Lunar AI Lab × Chimo - Automate the Work That Doesn't Scale

WHO WE ARE

7 years in tech marketing.
300+ projects. Now building AI.

Lunar AI Lab is the build team inside Lunar Strategy, a European growth marketing agency with 7+ years in tech and crypto. 300+ projects delivered. $60M+ in managed marketing budgets.

We build custom AI agent systems for marketing and revenue teams. We started by automating our own workflows, then brought the same approach to clients. We build for operators, not engineers.

Years in tech & Web3

Projects delivered

Managed budgets

Trusted by: Polkadot · OKX · Polygon · Cardano · DFINITY · BitMEX · Neo

WHERE YOU ARE TODAY

The manual work is the bottleneck.
Every hour spent here is an hour not spent on growth.

Current State

Monetization data scattered across Stripe, admin panel, and mail service. Each test cycle means hours of manual aggregation.
Critical insights like payment failure patterns only surface by accident.
Every new class action triggers the same copy-prompt-paste-into-CMS loop, repeated across 5+ lifecycle checkpoints.
Claim form logic is hand-coded from PDFs. Bugs hit production before they're caught.
A proprietary Google Ads bidding method runs across ~70 companies, 8+ times a month. Undocumented and entirely manual.

What Changes

One agent connects all data sources. Ask a question in plain English, get an answer in seconds.
Anomalies and trends flagged automatically. No more stumbling onto patterns weeks late.
New filings trigger automatic content generation and lifecycle updates. No manual input.
Claim forms get parsed, logic gets built, and QA runs across every branch before anything goes live.
Bidding methodology gets extracted, documented, and replicated. Human approval on every cycle.

THE SOLUTION

Four agents. Each one replaces
a specific manual workflow.

Monetization Analytics

Connects Stripe API + admin panel + mail service
On-demand cohort analysis, revenue projections
Anomaly flagging and trend detection
Natural language queries

2–3 weeks

Claims Content Engine

Monitors for new class action filings
Auto-generates SEO pages, social posts, triggers
Tracks lawsuit lifecycle checkpoints
Pushes to CMS or staging fallback

3–4 weeks

Form Logic + QA

Ingests claim form PDFs
Extracts conditional parameters
Builds user-facing claim flow logic
Automated QA across all branches

3–5 weeks

Ads Bidding Automation

Replicates proprietary methodology
Pulls data, applies logic, generates bids
Human approval gate, graduated autonomy
Full audit trail per cycle

2–3 weeks

R&D Exploration

Synthetic Split Testing

Chimo has too many pricing variables and not enough traffic to A/B test them all. We'll assess whether AI can simulate user behavior to provide directional signals on which offers to test in production. This is research, not a guaranteed deliverable.

Research existing approaches to low-traffic offer optimization.
Assess whether Chimo's user behavior data is sufficient to train a model.
Deliver a written go/no-go report with alternative approaches if no-go.

If feasible, the agent build gets scoped and priced separately in the full proposal.

DELIVERY ROADMAP

Four phases. 8–12 weeks.

Weeks 1–3

Phase 1: Monetization Analytics

Zero dependency on dev team
Stripe API + CSV exports from admin panel
Fastest win, immediate ROI visibility
Builds trust and working rhythm

Weeks 3–7

Phase 2: Claims Content

CMS write API spec delivered to dev team
Content generation works immediately
CMS push added when API is ready
Fallback: outputs to Google Doc / Notion

Weeks 6–11

Phase 3: Form Logic + QA

Highest complexity, builds on Phase 2 CMS
Requires staging environment access
Human approval on all deployments
Automated regression testing

Weeks 4–8

Phase 4: Ads Bidding

Independent from Chimo infrastructure, runs on its own timeline
Starts after knowledge extraction session
Human approval for 3+ cycles minimum
Methodology documented before build begins

INVESTMENT

What it costs. What it saves.

Agent	Est. Hours Saved / Month	What Changes	Scales With
Monetization Analytics	3–5 hours	Manual aggregation → on-demand queries	Test frequency
Claims Content Engine	5–10 hours	Copy-paste workflow → automated generation	Active claims onboarded
Claims Form Logic + QA	3–6 hours	Manual coding + bugs → auto-built logic with QA	Claims volume
Google Ads Bidding	4–6 hours	30 min/account × 8+ sessions → automated	Client count
Total	15–27 hours/month

Based on the workflows you described on our call. Actual savings scale with volume.

$15,000 – $20,000

Estimated total for all four agents. Final pricing confirmed in the full proposal after the workshop.

Add-ons

$1,000

Synthetic Testing: Feasibility Assessment

Written go/no-go report with alternative recommendations

Research existing approaches to synthetic A/B testing and low-traffic optimization.
Evaluate Chimo's existing user behavior data for model training viability.
Consult with technical lead on architecture feasibility.
Deliver written feasibility report with clear recommendation.
If no-go: provide alternative approaches (e.g., Bayesian multi-armed bandit testing).

PRICED AFTER FEASIBILITY ASSESSMENT

Synthetic Testing: Agent Build

Scoped and quoted only if the feasibility assessment confirms viability

Train agent on representative user behavior data.
Build simulation pipeline for offer variant ranking.
Integrate with Chimo's existing test infrastructure.
Deliver directional signal reports per test cycle.
Ongoing calibration as real user data validates predictions.

Terms

Ownership. All agents, code, configuration, and documentation belong to Evan upon deployment. We retain no access post-handoff unless ongoing support is agreed.

Infrastructure during development. All infrastructure costs (AI API, hosting, tooling) are covered by Lunar AI Lab during the build. Upon deployment, we hand over everything to Evan's accounts with full setup assistance and usage estimates.

Running costs post-deployment. Once agents are live, API token costs and hosting sit on Evan's accounts. We provide detailed estimates per agent so there are no surprises.

NEXT STEPS

How to get started.

Sign NDA

Both parties. Target: Monday, March 30.

Workshop: walk through all four use cases

Wednesday, April 1, 11:00 Lisbon. We'll send prep materials and requirements after the NDA is signed.

Full proposal and confirmed scope from Lunar AI Lab

Detailed architecture, final pricing, and delivery timeline based on workshop findings.

Review, confirm, and kick off the build

Align on scope and pricing, then we start.

Automate the Work That Doesn't Scale.

7 years in tech marketing.300+ projects. Now building AI.

The manual work is the bottleneck.Every hour spent here is an hour not spent on growth.