12.3 Comprehensive Case Studies: Altostrat Media & KnightMotives Automotive
Key Takeaways
- Altostrat Media already runs on GKE, Cloud Storage, BigQuery, and Cloud Run functions; its case study is about layering generative AI (recommendations, natural-language interaction, self-service, summarization) onto that estate while modernizing CI/CD and controlling storage cost.
- Altostrat's AI requirements include detecting and filtering inappropriate content and making AI systems auditable and explainable — mapping to Vertex AI custom models with Vertex Explainable AI, plus Model Armor and Sensitive Data Protection for generative AI safety.
- KnightMotives Automotive must modernize the in-vehicle experience across BEV, hybrid, and ICE models within five years while its unreliable build-to-order system strains dealer relationships — and dealers have no budget for new equipment, ruling out dealer-side hardware solutions.
- KnightMotives monetizes corporate data to fund AI investment, so a governed BigQuery-centered data platform (with Dataplex, IAM, CMEK, and audit logging) that unifies siloed vehicle, corporate, and dealer data is the core architectural answer.
- KnightMotives' security and compliance constraints — past data breaches and EU data protection rules for autonomous platforms — map to Security Command Center, Sensitive Data Protection, Assured Workloads, and resource-location Organization Policies; Cloud IoT Core is retired, so vehicle telemetry uses a partner MQTT broker feeding Pub/Sub.
Comprehensive Case Studies: Altostrat Media & KnightMotives Automotive
Architectural Objective: This section completes the four-case-study set with Altostrat Media and KnightMotives Automotive. (Mountkirk Games, TerramEarth, Helicopter Racing League, and Dress4Win were retired in the October 2025 exam refresh; questions on the current exam draw only from EHR Healthcare, Cymbal Retail, Altostrat Media, and KnightMotives Automotive.)
Case Study 3: Altostrat Media (In-Depth Teardown)
Company Overview & Solution Concept
Altostrat Media is a prominent media company with an extensive library of podcasts, interviews, news broadcasts, and documentaries. Its content management must keep pace with a dynamic media landscape. The solution concept is to modernize content management and user engagement using Google Cloud's generative AI: a platform that empowers customers with personalized recommendations, natural-language interactions, and seamless self-service support, while driving revenue growth through dynamic pricing, targeted marketing, and personalized product suggestions.
Existing Technical Environment (Already Substantially Cloud-Native)
- GKE runs the content management and delivery platform for scalability and high availability.
- Cloud Storage holds the media library across documents, audio, and video formats.
- BigQuery is the primary data warehouse for user behavior and consumption analytics.
- Cloud Run functions execute event-driven tasks — video transcoding, metadata extraction, personalized recommendations.
- Some legacy on-premises systems remain for content ingestion and archival (slated for modernization).
- Identity via Google Identity plus third-party identity providers; monitoring mixes Cloud Monitoring and open-source Prometheus, with email-based alerts.
Business & Technical Requirements Mapped to Architecture
| Altostrat Requirement | Google Cloud Architecture |
|---|---|
| Natural-language interaction with 24/7 user support | Conversational Agents (virtual agents) with Conversational Insights-class analytics on interactions |
| Automatically generate concise summaries of media content | Natural Language AI generic summarization plus Vertex AI domain-tailored summarization models, orchestrated with Workflows |
| Extract rich metadata from media (NLP + computer vision) | Video AI / Vision AI labels, objects, explicit-content detection |
| Detect and filter inappropriate content | Vertex AI custom models for harmful-content detection, hardened with Model Armor and Sensitive Data Protection |
| AI systems auditable and decisions explainable | Vertex Explainable AI plus governed model deployment (private endpoints, VPC Service Controls) |
| LLMs and conversational AI for personalized experiences | Conversational Agents with natural-language understanding for personalized assistance |
| Modernize CI/CD for containerized deployments with a centralized management platform | Cloud Build + Cloud Deploy + Artifact Registry, with GKE Enterprise/Anthos spanning on-premises and cloud clusters |
| Secure, high-performance hybrid connectivity for data ingestion | Dedicated/Partner Cloud Interconnect with HA VPN backup |
| Scalable, performant Kubernetes on-premises and in cloud | GKE + GKE Enterprise (Anthos) |
| Optimize cloud storage cost for growing media volumes | Cloud Storage Autoclass / lifecycle rules (Archive class for assets rarely accessed) |
Altostrat Target Architecture Highlights
- Generative engagement layer: Conversational Agents front the audience experience; Video AI, Vision AI, and Natural Language AI enrich the library; Vertex AI hosts custom moderation models with Explainable AI for auditability.
- AI safety perimeter: Model Armor screens prompts and outputs for injection and harmful content; Sensitive Data Protection prevents PHI/PII-class data leaking into prompts or training sets; secure model deployment uses VPC-scoped private endpoints.
- Cost and operations: Autoclass tiers cold media to cheaper storage classes automatically; centralized CI/CD retires per-team pipelines; real alert policies replace email alerting.
Case Study 4: KnightMotives Automotive (In-Depth Teardown)
Company Overview & Existing Environment
KnightMotives is a global vehicle manufacturer building autonomous, self-driving vehicles alongside BEV, hybrid, and internal combustion lines. Its BEV fleet has a modern in-vehicle experience, but hybrid and ICE models lag, hurting sales and satisfaction. KnightMotives wants to modernize the consumer experience across all vehicles within five years, positioning AI at the center of driving, shopping, buying, and service.
+-----------------------------------------------------------------------------------+
| KNIGHTMOTIVES EXISTING TECHNICAL ENVIRONMENT |
+-----------------------------------------------------------------------------------+
| HOSTING | Largely on-premises, some apps on major cloud platforms. |
| SUPPLY CHAIN | Outdated mainframe; outdated ERP makes promotions and dealer |
| | discounts hard to implement. |
| DEALERS | No budget for new equipment; unreliable online build-to-order |
| | system strains KnightMotives-dealer relationships. |
| CODE | Fragmented across vehicle lines; heavy backward-compat debt. |
| NETWORK | Connectivity to manufacturing plants and rural vehicle coverage |
| | are both weak. |
+-----------------------------------------------------------------------------------+
Business Requirements
- Personalized driver relationship and cohesive UX across every model — including legacy vehicles needing hardware/software updates.
- Better build-to-order model — less lot time, transparency for dealers and customers.
- Monetize corporate data to finance AI investment — current AI infrastructure is obsolete and corporate data is siloed.
- Security is paramount after past data breaches; EU data protection compliance is critical for the autonomous platform.
- Autonomous driving investment targeted first at regions with favorable regulation; employee upskilling and better business-technical communication.
Technical Requirements & Mapped Architecture
| KnightMotives Technical Requirement | Google Cloud Architecture |
|---|---|
| Consistent AI-powered in-vehicle UX across all models, reliable even in rural coverage gaps | Offline-tolerant vehicle design: edge buffering and asynchronous sync to Pub/Sub; Vertex AI models served via managed endpoints |
| Autonomous vehicle development and testing | Vertex AI with Cloud TPUs (AI Hypercomputer) for perception/sensor-fusion training plus simulation data pipelines at large scale |
| Vehicle telemetry at scale | Partner MQTT broker (Cloud IoT Core retired) -> Pub/Sub -> Dataflow -> Cloud Bigtable time series; BigQuery for fleet analytics |
| Data monetization with strict security and privacy | Governed BigQuery + Dataplex data platform unifying corporate, vehicle, and dealer data with IAM, CMEK, and audit logs; partner sharing via Analytics Hub / Apigee-managed APIs |
| Network upgrades between plants and HQ | Network Connectivity Center hub with Cloud Interconnect / Cloud VPN spokes; partner SD-WAN appliances for rural plants |
| Hybrid cloud strategy with gradual legacy replacement | GKE Enterprise (Anthos) consistent platform; Apigee façades in front of mainframe/ERP during phased modernization — never a one-step rewrite |
| Security framework after past breaches, incident response, awareness training | Security Command Center, Sensitive Data Protection, Binary Authorization for in-vehicle software supply chain, Google Threat Intelligence; CMEK via Cloud KMS |
| EU data protection for the autonomous platform | Assured Workloads, resource-location Organization Policies, EU regional processing and storage |
| Dealer tools and build-to-order modernization (dealers have no hardware budget) | Cloud-hosted dealer portal (App Engine/Cloud Run + IAP), CRM integration, and Apigee-managed partner/dealer APIs |
KnightMotives Target Architecture
+-----------------------------------------------------------------------------------+
| KNIGHTMOTIVES AUTOMOTIVE TARGET TOPOLOGY |
+-----------------------------------------------------------------------------------+
| |
| VEHICLES (rural-tolerant) PLANTS & HQ GOOGLE CLOUD |
| [ Edge buffer -> MQTT broker ] [ SD-WAN / Interconnect ] |
| | | |
| v v |
| [ Pub/Sub ] <--------[ Network Connectivity Center hub ] |
| | |
| v |
| [ Dataflow ] --> [ Cloud Bigtable telemetry ] --> [ BigQuery (BQML) ] |
| | |
| [ Dataplex governance / monetization ] |
| | |
| [ Vertex AI + Cloud TPUs: AV training & simulation ] v |
| v |
| [ Apigee dealer/partner APIs ] -> [ Dealer portal: |
| Cloud Run + IAP ] |
| SECURITY: SCC + Sensitive Data Protection + CMEK; Assured Workloads (EU); |
| Binary Authorization for vehicle software supply chain. |
+-----------------------------------------------------------------------------------+
The KnightMotives Exam Traps
- Do not propose a one-step rewrite of the mainframe and ERP estate — the case study demands a hybrid strategy with gradual modernization; façade legacy systems with Apigee and migrate incrementally.
- Do not spec dealer-side hardware — dealers explicitly have no budget; dealer tooling must be cloud-hosted.
- Do not cite Cloud IoT Core — it was retired in August 2023; vehicle telemetry arrives via a partner MQTT front end into Pub/Sub.
- Respect the EU constraint — autonomous-driving data for European operations requires residency controls (Assured Workloads, location Organization Policies, CMEK), not a global single-region data lake.
Case Study Comparison Matrix: Altostrat Media vs. KnightMotives Automotive
| Dimension | Altostrat Media | KnightMotives Automotive |
|---|---|---|
| Core Challenge | Generative-AI engagement and content intelligence on an already cloud-native estate | Five-year connected/autonomous vehicle transformation on a legacy-heavy estate |
| Existing Stack | GKE, Cloud Storage, BigQuery, Cloud Run functions | On-premises mainframe supply chain + outdated ERP; fragmented vehicle codebases |
| AI Workload | Summarization, metadata, moderation, personalized conversation | Autonomous perception training (Vertex AI + Cloud TPUs), simulation, in-vehicle AI UX |
| Data Platform | BigQuery warehouse + media library in Cloud Storage | Bigtable telemetry + BigQuery monetization warehouse governed by Dataplex |
| Security Emphasis | Model Armor, Sensitive Data Protection, Explainable AI | Security Command Center, CMEK, Binary Authorization, breach-response maturity |
| Compliance/Edge | Auditable, explainable AI decisions | EU data protection (Assured Workloads, location policies); rural offline tolerance |
[!IMPORTANT] Exam Watch: For Altostrat Media, think "generative AI layered on an existing GKE/BigQuery/Cloud Run estate": Conversational Agents for 24/7 natural-language support, Video AI/Natural Language AI for summarization and metadata, Vertex AI + Explainable AI for auditable content moderation, and Cloud Storage Autoclass for cost. For KnightMotives Automotive, the recurring answers are Vertex AI with TPUs for autonomous development, Pub/Sub + Dataflow + Bigtable for telemetry, a governed BigQuery platform for data monetization, Apigee dealer/partner APIs, GKE Enterprise for gradual hybrid modernization, and Assured Workloads for EU compliance.
Refer to the Altostrat Media case study. Altostrat wants natural language interaction with the platform and 24/7 user support, plus concise auto-generated summaries of its podcasts and documentaries, while keeping costs predictable. Which combination best satisfies these business requirements?
Refer to the Altostrat Media case study. The technical requirements demand AI-powered detection of harmful content and that AI systems remain auditable and explainable. How should the architect satisfy both constraints?
Refer to the KnightMotives Automotive case study. Dealers need modern build-to-order and service tooling, but dealers have no budget for new equipment, and the mainframe supply chain and ERP systems can only be replaced gradually. What should the architect recommend?
Refer to the KnightMotives Automotive case study. The autonomous vehicle program will launch first in European markets, corporate data monetization must obey strict privacy rules given past breaches, and millions of vehicles will stream telemetry from regions with unreliable rural connectivity. Which architecture addresses all three constraints?