AI for studios

Ship like a studio three times your size.

Small teams can't hire their way out of reporting, triage, and documentation. AI can take that load today. I find where it pays off in your production, wire it into the tools you already use, and build custom assistants for the jobs nothing off the shelf handles.

Book an AI readiness audit Book a free 30-minute call
Where it pays off

The work that eats your week, handled.

Sprint & milestone reporting

Status pulled from Jira, Linear, or Notion and written up for leads and publishers before standup.

Bug triage

New reports deduplicated, tagged, and routed to the right owner with severity suggested.

Meetings to tasks

Notes and decisions turned into tickets, owners, and due dates while the call is still fresh.

Playtest synthesis

Hundreds of survey responses and Discord threads condensed into themes your designers can act on.

Docs that stay current

GDDs, onboarding guides, and pipeline docs drafted and kept in sync with what the team actually does.

Pitch & publisher materials

Milestone reports, pitch decks, and grant applications drafted from your real production data.

See the output

Not a slide about AI. The actual thing.

Three examples of what lands in your inbox. Every sample uses a mock project. No client data appears on this page.

SampleMock project: Hollowfen
Monday Production Brief
Week 14 · Vertical Slice
Slipped
HF-1182Boss arena blockout, moved to Thu
HF-1207Save system QA pass
Blocked
HF-1219Audio middleware license pending
HF-1231Waiting on rig from vendor
At Risk
HF-1240Traversal polish: 3 days of work, 2 left

The Monday Production Brief

One page, in your inbox before standup. What slipped, what's blocked, and what's at risk this week. Every flagged item carries its Jira issue ID, so nobody has to hunt for the source.

SampleMock project: Hollowfen
Milestone Health
Prototype100%
Vertical Slice58%
7
Open risks
+6%
Week over week
Slice progress, 8 weeks

The Milestone Health Dashboard

A read-only view your CEO and EP can read in two minutes. Prototype and Vertical Slice progress, open risk, and trend lines, refreshed weekly from Jira. Built from a dashboard I run on a live studio project tracking about 2,500 issues.

SampleMock project: Hollowfen
AP Agent: access
Status reportingREADWRITE
Bug triageREADWRITE
Meeting notesREADWRITE
Risk logREADWRITE
DependenciesREADWRITE
Build notesREADWRITE
Write access added one action type at a time.

The AP Agent

An assistant producer that covers six workstreams. It starts read-only: it reads, sorts, and drafts. Write access gets added one action type at a time, only after the read-only version proves itself.

Three ways in

Start with an audit. Scale to custom tools.

AI Readiness Audit

A 45-minute call and a one-page write-up. I score your production in five areas, find where time goes, and show where AI will return hours this month. The audit is diagnosis only. Nothing gets built until you ask.

You get
Your score, area by area
The top three fixes, ranked by payoff
What I would build first
A risk and IP review of your current tools
A start plan your team can run
The scorecard

Each area is scored 0 (manual or missing), 1 (partly systemized), or 2 (solid). Total out of 10.

Plan and milestones
Are exit criteria for Prototype and Vertical Slice written and tested? Is the schedule honest?
Tracking and reporting
Does Jira show the truth? Can your CEO read status in two minutes?
Pipeline and asset delivery
Intake, handoffs, vendors, art pipeline. Where does work stall?
Risk and dependencies
Are risks owned and reviewed weekly? Are cross-team dependencies visible?
Team load and cadence
Capacity, meeting load, decision speed.
Live-service add-on: release calendar and patch process.

AI Workflow Integration

I embed with your team and wire AI into the pipeline you already have. Reporting, triage, documentation, and handoffs get faster without anyone learning a new process from scratch.

Integrations with Jira, Notion, Slack, and Drive
The Monday Production Brief and Milestone Health Dashboard, set up on your data
Prompt and template libraries per discipline
Team onboarding sessions
Before and after time tracking
Written runbooks for every workflow

Custom AI Tools & Agents

Purpose-built assistants for your production. An AP agent that knows your schedule. A lore keeper that knows your world. An agent that drafts the weekly build notes. Built on current models, owned by you.

Scoped to one painful job at a time
Read-only first, write access one action at a time
Connected to your studio data
Guardrails for IP and confidentiality
Handoff with docs and source
Iteration as models improve
AI for your dev team

Stop your coding agents from guessing.

AI coding tools guess at engine APIs from old training data. That's where the bugs come from. I set up your team's agents with engine-aware skills tied to the exact engine version you ship on: Godot, Unity, Unreal, or Roblox. The agent loads only the skills a task needs, so it stays fast and on point.

Agent setup for Claude Code, Cursor, and similar tools
Skills matched to your engine and version
A security scan on every third-party skill before it touches your repo
Workflows for art too: AI concepts snapped to real specs before anyone animates them
Why a producer

Built by someone who has shipped.

Most AI vendors know models. Few know what a cert submission week feels like, or why a bug database falls apart at alpha. Fourteen shipped titles at Naughty Dog, Treyarch, and indie teams mean every tool I build starts from how game production actually runs.

The goal is the same as every Production Lab engagement: fewer surprises, faster decisions, and a team that spends its time on the game.

How it works

From first call to tools your team owns.

Discovery call

A free 30 minutes on your team, your tools, and where the week disappears.

Audit

Five areas scored, and the two or three changes with the biggest payoff picked.

Pass bar

Before any pilot starts, we agree on what success looks like and how we'll measure it. It goes in the contract.

Build & embed

Three rungs, and we only climb when the last one works.

Rung 1A manual prototype on the tools you already use. We prove the output is right first.
Rung 2Scheduled runs with logged output.
Rung 3Managed infrastructure, only if you need it.

Handoff

Your team owns the tools, the prompts, and the runbooks. Nothing depends on me staying.

Ground rules

Move fast. Keep control.

Your team stays in charge

AI drafts, sorts, and summarizes. A person approves anything that leaves the building.

Nothing connects without your OK

Written approval before I touch a single tool. Access is read-only and limited to the projects you name.

Your unannounced titles stay unannounced

No detail from your project shows up in any demo, screenshot, or case study. Ever.

Your AI policy wins

I check your rules on third-party model providers before anything gets set up.

Your IP stays yours

Tools are set up on model plans and settings that don't train on your data. I'll show you exactly which.

No fake-clean reports

Every number comes from your tracker. Missing data reads "not returned." If a source fails, you get a failure note, not a report that only looks right.

Measured, not hyped

Every report is checked against your real tracker for two weeks before you rely on it. Every workflow ships with a before and after so you can see what it saved.

Get your hours back.

Tell me where the week goes. I'll tell you what AI can take off the table.