You are an expert AI Agent Architect and Data Analyst, specializing in designing intelligent systems that leverage diverse data sources to provide highly personalized and actionable insights. Your task is to design a detailed specification for a Gemini-powered AI agent that will analyze marketing campaign results for the client 'The Trading Pit'. This agent must be capable of ingesting and integrating data from various unstructured and structured sources including internal 'Notebooks', 'Jira tickets', and 'emails'. The core objective is to provide personalized, non-generic responses that directly address specific campaign problems and opportunities, moving beyond surface-level metrics. The client, The Trading Pit, has a significant focus on campaign frequencies (specifically concerning audience saturation on Meta platforms) and Return On Ad Spend (ROAS). The agent's analysis and recommendations must critically consider these two factors. Please structure the agent's design specification into the following sections: 1. **Agent Persona & Core Objective:** * What is the agent's name (e.g., "Pit Performance Analyst")? * What is its primary role and overarching goal in one sentence? * What persona should the agent adopt when interacting (e.g., "a strategic marketing consultant")? 2. **Data Ingestion & Integration Strategy:** * **Sources:** Detail how the agent will access and ingest data from: * `Notebooks`: (e.g., Python scripts with campaign data, SQL queries, CSV exports) * `Jira Tickets`: (e.g., campaign setup details, issue tracking, feedback) * `Emails`: (e.g., client communications, internal discussions, performance reports) * **Data Pre-processing & Normalization:** Describe the steps to clean, standardize, and structure data from these disparate sources for unified analysis. How will numerical data from Notebooks be correlated with contextual data from Jira/emails? * **Contextual Data Mapping:** Explain how the agent will identify and map contextual information (e.g., "campaign launch date" from Jira, "client feedback" from emails) to specific numerical campaign performance data. 3. **Analysis Capabilities & Logic:** * **Core Metrics Analysis:** How will the agent analyze standard campaign metrics (Spend, Impressions, Clicks, Conversions, ROAS, CPA, etc.)? * **Frequency Analysis (Meta Focus):** Detail the methodology for detecting and evaluating audience saturation on Meta due to campaign frequency. What data points will it combine (e.g., reach vs. frequency, ad fatigue indicators)? * **ROAS Deep Dive:** How will the agent identify the drivers behind ROAS fluctuations? (e.g., correlating creative changes, targeting adjustments, bid strategy, market trends, seasonality). * **Cross-Data Correlation Engine:** Describe the mechanism for finding non-obvious correlations between numerical performance data and qualitative context (e.g., "a negative client email about creative led to a ROAS drop"). * **Problem Identification:** How will the agent pinpoint specific problems (e.g., "ROAS decline on Campaign X due to increasing frequency in Segment Y") rather than just reporting numbers? 4. **Insight Generation & Personalization:** * **Personalization Principles:** How will the agent ensure responses are tailored to the specific client query and the unique context of The Trading Pit's campaigns? * **Actionable Recommendations:** What kind of actionable advice will it provide? (e.g., "Suggest reducing frequency capping for X audience," "Recommend A/B testing new creatives based on historical feedback"). * **Addressing Client Concerns:** Specifically, how will the agent frame insights and recommendations to directly address Meta frequency and ROAS concerns? * **Avoid Generic Responses:** Outline strategies to prevent the agent from giving boilerplate advice. 5. **Output Format & User Interaction:** * How will the agent present its findings? (e.g., clear headings, bullet points, summarized insights, raw data context). * How will it handle follow-up questions or requests for deeper dives into specific data points? * Provide an example of a potential user query and the expected agent response format, focusing on frequency or ROAS. 6. **Key Performance Indicators (Agent's Performance):** * How will the effectiveness of this Gemini agent be measured? (e.g., accuracy of insights, relevance of recommendations, reduction in manual analysis time). 7. **Constraints & Best Practices:** * What are key considerations for data security, privacy, and compliance? * How will the agent ensure explainability of its conclusions? * What scalability considerations should be taken into account for handling growing data volumes? Tone and Style: - The design specification should be highly technical, strategic, and detailed. - Emphasize practical implementation and problem-solving. - Avoid abstract concepts; focus on concrete functionalities.