Analyst Agent

Analyse this A/B test: Control [A METRICS] vs Variant [B METRICS]. Determine: (1) Statistical significance (p-value), (2) Confidence interval, (3) Effect size, (4) Sample size adequacy, (5) Winner recommendation with caveats, (6) Whether to continue testing.

Analyse this data and give me the top [NUMBER — 3 / 5] insights. For each insight explain: what the data shows, why it matters, and what I should do about it. Include a supporting chart or visualisation for the most impactful finding. Write for [AUDIENCE — management / the team / a technical audience].

Compare our metrics against industry benchmarks. Our data: [PASTE KEY METRICS]. Industry: [INDUSTRY]. For each metric: (1) Our value, (2) Industry benchmark, (3) Percentile ranking, (4) Gap analysis, (5) Whether the gap matters and what to do about it. Flag where we significantly lag or lead.

Break down the results in this data by [DIMENSIONS — region, product, customer type, department]. For each segment show: total, average, percentage of whole, and trend direction. Highlight the best and worst performing segments. Create a chart comparing the top segments.

Group users from this data into cohorts by [CRITERIA — sign-up month / first purchase / plan tier]. Track [METRIC — retention / revenue / engagement] over [PERIODS]. Show: cohort table, trend chart, strongest vs weakest cohort, and 3 actionable insights.

Create a visual report from this data for [AUDIENCE — leadership / the board / the team]. Include: (1) An executive summary of key findings in 3 bullets, (2) 2-3 charts showing the most important trends, (3) A summary table with key metrics, (4) Recommendations based on the data. Make the report clear enough to present without additional explanation.

Take these analysis results: [DESCRIBE FINDINGS]. Turn them into a data story for [AUDIENCE]. Include: (1) The headline insight (one sentence), (2) Why it matters (so what), (3) The evidence (charts and numbers), (4) What caused it, (5) What to do about it. Write for people who do not love data.

Combine data from [SOURCES] into executive dashboard. Calculate top-line KPIs with trends, department breakdown, 12-month charts, anomaly detection, next quarter forecast.

Scan this dataset for anomalies and outliers. Look for: (1) Values that are unusually high or low compared to the rest, (2) Sudden changes or spikes in trends, (3) Missing or suspicious data patterns, (4) Inconsistencies between related columns. For each finding, explain why it is unusual and suggest whether it is an error or a genuine anomaly worth investigating.

Based on this historical data, forecast [METRIC] for the next [PERIOD — 3 months / 6 months / quarter]. Show: (1) The predicted values with confidence range, (2) The method and assumptions used, (3) Key factors that could change the forecast, (4) A chart with the historical trend and projected values. State your confidence level.

Analyse this data and identify the biggest drivers behind [METRIC — revenue growth / ticket volume / churn rate]. For each driver: quantify its impact, explain the mechanism, and suggest what we can do about it. Rank drivers by impact size. Create a chart showing the contribution of each driver.

Classify feedback as positive, neutral, or negative. Show: sentiment distribution, positive themes, negative themes, sentiment trend, priority action areas.