Strategy without execution artifacts is just conversation. The samples below are tactical deliverables — assessments, frameworks, architectures, roadmaps — that emerged from strategic engagements and workshops. They're shown here not because Signal & Horizon builds things, but because they illustrate what rigorous strategic work makes possible. Identifying details have been removed. The substance has not.
Governance Architecture
Benefits Enrollment Automation
Insurance — Fortune 500
A major supplemental insurance carrier was sending group benefits case setup materials — rates, product rules, brochures — to third-party enrollment platforms via email and manual SFTP. The process was error-prone, slow, and impossible to scale. The work involved scoping a GenAI-enabled data exchange architecture to automate the handoff between the carrier and five major benefits administration systems.
- Mapped current-state workflow across the full enrollment lifecycle from case setup through file processing
- Defined ideal-state business outcomes: real-time data exchange, automated confirmation, zero manual copy/paste errors
- Designed evaluation criteria for vendor solutions across feasibility, scalability, and regulatory compliance
- Developed a reverse-pitch framework that inverted the vendor relationship — carriers defining requirements, vendors proposing solutions
Scale: $100B+ carrier — 5 third-party platform integrations
Evaluation Framework
Product Schema Evaluation
Insurance / Technology — Enterprise
An enterprise organization needed to evaluate whether a large language model could reliably extract and structure complex insurance product rules from unstructured source documents. The engagement required defining schema architecture, establishing accuracy benchmarks, and building an evaluation framework that would hold up to compliance scrutiny before any production investment.
- Designed a structured product schema capable of capturing eligibility rules, rate logic, and regulatory requirements
- Established a model evaluation framework with accuracy thresholds tied to business risk tolerance
- Produced an executive summary positioning the findings for board-level capital allocation decisions
- Identified edge cases and failure modes that would require human-in-the-loop governance before production deployment
Output: Board-ready executive brief — capital allocation decision
API Governance Model
Automated API Generation on the Fly
Insurance — Group Benefits Division
A group benefits division was evaluating whether to build a modern API layer to expose internal data assets to broker partners and third-party platforms. The work required designing the API governance model, evaluating partner integration patterns, and translating technical architecture options into business-level investment decisions for a non-technical leadership audience.
- Developed an API capability assessment framework that mapped technical maturity to business value creation
- Designed partner integration tiers based on data sensitivity, volume, and compliance requirements
- Produced a readout structure that enabled cross-functional leadership alignment on build vs. buy vs. partner decisions
- Established a governance model for API lifecycle management and third-party access controls
Scope: Enterprise API layer — multi-partner data exchange
Delivery Modernization Roadmap
Legacy Code Modernization Roadmap
Enterprise Technology — Large-Scale IT
An enterprise IT organization was evaluating AI-assisted code generation tools to accelerate software delivery and reduce technical debt accumulation. The engagement involved assessing readiness across the SDLC, evaluating tooling options against delivery velocity benchmarks, and designing a phased adoption roadmap that balanced speed with code quality and security requirements.
- Benchmarked current SDLC velocity and identified highest-value automation insertion points
- Designed a tool evaluation framework that weighted security, compliance, and developer experience equally with output quality
- Developed a phased rollout model that isolated risk to non-production environments in early phases
- Established guardrails for AI-generated code review that preserved human oversight at critical checkpoints
Scale: Enterprise software delivery — 500+ developer environment
Operating Model Design
Multi-Tenant Industrial Platform Design
Aerospace / Manufacturing — Global Scale
A global aerospace and industrial technology organization needed to design a multi-tenant digital platform capable of ingesting sensor data from 110 million flights and 500,000+ IoT devices. The work involved designing the data ingestion architecture, defining the operating model for four delivery teams, and structuring the analytics layer for both internal operations and airline customers.
- Defined the four-team operating model: airlines onboarding, back office, L2 support, and core API development
- Designed the data pipeline architecture from raw sensor ingestion through transformation to customer-facing dashboards
- Established the governance model for multi-tenant data access across 215 billion km of flight data
- Created the training and knowledge base structure for platform-wide adoption across airline customers
Scale: 12B passengers — 500K+ IoT sensors — Airbus/AWS cloud infrastructure
Predictive Analytics Platform
Retail Demand Forecasting
Retail / Economic Research — Japan
A $4.5B economic research and consulting firm needed a buying demand forecast system across 40+ stores, 89 product categories, and 500+ SKUs. The engagement involved selecting and implementing the model architecture, managing the training of 54 AI models across multiple regression and time-series approaches, and building the analytics layer for operational use.
- Designed and implemented a Dataiku-based ML platform deployed on AWS for scalable model training
- Trained 54 AI models using Random Forest, Ridge/Lasso regression, ARIMA variants, and moving average methods
- Achieved near-zero MAPE on select categories; 29 categories achieved under 30% error on monthly basis
- Integrated COVID-19 data to cross-validate and extend model accuracy during atypical demand periods
Scale: 10B records — 16M customers served — 6,300+ employees
All project descriptions reflect real work. Identifying details have been removed. The sector, scale, and outcome information is accurate to the engagement.
