Ceballos/Works

We are an AI-native consulting firm for product, program, and operations.

The critical-thought layer is human.
Fifteen specialized agents handle delivery.

What we are

A consulting firm whose entire roster is AI agents.

CeballosWorks runs as a roster of fifteen specialized AI agents. Each agent has an explicit identity, decide-vs-escalate boundaries, a behavioral contract, scoped tool and MCP access, and stated limitations. They handle research, diagnosis, planning, financial modeling, change management, contracts, business development, product strategy, and engineering.

Quality gates are human. The critical-thought layer is human. The agents own throughput.

An AI agent drafts every artifact the firm produces. A human decides and edits anything that carries judgment, recommendation, or commitment.

How we operate

Five operating commitments.

  1. Human judgment where judgment matters.

    The critical-thought layer is human. Agents draft and execute. Product judgment, architectural decisions, prioritization, escalation, and any commitment to a client clear human review. Routine outputs ship on agent confidence with reviewable artifacts.

  2. Every agent ships with a written behavioral contract.

    Identity, decide-vs-escalate boundaries, tool and MCP access scopes, communication style, stated limitations. No agent runs without one.

  3. Default to the agent.

    When work arrives, the question is which agent. New work without an agent fit is a signal to design one.

  4. Artifacts are the source of truth.

    Every agent reads its own profile plus a shared firm context document at session start. No implicit knowledge. Outputs are reviewable.

  5. Automation includes the firm itself.

    The same standard applies to BD, intake, billing, and internal ops. Client delivery is one workstream of many.

The roster

Fifteen specialized agents. Each with its own written persona.

How an engagement runs

Five phases. Human judgment at every gate.

  1. Pre-Engagement Weeks 1 to 2

    Industry briefing, intake, contracts.

  2. Discovery Weeks 2 to 4

    Diagnostic, interviews, current-state, financial baseline, AI readiness.

  3. Strategy/Design Weeks 4 to 6

    Program plan, KPI framework, change strategy, tech recommendations, build specs.

  4. Execution Weeks 6 to 24+

    Program management, modeling, change comms, build and deploy.

  5. Closeout Final 2 weeks

    Deliverable packaging, knowledge transfer, reusable asset capture.

Principal

Rodrigo Ceballos

Founder & Principal

Senior Product Leader with 10+ years shipping AI-enabled commerce at scale. At Amazon, defined and shipped GenAI content workflows delivering $512M annualized for premium selection during Black Friday and Cyber Monday; led the 0→1 Nike and Converse launch across 27 customer touchpoints, hitting $130.4M one month ahead of schedule; lifted cosmetics conversion +1.32% via a multi-skin-tone Computer Vision program.

Engineer by training (ITAM, Telecommunications & Informatics). Executive MBA at UCLA Anderson, Class of 2028.

Based in Los Angeles · linkedin.com/in/rodrigoceballos

Contact

Currently scoping select engagements.

[email protected]

Best fit is small and mid-market businesses. AI-native delivery makes deeper diagnostics, faster turnarounds, and more iterations economical at this scale.