Web Summit Rio 2026 featured a new solution designed to monitor content generated by generative artificial intelligence. Selected as an ALPHA Startup, Teia Studio presented its Teia GEO IGO platform on June 9, built to audit real-time information provided by models like ChatGPT, Gemini, Claude, and Perplexity regarding brands, institutions, and services. During the event, the company also revealed results from a pilot program that uncovered hallucinations and significant discrepancies among competing AI systems.
Web Summit Rio connects founders from Brazil and Latin America with international investors, journalists, and tech leaders. The event is part of the global Web Summit network, which includes flagship conferences in Lisbon, Doha, Vancouver, and RISE in Asia.
At the 2026 edition in Rio de Janeiro, Teia Studio was chosen for the ALPHA Startup Program. The firm exhibited at Stand A3-30 on June 9 and competed in the Web Summit PITCH Competition.
Platform Audits Responses Across Four Major AI Engines
Teia GEO IGO addresses the growing reliance on generative AI tools as primary information sources. According to the company, platforms like ChatGPT, Gemini, Claude, and Perplexity are routinely queried for data on companies, healthcare facilities, products, services, and public figures.
The platform continuously tracks generated content and analyzes how different models answer identical prompts:
- Baseline Source: The process begins with a client’s official website, which serves as the ground-truth benchmark.
- Automated Testing: Standardized prompts are simultaneously routed to all four AI models.
- Data Ingestion & Analytics: Complete responses are logged with timestamps and model metadata. The platform then evaluates brand visibility, factual accuracy, hallucination rates, cross-model consensus, and content drift over time.
Insights are compiled into interactive dashboards and executive reports to support continuous audit tracking.
Key Algorithmic Performance Indicators (KAPIs) and IGO Framework
To convert audit data into actionable business metrics, Teia Studio introduced proprietary KAPIs (Key Algorithmic Performance Indicators).
The platform operates on a framework termed IGO (Intelligence Governance & Observability), which expands on GEO (Generative Engine Optimization). While GEO focuses on optimizing how brands appear in AI-generated answers, IGO continuously monitors and audits what AI models actually tell end users.
According to the company, the underlying methodology is protected by a patent application filed with Brazil’s National Institute of Industrial Property (INPI) under registration number BR 10 2026 001032-4, and is grounded in a peer-reviewed paper published on Zenodo (DOI: 10.5281/zenodo.19765674).
Pilot Program Reveals Severe Hallucination Rates and Discrepancies
Ahead of its Web Summit launch, Teia Studio conducted a pilot program from March 9 to April 7, 2026, evaluating four real-world brands across four industry sectors over weekly audit cycles.
In one public health case study involving 640 logged interactions over two audit cycles, the platform detected 58 hallucinations, 18 of which were categorized as high severity.
- Surge in Errors: Hallucination occurrences jumped 1,250% between the first and second weeks (rising from 4 incidents to 54).
- Perplexity Risk Concentration: Perplexity accounted for 83% of the severe hallucinations in this specific case. Identified errors included non-existent medical license numbers, incorrect geographic coordinates, broken domain links, fabricated scientific studies, and fake contact information.
Cross-Model Inconsistencies
The pilot study also highlighted major response discrepancies:
- Contradictory Output: In 1 out of every 3 queries, the four AI models yielded contradictory answers to the exact same prompt.
- Low Brand Citation: On average, the target brand was cited in only 44% of responses—even when prompts directly named the institution.
- Lack of User Disclaimers: None of the evaluated AI models issued warnings to users regarding potential inaccuracies in their outputs.
Continuous Monitoring and Algorithmic Optimization
Teia GEO IGO operates on continuous weekly audit cycles. Beyond diagnostics, the platform recommends technical adjustments for official domain structures and digital channels to help AI models recognize official content as primary source material.
According to José Vásquez, Founder and Chief Data & AI Officer (CDAO) at Teia Studio, the platform offers enterprises a structured method to govern how AI models communicate their brand narrative while benchmarking performance against competitors.
Teia Studio’s presentation at Web Summit Rio underscores a growing priority in the enterprise software ecosystem: providing specialized audit tools to measure, track, and verify information generated by artificial intelligence.
