Live-ops intelligence
Offer targeting, event cadence and economy tuning, automated on Alexandria and wired into your live-ops data pipeline.
Your live-ops calendar is fuller than the team can staff. The AI tools you bought to help mostly sit unused. We've been in that seat. Running a live-ops hit is the job we left to start this firm, and enterprise AI needs the same discipline. Your studio already has it. What's missing is knowing where to point it.
Patrick McGrath has led products at PopCap (EA), NaturalMotion (Zynga) and Wooga (Playtika), including Plants vs. Zombies 2, CSR Racing and June's Journey. The products he's shipped have earned more than $4 billion in lifetime revenue. Separately, he built a Berlin studio from zero for Ten Square Games. Nobody has to explain a live-ops build meeting to us.
We start with the product question: where players leave, or where the offer model leaks. Then we automate the workflows around the answer. In a gaming studio that work comes in four kinds, and we've run all four as operators: live-ops intelligence, content pipelines, churn and LTV forecasting, and UA optimization.
Live-ops tooling we built for a top-20 studio
- Design
- Approve
- Version
- Promote
- Target
- Measure
Where we help
Four workstreams we keep seeing
Content pipelines
Asset generation, localization and narrative iteration, with clear rules on what the AI can ship alone and what a human checks first.
Churn prediction and LTV forecasting
First find the moment players leave, then build the models that watch it: churn, LTV, retention and segmentation. We report honestly on how much of the result the models can take credit for.
UA optimization and creative testing
Ad creative testing, creative fatigue models and paid-channel decisions. Your UA team measures in hours, so the tooling does too.
Proof of work
The first of those four, in production
A top-20 mobile games studio designs its live-ops events with tooling we built.
We built their live-ops tooling as a platform. It simulates every event against real player baselines, values it on net lifetime value, checks it against the studio's hard economy rules, and compiles it to their game config. A ship layer versions that config, targets player segments and runs the A/B test.
It was backtested against shipped events and real history before it touched the live game, and it has 490 automated tests. It runs on their keys, in their stack. Their team runs it without us.
See how the platform is built. Reference on request.
What AI changed in a live-ops team, and what it didn't: Live ops after AI.
Build or buy
We'll tell you when to buy a platform instead
Live-ops platforms publish their prices and ship fast. If one fits your game and your stack, buy it. We'll say so on the first call, and it costs us the build.
Their AI works inside their own backend. Most studios run on something else: an in-house server, tools bought over the years, and data spread between a warehouse, an attribution partner and the design team's spreadsheets.
That's where we build. The live-ops tooling and pipelines above run on the stack you already have, so nothing has to migrate before something ships.
The longer argument, with this month's platform prices and the numbers from our builds: Live-ops AI: buy the plumbing, build the judgment.
Fit
When our team is the right call
- Live-service mobile, PC or web games with real DAU and transaction volume
- Studios whose retention or monetization has stalled and want an operator's read before buying any AI
- Studios that want automation built around how their live-ops team already works
- Publishers pulling AI together across a portfolio who need a platform decision they can defend
- Studios working on interactive narrative or generative systems, where editorial calls still matter
If none of these sound like you, we'll say so on the first call and keep it short.
In practice
Structures that have fit here
This is where past engagements started. It isn't a menu. We shape each one around your problem.
Diagnose
Product & Growth Audit
A month on retention, liveops and monetization for a live title. You get build-ready specs and a 90-day roadmap.
Build
AI Build Program
We automate workflows inside your production pipeline and keep running them with you after go-live.
Bet
Outcome-Based Build
When the target is ARPDAU, retention or LTV and your measurement is clean, we'll take the fee out of the uplift.
Next step
A 30-minute call with Patrick
You'll be talking to someone who has sat in your live-ops build meeting. Bring the KPI that's stuck.
Further reading
Churn prediction is the easy part
Most players decide to leave in their first sitting, so a churn score that fires after a week of silence reports a decision already made. Three games this year showed us how to find that moment, and how to prove a fix worked.
Live-ops AI: buy the plumbing, build the judgment
Every live-ops platform now sells AI, and the prices are public. Buy the part that ships config. Build the part that decides what to ship, because it runs on your game's own history.
Live ops after AI
The AI that paid off in live ops sat under the calendar, in churn and LTV forecasting, where it puts a decision in front of a person days sooner. The talking NPCs and self-filling calendars vendors pitch first are the wrong place to start.