Study indicates that 74% of companies face unexpected costs with AI and identifies governance as a critical factor for scaling the technology
A survey conducted by Sensedia and Squadra in partnership with MIT Sloan Management Review Brasil indicates that the competitive advantage is no longer about adopting AI, but rather its integration and governance at scale.
SÃO PAULO, August 4, 2026 - PRNewswire
Artificial Intelligence (AI) has become a strategic priority for companies, but the transition from pilot projects to large-scale implementation remains one of the biggest challenges for organizations. This is revealed by a study developed by Sensedia and Squadra, in partnership with MIT Sloan Management Review Brasil, which shows a significant gap between intention and execution: although 66.2% of executives consider AI a business priority, only 7.4% have managed to integrate it across their entire operation.
The study shows that the challenge is no longer experimenting with AI, but scaling it efficiently and with a clear strategy. Decentralized adoption and lack of governance have led 74% of organizations to face unforeseen costs, technological redundancies, and increased compliance risks.
Decentralized adoption and lack of governance have led 74% of organizations to face unforeseen costs, technological redundancies, and increased compliance risks. The report is based on a digital maturity survey conducted during the APIX 2026 event, complemented by interviews with leaders from companies such as Bradesco, MAPFRE, Inter, Hcor, Sensedia, Squadra, and Klabin, as well as an analysis of surveys from global companies like Bain & Company, Gartner, McKinsey, and Deloitte.
Among its main conclusions, the study points out that isolated improvements in productivity and automation do not, in themselves, translate into significant financial impact or sustainable competitive advantage. The true differentiator lies in the ability to integrate AI into the organization's technological architecture and align it directly with strategic business objectives.
The challenges of infrastructure and data control
The study exposes serious limitations in companies' data infrastructure. For 54.4% of technology professionals, access to the corporate data needed to feed AI solutions is still unstructured. Of these, 31.6% rely on manual context insertion into tools, while 22.8% use solutions disconnected from company information.
Added to this is the lack of control structures, with 41.4% of the technical team working without formal risk management processes and only 4.8% of executives reporting having consolidated governance models, supported by specialized platforms.
Furthermore, 80% of organizations lack structured agent architectures, and 68% have not implemented the Model Context Protocol (MCP) in production environments — essential components for the next generation of intelligent agent-based applications.
In this regard, José Gómez Amador, Regional Business Manager at Sensedia, highlights that AI adoption is no longer the main challenge. As intelligent agents take over critical processes, companies need an architecture capable of connecting data, systems, and models under a single governance layer. Without it, costs, operational failures, and scalability become more difficult to control.
In this new phase of maturity, the market is migrating from the first wave of copilots and ad hoc automations to scalable operations, integrated into core processes and driven by business results. Enterprise architecture, integration, and data governance are becoming strategic assets. Companies that maintain isolated initiatives will achieve only marginal efficiency gains; on the other hand, those that integrate AI into their core operations will achieve sustainable competitive advantage.
The report also projects rapid evolution towards agent-based AI ecosystems and multi-agent architectures, where applications cease to be mere assistants and begin to perform tasks, make decisions, and interact autonomously with corporate systems. This requires agent-ready infrastructures that ensure secure, scalable, and contextualized access to APIs and data.
The strategic role of APIs and success stories in the region
With the advancement of these tools in daily operations, APIs and integration platforms take a leading role, becoming the infrastructure that provides context, traceability, security, and governance. Amador emphasizes that APIs are the layer that connects intelligence to decision-making and processes, proving essential in an agent-dominated environment to ensure observability, cost control, and governance at scale.
The organizations leading the AI-driven transformation today have something in common: they have not built their success on isolated projects, but rather on a robust data and integration architecture that allows them to scale AI safely and efficiently. Bradesco is an excellent example. The company already operates over 600 AI applications in production, a strategy that generates an estimated value of over US$45 million annually, demonstrating that the true return on AI investment is achieved when it is integrated into the core business, not just used as isolated initiatives.
Similarly, Inter operates over 500 models that support critical decisions in credit analysis, fraud prevention, and personalization, while MAPFRE is implementing intelligent agents in customer service and claims management. Meanwhile, institutions like Hcor and Klabin are advancing in health data interoperability and multi-agent platforms with centralized governance.
By 2028, it is projected that most companies will replace isolated AI tools with end-to-end intelligent automation workflows. Gartner warns that companies attempting to overlay AI onto legacy architectures without structural modernization will see their profit margins drastically reduced.
In this context, Amador concludes that competitive advantage will not be determined by the number of AI models implemented, but by the ability to orchestrate agents, data, and APIs within a secure architecture. This infrastructure is what transforms proof-of-concepts into real operational intelligence.
Finally, Rômulo Cioffi, Innovation Director at Squadra, highlights that generative AI has begun to commoditize aspects of software engineering, shifting strategic advantage to decision architecture. Therefore, leading companies will be those capable of integrating intelligence directly into their operational workflows through agents that can observe, interpret, act, and learn continuously.
The full study is available here. Download and access it.
About Sensedia
Sensedia is a multinational provider of API management platforms, agent governance, and integration solutions, as well as professional services. The company drives the digital transformation of large enterprises through more agile, modern, and scalable architectures, enabling its clients to offer digital products and experiences to their entire ecosystem. Learn more at sensedia.com.es.
About Squadra
Squadra is one of the largest technology consultancies in Brazil and a digital partner for brands such as Inter, Vale, VLI, Unimed, Claro, and MRV. It combines comprehensive design, applied AI, and technical excellence in development, operations, and data to build modern, intelligent, and scalable digital platforms. It brings together over 700 technology professionals working throughout Brazil and on international projects. With its proprietary AI platform, Genius, Squadra enhances results and accelerates digital transformation with real impact. Learn more at squadra.com.br.
Available at: Estadão | PR Newswire
