Claude, on Bedrock, in production.
Claude is the model VeUP reaches for when reasoning quality and guardrail maturity carry the workload — and the model VeUP's own delivery engine runs on every day. It still has to win the benchmark on each engagement.
The engagements whose records name these models.
Every card below is a published case study whose record names a Claude model. The chip says what Claude actually did there — shipped it, backed it up, or contested it.
Three agents, three models. Claude 3.5 Sonnet carries the reasoning steps and Claude Haiku the latency-sensitive ones — a split settled by benchmark against Amazon Nova before launch, then hardened with Bedrock Guardrails.
Claude on Bedrock behind MCP tool servers over the CRM's existing REST APIs, with per-tenant context isolation and a human in the loop on v1.
Claude Sonnet is the top rung of a three-step ladder — Nova Lite triages, Nova Pro answers by default, Claude Sonnet takes the questions that earn it. Answers land in 1–2s against a 5s SLA.
Claude Opus was one of five candidates scored on an identical labelled dataset, all invoked through Bedrock. Qwen 3 VL won this workload — which is the point of running the bake-off.
Claude on Bedrock is the qualified alternate foundation model behind a Nova primary. The fallback is pre-tested, not hypothetical — tool-call accuracy was the deciding measure.
Claude Opus runs inside the evaluation harness: 12,000 real production prompts, LLM-judge scoring, hallucination tracking. The harness stopped a blind model switch before it shipped.
Claude Code in VeUP's own delivery loop — part of the three-week build behind a connected-vehicle model at 75B telemetry points a day.
Reasoning depth, tool-call accuracy, guardrail maturity.
Across the published engagements a consistent shape shows up: Claude takes the steps where a wrong answer is expensive, and a cheaper model takes the rest.
- Premium tier of a model ladder — Claude answers what the cheap tiers escalate (Cluster Systems).
- Reasoning agent in a multi-agent split, with Haiku on the latency path (MyCena).
- Qualified alternate foundation model, chosen on tool-call accuracy (Sofvie).
- MCP tool-calling over existing REST APIs with per-tenant isolation (SalesNOW).
35 accounts where Anthropic Claude is on the table.
These are customer accounts VeUP runs where this model family came up on a recorded customer call — architecture sessions, evaluations, roadmap conversations. That is engagement evidence, not a statement that the customer runs these models in production. Promotion to a production claim means a published case study, which is what the section above holds. 9 of these 35 are named here because they already have one. Every other row stays anonymized — including customers who do have a published case study but have not agreed to be named, who appear by description only.
- Identity held on fileTelecommunicationsBuildIgniteSellOperate28 AWS accounts
- Shipium Corp.Logistics and Supply ChainIgniteSellOperate20 AWS accounts
- An AI meeting-intelligence SaaS companySoftware DevelopmentBuildIgniteSellOperate13 AWS accounts
- An energy-sector fibre-optic-sensing technology companyOil & EnergyIgniteOperate13 AWS accounts
- Identity held on fileInformation Technology and ServicesSellOperate11 AWS accounts
- BrontoSoftware DevelopmentSellOperate10 AWS accounts
- Identity held on fileFinancial ServicesBuildIgniteSellOperate10 AWS accounts
- Identity held on fileComputer SoftwareBuildSellOperate9 AWS accounts
- Cluster Systems (product: ClusterPOS)DocumentedSoftware DevelopmentBuildIgniteOperate9 AWS accounts
- FerootComputer & Network SecuritySellOperate9 AWS accounts
- Identity held on fileConstructionSellOperate8 AWS accounts
- A multi-tenant GRC SaaS providerSoftware DevelopmentSellOperate6 AWS accounts
- Identity held on fileElectrical/Electronic ManufacturingIgniteOperate4 AWS accounts
- Identity held on fileVenture Capital & Private EquitySellOperate4 AWS accounts
- SalesNOW (Interchange Solutions Inc.)DocumentedSoftware DevelopmentSellOperate3 AWS accounts
- Identity held on fileSoftware DevelopmentSellOperate3 AWS accounts
- A consumer connected-home-security IoT platformComputer & Network SecuritySellOperate2 AWS accounts
- Identity held on file—Operate2 AWS accounts
- Identity held on fileSoftware DevelopmentIgniteSellOperate2 AWS accounts
- Identity held on file—SellOperate1 AWS accounts
- TenderdComputer SoftwareSellOperate1 AWS accounts
- Identity held on fileSoftware DevelopmentBuild0 AWS accounts
- Identity held on fileComputer SoftwareIgniteSell0 AWS accounts
- Identity held on fileManagement ConsultingSellOperate0 AWS accounts
- Identity held on fileFinancial ServicesBuild0 AWS accounts
- Identity held on fileSoftware DevelopmentSell0 AWS accounts
- Identity held on fileComputer SoftwareBuildIgniteSell0 AWS accounts
- MyCena Security SolutionsDocumented—Sell0 AWS accounts
- Identity held on fileMedia ProductionBuild0 AWS accounts
- Identity held on fileBiotechnologyIgnite0 AWS accounts
- Identity held on file—BuildIgnite0 AWS accounts
- Swftbox—Sell0 AWS accounts
- A travel-technology company building a governed traveler-data platformLeisure, Travel & TourismBuildIgnite0 AWS accounts
- Identity held on file—Build0 AWS accounts
- ZenhubDocumentedProgram DevelopmentBuild0 AWS accounts
Every account here is a VeUP customer where this family is in the work. The chips are the engagement motions VeUP runs at that account — Build means VeUP builds and operates there, Ignite that VeUP funds and resells the estate, Sell and Marketplace the co-sell and listing motions. Operate marks the estates where VeUP is the AWS payer of record. Accounts are resolved from the customer registry and the meeting archive on participant email domain. Counting conversations is not counting deployments — the number is here as scope, never as a claim. Linked AWS accounts describe estate size, not spend.
