Gen AI and The AEC Industry:
spotlight in latin america

Report Summary

‍

The global report places SMEs at around 90% of the AEC sector, and considers this composition a defining characteristic rather than a deficiency. Mexico is no exception to this description; it intensifies it, and this intensification modifies how

the adoption of these practices manifests itself.

‍

Mexico has 8,850 registered architecture, engineering, and related services companies; 79% employ ten people or fewer, and 61.7% of the sector’s 140,000 workers are informal, meaning they lack social security registration. Informality in construction is 81.5%, compared to 54.8% for the economy as a whole. In all sectors, 95.3% of the country’s 5.47 million business units are micro-enterprises; they employ 41.4% of the workforce and generate 16.1% of the added value.

‍

‍

AI is Used More Freely

But not adopted more quickly

‍

A preliminary version of this document argued that Mexico would adopt AI more quickly than CAD or BIM. The framework of the global report demonstrates why this assertion needs to be reformulated, and this reformulation forms the core of the argument.

‍

Generative AI has a lower barrier to entry than any previous AEC technology: a browser and a few minutes, no license, no training, no permissions. This explains the individual figures. The 20.1% of people aged 15 to 64 in Mexico who use a generative AI product is on par with Colombia (24.5%), Chile (22.7%), and Argentina (21.9%), and ahead of Brazil (19.1%). This measure is derived from aggregated device telemetry and records usage, not value. It is a regional pattern driven by accessibility, not an achievement unique to Mexico.

‍

Shadow IT

Without the functions that contain it

‍

The global report notes that with each technological wave, users discover free and accessible tools before understanding the underlying business models or the implications of sharing project information, and that larger organizations have cybersecurity, privacy, legal, and data governance functions that smaller companies lack. The evidence that this is happening with generative AI within large organizations is compelling. A 2024 benchmark study found that 27% of large companies had banned generative AI, 63% restricted the data that could be input, and 61% restricted the tools that could be used, citing intellectual property (69%) and disclosure to competitors (68%).

‍

Specific Risks of Mexico

Risks and challenges

‍

The five risk categories in the global report are applicable. Two of them are more pressing in Mexico because the exposure lies with a client and a signatory, rather than a company.

• Customer Data
• Liability and Insurance

Who Formalizes The Market?

• The Capital
• The State
• Real Estate Investment Trusts (REITs)
• Europe as a benchmark
• Regional standard

Conclusion and Perspectives

‍

The divide within the Mexican architecture,engineering, and construction (AEC) sector is evident. The informal majoritywill continue to generate short-term productivity thanks to readily availabletools, while accumulating risks they cannot perceive. The minority thatconducts qualified reviews and manages client information commercially will bethe ones prepared to work with institutional capital. Both use the same tools;the difference lies in whether a qualified professional is visibly involved inthe process, whether client information is stored where agreed upon, andwhether any of these facts can be proven.

‍