John Price
Managing Director
AMI
As artificial intelligence (AI) transforms industries with its ability to process massive datasets, detect patterns, and automate tasks, businesses are increasingly tempted to use AI for high-stakes strategic functions—especially in market research, competitive intelligence (CI), and political-economic risk analysis – the three pillars of market intelligence. On the surface, it appears logical: AI promises speed, scale, and objectivity. But in emerging markets—where data is patchy, institutions are informal, and change is nonlinear—this promise breaks down.
Strategic success in such environments requires more than processing power. It requires context, cultural fluency, trust-building, and creative judgment. This article outlines why AI is no substitute for human-led market intelligence in emerging markets and related strategic domains. It also makes the case for a hybrid approach that blends the best of AI with irreplaceable human insight.
How to Size Markets in Latin America
A practical, research-based approach to accurately size markets in Latin America, including TAM, SAM, SOM, and local insights.
1. Data Scarcity and Structural Informality
AI thrives on structured, high-quality, and voluminous data—none of which are guaranteed in emerging markets like Latin America. Government statistics may be outdated, inconsistent, or politically manipulated. Some estimate that the amount of public market related data published by even the most transparent democracies in Latin America is less than 5% of that found in advanced OECD countries. In many sectors, the cash-based informal portion of the sector may be larger than the formal component so any published data is grossly under-representing the full market; examples include retail, food service, agriculture, and transportation. Data pipelines—where they exist—are fragmented and lack reliability.
Even private sector sources in emerging markets are extremely limited. In markets like Western Europe, industry associations act as trusted information brokers working with their industry membership to collect and amalgamate sensitive company data and additionally invest in sector wide independent research. Their publications are substantial, insightful and transparent. By contrast, industry associations in Latin America, with few exceptions, fail to instill enough trust in their members to convince them to share sensitive information. Their industry reports vary in quality but almost never match the depth and rigor of their counterparts in Western Europe.
AI systems, which require clean, comprehensive inputs, flounder in the data-sparse conditions of emerging markets. In contrast, human researchers can:
- Conduct fieldwork and in-person interviews,
- Use proxies and triangulation to estimate market conditions,
- Interpret ambiguous or conflicting data,
- Build mental models of informal or opaque systems.
Experienced, human analysis in emerging markets is essential where insight must be constructed, not merely extracted.
2. Blindness to Informal Power Structures
AI operates within the limits of observable, recorded data. But in many emerging economies, true decision-making power lies outside formal institutions and their rules. Regulatory enforcement may depend on local patronage. Business permits may require negotiation with municipal officials, illicit leaders, or party loyalists. Corruption, favoritism, or informal fees often go unrecorded.
These nuances are invisible to AI that reads data at face value. Algorithms cannot navigate opaque hierarchies, decode clientelism, or detect behind-the-scenes negotiations. Human intelligence professionals—particularly those with local networks—can map these dynamics and reveal who actually holds power, how influence operates, and what it takes to get things done.
For industries operating in remote locations (extractive sector) or those that are highly regulated and thus exposed to political interference, or those that compete with state-owned enterprises, this dynamic is particularly relevant.
3. Inability to Interpret Context, Intent, and Emotion
AI can analyze behavior but cannot grasp motivation. A competitor opening a new R&D center might be pursuing genuine innovation—or chasing tax breaks. A regulatory reform could signify liberalization—or a veiled power grab.
Natural language processing (NLP) tools misread sarcasm, slang, and culturally loaded expressions. AI might identify increased online discussion of “green products” as evidence of eco-conscious consumption—missing that these mentions may be aspirational, not actionable, in low-income communities. In political environments like those found in much of Latin America, public political discourse and written pronouncements are more often than not miles from the truth and how events will transpire.
Humans interpret tone, body language, and unspoken subtext. They understand symbolic gestures, emotional triggers, and cultural taboos—elements critical to understanding how markets behave, why policies shift, and what consumers really want.
In markets like Latin America, published sources vary dramatically in terms of trustworthiness and even within one source, say a newspaper, some journalists will maintain high standards while others will bend the truth in return for financial gain from an external party or to abide to the threat of harm. AI is not equipped to measure the integrity of one written source versus another.
4. Failure to Detect Weak Signals and Emergent Trends
Disruption often begins as a whisper. Before trends become data, they emerge as anomalies: a subtle shift in language, an unusual consumer workaround, a string of quiet executive exits.
AI needs patterns to detect significance. But early indicators—such as discontent with a local product, whispered rumors of regulatory pressure, or newfound enthusiasm for a niche lifestyle trend—often lack volume or structure.
Latin American economies are notoriously volatile. As market conditions abruptly change, the patterns detected over the last 1-3 years have no bearing on the year ahead. Economic volatility in turn influences the demand and competitive dynamics of all industries as well as consumer behavior, purchase trade-offs, etc.
Human researchers, especially those embedded in local markets, identify these signals of change through:
- Observation and ethnography,
- Informal interviews and insider conversations,
- Experience-based inference and intuition.
Such early warnings are invaluable in competitive intelligence and risk analysis as well as forecasting—areas where being first is often more important than being precise.
5. Bias and Blind Spots in AI Models
AI reflects the data it’s fed—and inherits its flaws. If ‘training’ data disproportionately represents only a slice of the population (urban, middle & upper class, formally employed), then AI will mirror and magnify those biases.
This skews market research, underrepresents rural or economically marginalized groups, and overlooks key segments. Similarly, Latin America’s robust informal & illicit competitors may operate entirely outside of the digital signals that AI tools monitor.
Political risk models may falsely indicate stability in authoritarian states simply because censored media and rigged data sets show a untruthful utopia. The Kirchner-led administrations of past in Argentina famously published fantastical inflation data to keep unions and workers wage expectations at bay. The Bukele administration has worked hard to sanitize the media’s coverage of its judiciary actions. The Arce administration in Bolivia stopped reporting its foreign reserves levels in early 2024 when they grew dangerously low. AI fails to detect such political interference in published data.
Meanwhile, human analysts can detect hidden cracks through cross-referenced reporting, whistleblower networks, or off-the-record briefings.
6. Lack of Hypothesis-Driven Thinking
AI systems are query-based. They answer what they are asked—but do not ask new questions. Strategic disciplines are inherently hypothesis-driven. Analysts must pose open-ended inquiries like:
- “Why is our competitor altering its product line?”
- “Could this policy shift signal future trade barriers?”
- “What unmet emotional needs are driving these new behaviors?”
These questions evolve as more information emerges. Trained human researchers adjust lines of inquiry, test theories, and refine their models iteratively—something algorithms cannot do without constant human steering.
7. Inability to Model Human Behavior and Power Dynamics
Strategy is shaped by human behavior: ambition, fear, trust, revenge. These emotional and political undercurrents drive decisions—from boardrooms to parliaments.
AI cannot predict whether a judge will be influenced by political pressure – which we can expect more of going forward in Mexico, for instance; whether a CEO will panic-sell assets, or whether a strike will be called off following a midnight phone call from a party leader. Nor can it detect when firms deliberately manipulate data – as they might do when under pressure from organized crime with whom they are colluding, pursue symbolic partnerships, or engage in reputational feints.
Humans excel at reading between the lines—assessing credibility, discerning hidden motives, and contextualizing decisions. This interpretive capacity is indispensable.
9 Blunders Made with Strategic Planning in Latin America
Where companies go wrong, why this matters and how they adjust these mistakes to maximize success.
8. Missing the Role of Trust, Networks, and Reputation
In many emerging markets, valuable intelligence is not online—it’s relational. Information flows through networks, not dashboards. Access depends on trust, credibility, and reputation.
Human researchers:
- Build rapport with local leaders and stakeholders
- Gain access to insider perspectives at conferences or field visits
- Conduct candid interviews with regulators, vendors, or consumers
- Know which “facts” are political theater and which are real
AI cannot replicate relationship-based intelligence gathering. It cannot sense when a source is hesitant, angry, or hiding something. It cannot ask a sensitive follow-up question or navigate interethnic, class, or religious dynamics during an interview.
9. Ethics, Privacy, and Reputational Risk
AI-based methods—especially those that scrape digital data—can raise ethical and legal red flags. In regions where privacy norms vary or where surveillance evokes authoritarian associations, misuse of AI can damage a brand’s reputation.
Human-led intelligence gathering, when done ethically, is more transparent and respectful. It allows companies to operate in accordance with local laws and cultural expectations. It also builds goodwill and fosters deeper understanding.
10. Strategic Misjudgments from Over-Reliance on AI
Trusting AI too much can lead to costly errors:
- Launching a product based on inaccurate demand indicators
- Entering a market just before a political or currency crisis
- Missing an insurgent or informal competitor because it flew under AI’s radar
- Misjudging reputational risk due to a lack of documentation or undetected public sentiment shifts
AI may spot “what” is happening but not “why”—and certainly not “what’s next.” In complex, high-risk environments, misinterpretation can be fatal.
By contrast, human analysts add perspective, challenge assumptions, and incorporate soft variables—delivering nuanced conclusions that algorithms alone cannot.
11. AI as an Enabler, Not a Replacement
This is not to say AI is useless. On the contrary, AI can be an immensely powerful tool when deployed strategically:
- In market research, AI can process large-scale surveys, monitor sentiment, and identify behavioral trends from professionally gathered data
- In CI, it can track competitor filings, flag sudden changes, and automate media monitoring
- In risk analysis, it can aggregate macroeconomic indicators, monitor social media for unrest signals, and visualize cross-border policy shifts
Used well, AI boosts human productivity. It saves time, enables scalability, and surfaces patterns that might otherwise go unnoticed. But the key insight still comes from people—from their ability to connect the dots, imagine alternatives, and anticipate the unexpected.
In the End, Insight Is Human
Artificial intelligence is revolutionizing the mechanics of analysis—but not its meaning. The essence of market intelligence lies not in data alone, but in the ability to interpret, contextualize, and act upon it. In emerging markets like Latin America—where ambiguity, informality, and volatility are the norm—this ability is uniquely human.
Strategic success comes not from replacing analysts with algorithms, but from combining both in a hybrid model. Let AI handle the noise, scale the signal, and flag anomalies. Let humans test hypotheses, explore context, and generate insight.
Organizations that understand this balance—who invest in human capital even as they adopt advanced tools—will not only avoid blind spots but will unlock deeper opportunity, foresight, and competitive advantage in a complex and fast-changing world.
Next Steps
Contact us when your company is trying to answer strategic business questions about Latin America, such as how to size a market, researching an export market in Latin America, the competitive landscape in a particular market, understanding risk in a specific market or industry in the region, and much more. While AI can help in certain areas, a custom study built to Latin America’s information flow and idiosyncrasies will provide the answers your company needs to guide its investments in the region.
