve their fluency and vocabulary in person and online with minimal teacher supervision.

Learn more on our website

Framed & Matted launched in 2013 to change the way people custom frame their favorite prints, posters, photos, and most works on paper. They consistently deliver elegant, simple, and classic picture frames.
 

Buying a picture frame for a poster, print, or artwork you own is expensive and limiting. Framed and Matted wanted to design a solution that would allow their customer to customize a picture frame all the way down to the nitty-gritty details.
 

We helped them zero in on the best user experience for building a custom frame builder that took the user down a path towards creating the perfect custom frame. This involved a number of user experience studies and research until we found an ideal solution.
 

Spiral Scout built the entire site, including an eCommerce gateway, an easy to use page editor, an integrated shipping system (via the FedEx API), a complex CMS to manage user-created frames, a reseller's API and referral program, and a robust customer relationship dashboard to manage all members and orders.
 

TECHNOLOGIES: FedEx API, Stripe, Image Magick, Spiral CMS, PHP5, PayPal Integration, Apache, MySQL, Spiral Framework, HTML5 Canvas, CSS3, jQuery
 

Visitors can customize their frame via an online frame builder that allows them to drag and slide their frame to the correct size, add matting to the borders and choose their frame style.

Learn more on our website

After an earthquake occurs, licensed structural engineers and city inspectors are called on to assess buildings and identify structural damage as well as clear the buildings for reoccupation. As part of the STIR program, the San Francisco Public Works department wanted to bring their vision of a natural disaster app for government bodies to life to help reduce the time and costs surrounding these types of building inspections and streamline the efforts of first responders. 

Decovery also facilitated more organization around assigning areas of a city to inspectors and gathering building evaluations to create robust damage reports. Ultimately, Decovery was created for use by any municipality or large business to review their buildings and create actionable building reports.

CHALLENGE
Over 16 weeks, Spiral Scout worked with the Department of Public Works and an outstanding team of stakeholders there to design and develop a disaster recovery app that would change the way the city would react to a manmade or natural disaster, like an earthquake for example. Decovery is short for Disaster Recovery.

The Spiral Scout team knew that developing mobile apps for city government use meant creating a solution that worked quickly and effectively both on a computer and a mobile device. So we designed an “in the field” mobile app and a responsive admin website.

The app allowed field inspectors to quickly compile an electronic inventory of city buildings and structures, record important information about a building (pre and post-natural disaster), and send that data back to the admin dashboard. The data collected would inform reports that could be generated in real time for city officials about the buildings and areas in a disaster zone.

RESULT
The Spiral Scout team built Decovery as a part of the prestigious STIR (Startups in Residence) program working with the city of San Francisco’s Public Works department. STIR connects SF government agencies with technology startups like Spiral Scout to develop software products that solve civic challenges.

Learn more on our website

Destination 4 Rent’s online vacation rental marketplace makes it easy for people to book short-term luxury vacation rental properties in the U.S. and Caribbean.
 

Destination 4 Rent was having difficulty converting visitors into customers with their original rental marketplace WordPress site. They wanted to both improve their website’s SEO and increase traffic as well as resolve technical bugs that were negatively impacting the user experience.
 

Though they had been able to create a real estate website with WordPress, the Destination 4 Rent team lacked the technical expertise in WordPress website development to be able to optimize their site for a smoother and more successful user experience.
 

Our initial approach was to involve them in a branding and design exercise that would provide insights into their website’s current UI/UX. From there we gathered additional user feedback and conducted our own quality assurance testing to identify the most problematic areas of the website. We worked with them to strategize key updates to their web and mobile sites that would more aptly address the needs of a user looking to book a vacation rental.
 

Because the Destination 4 Rent team was most experienced with creating a business website in WordPress and had a limited budget, our developers built out their revamped website on top of WordPress. 
 

We crafted and customized various plugins and pulled in search results via a third-party provider called VRP Connector that was using the HomeAway API. We then integrated Stripe, their payment processor, into the new site so reservations could be seamlessly booked and sent to the third-party booking agent who would then notify the listing agents.
 

We knew that building customer confidence and trust was key to the redesign. We put enhancing the search tools and engineering a smoother and more secure user checkout at the top of the list.

Learn more on our website

Attracting enterprise clients and saving $100k+ each year. Demos are critical to the buying experience. But pre-sales teams often struggle to cope with increasing demand for live demos, and growing a team of expert solution consultants takes time. 

Consensus helps businesses create a more successful and tailored sales process through video asset management solutions and interactive video demos. The custom solution allows their customers to deliver a fully customized and interactive video demo on demand to each new prospect in the sales process. 

Then, as an internal “champion” shares a Consensus video with other stakeholders at the company, the system automatically discovers each of their unique interests and re-configures the sales video to educate the new stakeholders in a more personalized way.

The key to scaling for Consensus was attracting more enterprise-level clients like Oracle, IBM, and HP. The demands of these types of Fortune 500 companies required Consensus to enhance their demo automation platform with a more comprehensive website system architecture, higher-level security procedures, compliance with European laws, and features like single sign-on.

While we supported, maintained, and fixed the existing application focused on digital asset management for video production, we were also hard at work estimating, planning, and creating a new system architecture design to replace it.

Right from the start, the upgrades our team made helped them eliminate the need for a costly third-party platform, saving them over $100k a year, as well as lowered their overall spending associated with servers, hosting, and video storage by a few thousand dollars a month.

TECHNOLOGIES:

Amazon AWS, PHP5/PHP7, Node.js, Codeigniter, Golang, Spiral Framework, MySQL, Influxdb, SQL, ORM, RPC, auto-scaling video processing with FFMPEG, Elixir

Learn more on our website

Building the world’s largest library of eLearning assets. Before eLearning Brothers was the global provider of custom eLearning design, eLearning templates, and online training for eLearning professionals, they were a couple of brothers from Salt Lake City, UT with a dream - to make online learning more engaging, more inspiring, and more awesome.

eLB simultaneously needed a customer relationship management service that would organize all their customer's data for their sales team and included sales-tracking history with helpful alerts sent to salespeople when a subscription was about to expire.

Spiral Scout engineered a robust, subscription-based digital asset management system (DAM) with a scalable asset library and admin dashboard. Why was a DAM system the answer? The business case for building a digital asset management system was all about saving time, boosting efficiency, and improving security. 

The custom DAM system Spiral Scout built provided for both structure, organization, and a secure central repository that equipped any eLB site admin with the digital asset management tools they needed to upload multiple files, associate metadata with them, and offer them to customers for immediate download. 

Since 2012 when eLearning Brothers was just a handful of employees, their team has grown to over 90 members, consistently being highlighted as one of the fastest-growing companies in Utah and the U.S. (most recently being honored in the Inc. 5000 Hall of Fame).

TECHNOLOGIES:

SphinxSearch, ImageMagick, CMS, PayPal, USAePay, PHP5, Apache, MySQL, Spiral Framework, HTML5, ECMAScript 5, CSS3, Spiral JavaScript toolkit, LESS CSS preprocessor. Template driven, Two-way data-binding development with Ractive.js, Audio library SoundManager 2.

Learn more on our website

Drew Altizer Photography is a leading full-service photography agency specializing in event and portrait photography, video production, PR, and media placement in San Francisco and the surrounding Bay area. 

Like what happens with so many businesses when they start to experience substantial growth, Drew’s legacy platform for his website was unable to sufficiently scale as business picked up nor could it handle the technical demands of things he wanted customers to be able to do, like purchase photos from different high-profile events he was shooting around San Francisco. 

He was losing sales daily and lacked the internal technical resources to create a photo-sharing website with a bunch of extra bells and whistles.

Drew needed a web development company in San Francisco that specialized in digital asset management systems. 

He was looking to engineer a brand new responsive website that both communicated important information about his services as well as allowed for millions of event photographs to be stored, shared, purchased, and distributed to the press, media clients, event guests, and the people throwing the event. He also needed to seamlessly allow website visitors to purchase digital downloads and order prints from a printing service he was using in the Bay Area.

Drew had an additional unique challenge in that he needed a seamless way for his photo editors and media clients, both those who were asking him to shoot events and those looking to buy images from the events, to edit and tag images in real-time while working in different locations. This would allow them to immediately distribute photos to guests and clients.

The Spiral Scout web development team kickstarted the process by drafting a technical review document that outlined the entire scope of the project. Knowing that the web design for the photography business needed to be artistic, intuitive, and on-brand, we worked closely with Drew to clearly articulate his vision and business goals. 

Using that document as the blueprint, Spiral Scout assembled a team to execute the responsive website; the team included a project manager, a web designer, three engineers, and a quality assurance engineer.

We essentially ended up building four admin interfaces or websites into one web application. The fully customized solution included a robust photo digital asset management component, a content distribution system for clients and the media, a fully custom eCommerce platform with a built-in customer relationship management tool for handling users and online purchases, and a collaborative editing workspace for clients, editors, and photographers so they could work in real time once an event was over to get the images out to press. We also tackled the sizeable task of migrating over all of Drew's existing photos and metadata.

In the end, Spiral Scout’s web development expertise and fast execution helped deliver the final product on time and under budget based on a clearly defined spec that was written upfront and our client's feedback.

Not only was it easier for Drew Altizer Photography to work and collaborate with key stakeholders that helped generate profit for his business, but the integrated custom-built eCommerce website development also drastically improved his online sales while generating more useful metrics and sales reports.

Learn more on our website