Selected work

Operating systems, not isolated deliverables.

The examples below show the pattern behind the work: recover context, model ownership, connect systems, control automation, and make execution visible.

Customer operations

AI-ready customer service operating system

Unified email, messaging, social, commerce, knowledge, routing, QA, reporting, and AI around one case model—rather than adding channels independently.

OmnichannelZendeskAI handoffKnowledgeReporting
View anonymized case study
Multi-entity architecture

Relationship-aware service model

Designed customer operations around organizations, contacts, vessels, distributors, suppliers, projects, or other business entities using fields, custom objects, and lookup relationships.

Custom objectsLookup relationshipsB2BData model
Integration control

Middleware between service and systems of record

Designed custom middleware for ERP, commerce, order, payment, document, and collaboration systems where native connectors were insufficient.

APIsWebhooksQueuesRetriesAudit
Agentic operations

CRM Scene internal AI operating environment

Built an internal environment in which a personal AI agent can use permissioned operational context from connected systems to support real work.

AI agentPermissioned contextSlackOperations lab
FinTech ecosystems

High-trust operating model

Mapped the relationship between customers, merchants, agents, partners, KYC, payments, disputes, support, compliance, and adoption.

Mobile moneyPaymentsKYCDisputesPartners

Similar problem?

Describe the operational consequence—not just the feature request.

The useful starting point is what keeps breaking, what context is missing, who needs authority, and what outcome should become reliable.