Aaron Levie — “The Agent Deployer” Role JD (Apr 14, 2026)

Why this is in the vault

Levie (CEO of Box) wrote a 341K-impression post describing a new enterprise role — the “agent deployer and manager” — and listed the exact skills and responsibilities. This is a tweet-length validation of RDCO’s consulting positioning. Levie is describing, almost verbatim, what the founder does. Worth filing as a reference for the MAC content series and any consulting proposal we draft.

The JD verbatim (key passages)

Full post at https://x.com/levie/status/2043883641366032638

The role:

“The more enterprises I talk to about AI agent transformation, the more it’s clear that there is going to be a new type of role in most enterprises going forward. The job is to be the agent deployer and manager in teams.”

Responsibilities:

Figure out highest-leverage workflows where agents drive 100X speed or volume
Map structured and unstructured data flows
Figure out the ideal workflow
Get the agent the right context
Figure out where humans interface with the agent and at what steps
Manage evals and reviews after any major model or data change
Run and manage agents on ongoing basis, tracking KPIs

Required skills:

Strong at process mapping, understanding where value could be unlocked
Relatively technical
Comfortable with skills, MCP, CLIs
Good operationally and at business
Full autonomy to connect business systems and drive automation

Org structure:

“There will likely need to be one or more of these people on every team”
Not a centralized role per se
May roll up to IT or an AI team, or live in the function with checkpoints to a central team
“May be an existing person repositioned, or a totally net new person”

Target candidates:

“Fantastic job for next gen hires who are leaning into AI, and are technical”
Future of the engineering career path

The supporting post (2026-04-12)

Levie’s separate thread from 2 days prior (5,101 likes) adds context:

“Despite Silicon Valley’s sense that AI has made hard things easy, the most powerful ways to use agents is more ‘technical’ than prior eras of software. Skills, MCP, CLIs, etc. may be simple concepts for tech, but in the real world these are all esoteric concepts that will require technical people to help bring to life in the enterprise.”

And:

“Engineers may not be ‘writing’ software, but they will certainly be the ones to setup and operate the systems that actually automate most work in the enterprise.”

Mapping against Ray Data Co

The founder IS the agent deployer role — he’s done every single bullet in the last 72 hours:

Identified highest-leverage workflows (newsletter ingestion, data quality testing, board-driven orchestration)
Mapped data flows (vault structure, silver/gold layer architecture discussion yesterday)
Built agents with skills, MCP, CLIs (22 skills, PostGrid API, Gmail MCP, Notion MCP)
Manages evals (MAC framework — the evaluation-layer thesis)
Tracks KPIs (Notion board, self-review scoring)

The two consulting postures this unlocks:

Posture 1: Embedded Fractional Agent Deployer

Client does NOT hire; RDCO fills the role as a service
Retainer-based, long-term engagement
RDCO operates their agent stack on their behalf
Pros: high-value per client, deep state ownership (matches state-moat thesis)
Cons: scales poorly (N clients = N fractional Bens), bus-factor risk (everything depends on founder)

Posture 2: Playbook + Coaching Model

Client hires or repositions their own “next-gen technical hire” for the role
RDCO provides the playbook (MAC framework + state-ownership architecture + audit-model skill + onboarding curriculum)
RDCO coaches the new hire for 3-6 months
Pros: scales better, matches the drip course/lead magnet thesis (Cole playbook), leaves the client with owned state
Cons: lower revenue per client, requires good curriculum

Posture 2 aligns with everything else we’ve built:

Cole $100K paid newsletter playbook → MAC drip course trains the next-gen hire
State-ownership architecture → client owns their vault + skills + state
40RTY Shopify lead-magnet pattern → open-source skills qualify leads
Sanity Check newsletter → top-of-funnel for coaching/consulting engagements

Hybrid: start with Posture 1 to prove the model, then productize into Posture 2

Early engagements: do the work yourself, demonstrate outcomes
Document the playbook in real-time
Transition to Posture 2 once the MAC drip course + coaching package is proven

2026-04-13-moura-entangled-software-agent-harnesses-dead — Moura’s entangled-software thesis; the agent deployer is the person who builds the entanglement
2026-04-13-jaya-gupta-ai-lock-in-state-moat — Gupta’s state layers; the agent deployer manages all four
2026-04-13-cole-100k-paid-newsletter-playbook — the content-business playbook that funnels into the coaching engagement
../04-tooling/rdco-state-ownership-architecture — the architectural framework the agent deployer would implement
competitive-review-40rty-shopify-skills-2026-04-13 — lead-magnet pattern (skills as qualifier)
commentary-tan-fat-skills-thin-harness-2026-04-14 — the three-layer architecture the agent deployer enforces
2026-04-23-technically-when-not-to-vibe-code — Gage’s blast-radius/ownership gate; the entry-criteria rule for what an agent deployer ships vs. defers
2026-04-26-alphasignal-deepseek-v4-kimi-k26-agentic-ai — open-weight tier-2 model substrate that makes the agent-deployer role economically defensible (DeepSeek-v4, Kimi-K2.6, Qwen3.6 as swappable underpinnings)