AI
How our agency uses AI in each phase of the 3-month method
Month 1 analysis, month 2 implementation, month 3 optimization. Where AI comes in, what stays a human decision, and which tools we use in each phase.
Every agency today claims to use AI. The question that changes the outcome is: where, exactly. This piece shows, phase by phase, how our agency applies AI inside the MADS 3-month method, which tools enter at each moment, and what remains a human decision. More than 571 companies have gone through this flow.
Premise: AI accelerates process, does not replace method
The MADS method existed before AI. AI came in to accelerate each phase, not to rewrite the script. If the sales process is unclear, AI becomes expensive waste with a fancy name. When it is clear, AI multiplies the team.
Month 1. Strategic analysis with AI
The first month is diagnosis and structuring. AI shows up in four places:
- Reading the account and the competition. Claude cross-references campaign history, past creatives and competitor moves. Out comes a map of what has been tried and what makes sense to test.
- Setting the 70/30 strategy. Based on the business, AI proposes the initial split between prospecting and conversion. The human consultant adjusts for context (seasonality, niche constraints, real budget).
- Creative hypotheses. AI suggests angles, benefits, audiences. The human creative selects and produces.
- Account structuring. The initial Google Ads and Meta Ads structure is drafted by AI from the brief. The media manager finalizes.
Human decision in month 1: strategic priority, positioning, real budget allocation.
Month 2. Smart implementation
The second month is execution with process. AI enters to scale:
- WhatsApp agents. Triage, first response and qualification deployed following AI in WhatsApp support. Reception starts receiving warm leads.
- Creative routine. Reels scripts, ad copies and captions produced in queue. The human team edits, approves and publishes.
- Structured follow-up. Daily queue of messages for warm leads and pending patients. Reception only approves.
- Living documentation. Every operational decision becomes a file in Claude Cowork. Context compounds, agents get better.
Human decision in month 2: creative approval, tone adjustments, handling exceptions in support, managing the client's internal team.
Month 3. Continuous optimization with AI
The third month is scale and maturity. AI becomes a daily instrument:
- Daily account analysis. AI compares variations, flags anomalies, suggests pausing a tired ad, suggests raising a rising creative. The manager decides.
- Automatic reports. The monthly report flow runs with senior review.
- Structured tests. Every two weeks, new creative hypotheses enter rotation with a defined read plan. AI helps list hypotheses, the team prioritizes.
- Scaling what works. AI spots the pattern of a winner and suggests replicating it across new audiences and channels. The consultant decides the pace of the ramp-up.
Human decision in month 3: when to scale, when to cut, when to renegotiate budget, what becomes a fixed program.
What is AI, what is human in each phase
| Phase | AI role | Human role |
|---|---|---|
| Month 1 (Analysis) | Data reading, hypotheses, draft structure | Strategy, positioning, investment decision |
| Month 2 (Implementation) | Support at scale, creative in queue, documentation | Approval, tone, exceptions, people management |
| Month 3 (Optimization) | Alerts, suggestions, reports, scale patterns | Scale, cut and renegotiation decisions |
Tools we use in each phase
- Claude: base model for analysis, consultative writing and review. Present across all three months.
- Claude Code: automation by conversation, especially in account structuring in month 1 and reports in month 3.
- Claude Cowork: per-client folders with living context. Running from day 1.
- MCP: connectors between Claude and spreadsheets, CRM, WhatsApp, Google Ads, Meta Ads. Backbone from month 2 onward.
- Complementary tools: ChatGPT and Gemini for research and brainstorming, per the AI tools hub.
What AI did not change in the method
- The client still needs real minimum media investment. AI does not generate leads without budget.
- Reception still needs training. AI does not sell on its own.
- 70/30 is still the base. AI accelerates, does not replace the principle.
- The human consultant still decides what matters. AI proposes, human disposes.
Next steps
How we implement AI with clients: AI implementation for companies. To run this in your operation, start with the free diagnosis. Related: real AI use cases. Traffic is the means, the sale is the end, and AI serves both.
Ready to apply this in your business?
Free diagnosis before any proposal.
8 quick questions and an initial action plan tailored to your case.
Book my free diagnosis
About the author
Mattias Custodio
CEO and founder of Mads Acelerador. Over 9 years running paid traffic, Google Partner since 2019. Personally leads the MADS methodology that accelerated more than 571 companies in Brazil, with Claude as the standard across the AI layer.
Keep reading