Mastering the craft of communication with artificial intelligence
Prompt engineers who work in finance and business intelligence quickly discover that macroeconomic context is not just background noise — it is essential domain knowledge for generating reliable analysis. A model asked to assess an equity valuation without understanding the interest-rate environment will produce plausible-sounding but misleading output. Understanding the core macro signals, and how they interact, makes your prompts sharper and your evaluations of AI-generated analysis more rigorous.
Begin with the bond market, which often prices economic expectations before equities do. An inverted yield curve — when short-term Treasury yields exceed long-term ones — is the most closely watched recession indicator in professional finance. The inversion signal works because it reflects the collective judgement of bond traders: they are willing to lock in lower long-term rates because they expect the economy to slow and the Federal Reserve to eventually cut short-term rates. Nearly every U.S. recession since 1970 was preceded by an inversion, typically 12 to 18 months earlier. When prompting AI systems to analyze market conditions, including yield-curve status as context dramatically improves the relevance of the output.
Labour market data provides a complementary signal from the real economy. The headline unemployment rate has a well-known limitation: it only counts people actively seeking work. The labor force participation rate — the share of the population that is either employed or looking for work — provides a fuller picture. After the 2008 financial crisis, participation fell structurally as discouraged workers left the labour force; the unemployment rate looked healthier than the underlying reality. Prompting AI to analyze labour conditions using both metrics, not just the headline rate, produces substantially more nuanced conclusions.
Participation interacts directly with how fast workers expect pay to rise. When the labour market is genuinely tight — not just low unemployment but also high participation — workers gain bargaining power and wage-growth expectations rise. Rising wage expectations matter for monetary policy because they can create self-fulfilling inflation: workers who expect their pay to rise demand higher wages, companies pass costs through into prices, and the expectations become reality. The Fed watches wage-growth surveys closely as a forward indicator of whether inflation is likely to accelerate or decelerate.
The sustainable basis for wage increases without inflation is rising labor productivity — the output produced per worker hour. When productivity grows, companies can afford to pay workers more without raising prices. The AI-driven productivity story is potentially significant here: if large language models and automation genuinely raise output per hour across knowledge work, the economy could sustain higher real wages without the inflationary pressure that would otherwise prompt the Fed to tighten. For AI practitioners thinking about the macro context of their own industry, productivity statistics are one of the few data series that directly capture whether AI's economic promise is materializing.
All of these variables ultimately interact with the M2 money supply — the broadest commonly cited measure of money in circulation, including deposits and money-market funds. Rapid M2 expansion, as occurred during 2020-2021, floods the economy with purchasing power and tends to produce asset-price inflation followed by consumer inflation. The subsequent contraction of M2 growth — as the Fed raised rates aggressively — tightened credit conditions and repriced long-duration assets downward. For AI models generating financial analysis, M2 trends provide essential context for understanding why markets moved the way they did in any given period.
The craft payoff here is real: when you include yield-curve status, participation rates, wage expectations, productivity readings, and M2 direction in a financial-analysis prompt, you give the model the context it needs to produce analysis that a professional economist would recognize as properly grounded. Macro signals are not a separate domain from AI; they are the contextual layer that separates credible AI-assisted analysis from sophisticated-sounding noise.