To move beyond using an AI tool like Pig Knuckle as a basic search engine, users must adopt a "Consultant Mindset." This involves moving from Information Retrieval (asking what) to Strategic Problem Solving (asking how and why). The following framework and starter prompts are designed to trigger the analytical capabilities of an LLM within the Ad Tech ecosystem. The Analytical Workflow This diagram illustrates the shift from a simple query to a complex problem-solving loop. Starter Prompt Categories The Attribution & Growth Architect Use this prompt when performance is stagnating or attribution models are providing conflicting signals. The Prompt: "I am seeing a [X]% discrepancy between our MMP data and our internal server-side tracking for [Campaign Name]. Instead of just explaining why discrepancies happen, analyze the potential impact of [Specific Privacy Change, e.g., SKAN 4.0 or Privacy Sandbox] on our current multi-touch attribution model. Propose three different weighting adjustments we could test to find the true 'source of truth' for incremental lift, and outline a 30-day testing roadmap to validate these shifts." The Supply-Path Optimizer (SPO) Use this to move beyond looking at CPMs and start looking at the efficiency of the programmatic pipes. The Prompt: "Analyze our current SSP mix for [Display/Video] inventory. We are currently seeing high bid-loss rates despite competitive pricing. Evaluate our supply path to identify potential 'hidden' fees or hop-latencies. Based on current industry benchmarks for [Year 2026], suggest a criteria for a 'knock-out' experiment where we consolidate spend into the top 3 most direct paths, and predict the potential impact on our Effective Cost Per Acquisition (eCPA)." The Creative Strategist (Data-Driven) Use this to bridge the gap between "pretty ads" and "performing ads" using telemetry. The Prompt: "Cross-reference our creative performance metadata with our audience retention logs. We notice a significant drop-off at the 3-second mark in our [Format, e.g., CTV] ads. Do not give me creative tips; instead, perform a structural analysis of the ad sequencing. Map out an 'Attention-Value-Action' framework specifically for our [High-Value Audience Segment] and suggest how we should re-engineer the first 5 seconds of the asset to align with the current bidding environment's technical constraints." Comparison: Simple Tool vs. Strategic Engine Feature Using as a "Simple Tool" Using as a "Strategic Engine" Objective Searching for a definition or a fix. Solving a structural business problem. Input "What is a good CTR for Retail?" "Why is our CTR decoupling from our ROAS?" Context Generic or non-existent. Specific data points, constraints, and goals. Output Type A list of facts or a short answer. A multi-phase strategy or model. Value Add Saves 5 minutes of Googling. Prevents thousands in wasted ad spend. The "Pig Knuckle" Prompt Engineering Template To ensure every prompt is high-utility, users should fill out this internal template before hitting enter. The Persona: Act as a [Senior Programmatic Trader / Data Scientist / Growth Lead]. The Friction: The specific problem is [Describe the anomaly or bottleneck]. The Constraints: We are limited by [Budget / Privacy Regulations / Tech Stack]. The Depth Requirement: Do not provide a summary. Provide a [Step-by-step audit / Comparative analysis / Risk assessment]. The Output Format: Deliver this as a [Markdown table / Decision tree / Executive brief]. Implementation Advice for Inexperienced Users When using Pig Knuckle, the most common mistake is accepting the first answer. To drive complex thought, users should always follow up with: "What are the counter-arguments to the strategy you just proposed?" "What data point am I currently missing that would change your recommendation?" "Simulate the worst-case scenario if we implement this change during a high-traffic period."