Jan. 29, 2026 /Mpelembe Media/ — Thoughtworks introduced their latest Looking Glass report, which outlines necessary structural changes for businesses in the agentic AI era. The report emphasizes that companies must move beyond minor experiments to rewire their core architectures, ensuring that data, platforms, and security are fully integrated to support autonomous intelligence. Key focus areas include modernizing software delivery, transitioning to intent-driven user experiences, and establishing federated data ecosystems that provide real-time information to AI agents. Additionally, the report highlight the importance of computational governance, suggesting that safety and privacy regulations should be built directly into a company’s technical foundation. This overview also mentions the Mpelembe Network, a platform that distributes such technology news and analysis to help leaders navigate rapidly evolving industry trends. Together, these documents serve as a strategic guide for enterprises aiming to maintain architectural integrity while adopting transformative digital innovations.
To rewire their systems for the era of agentic AI, enterprises must move beyond isolated experiments and undergo a structural reconfiguration of their core architectures, workflows, and data ecosystems. According to the sources, this transition involves several critical shifts across the technical and operational landscape:
Rebuilding Core Systems with AI-First Software Delivery (AIFSD): Enterprises should modernize legacy estates by adopting AI-First Software Delivery. This approach requires a “co-construction” model where human engineers provide rigorous oversight to ensure that AI-generated code maintains architectural integrity and does not introduce security vulnerabilities or technical debt.
Breaking Down Data Silos for Agentic Access: For agents to function effectively, they must be able to access information from anywhere within the organization. This requires rebuilding workflows to eliminate human bottlenecks, allowing agents to deliver work with transparency and continuous improvement through established guardrails.
Transitioning to Data Mesh 2.0: Legacy data lakes are no longer sufficient; instead, organizations should evolve toward AI-ready data ecosystems. These are product-centric, federated environments—often referred to as Data Mesh 2.0—that provide trustworthy, real-time data to both human users and intelligent agents.
Adopting Intent-Driven Interaction Choreography: The sources suggest a shift from traditional screen-based interfaces to intent-driven experiences. Systems should act as “cognitive exoskeletons” that use multimodal inputs (such as voice and gesture) to anticipate user goals and adapt to specific contexts through interaction choreography.
Implementing Computational Governance: As regulations increase, enterprises must move toward “computational governance.” This means codifying safety, privacy, and security requirements directly into the architecture and software delivery pipelines rather than treating them as aspirational goals.
By integrating these shifts, the sources indicate that technology will no longer be a set of disruptive tools but a reconfigured foundation that creates value across the entire enterprise.
While the transition to AI-First Software Delivery (AIFSD) offers the potential to accelerate development cycles and modernize legacy systems, the sources identify several significant risks associated with using AI in software development:
Compounding Technical Debt: Generative AI tools, if used without rigorous oversight, risk increasing an organization’s technical debt, which can make future maintenance and updates more difficult.
Subtle Security Vulnerabilities: One of the primary concerns with AI-generated code is the introduction of subtle security vulnerabilities that may be overlooked during the development process.
Loss of Architectural Integrity: There is a risk that organizations may prioritize speed over quality. Without a “co-construction” model where human engineers provide rigorous oversight, AI-generated code can compromise the overall architectural integrity of core systems.
Evolving Regulatory and Safety Risks: As technology advances, organizations face the challenge of tightening regulations. Failure to move toward “computational governance”—where safety, privacy, and security are codified directly into the delivery pipeline—can leave an enterprise vulnerable to compliance and ethical failures.
To address these risks, the sources emphasize that success depends on humans ensuring AI-driven processes maintain high standards rather than just focusing on rapid output.
The integration of AI is driving a structural reconfiguration of both human interactions and data ecosystems, moving them away from static, siloed models toward adaptive, intent-driven environments,.
Changes in Human Interactions
AI is fundamentally reimagining how humans engage with technology by shifting the focus from screen-based interfaces to agentic, intent-driven experiences. Key developments include:
Interaction Choreography: The sources describe a shift toward systems that act as “cognitive exoskeletons.” These systems use multimodal inputs—such as voice and gesture—to anticipate user goals and adapt to the specific context of the interaction.
Co-construction in Workflows: In specialized fields like software development, the interaction model is evolving into a “co-construction” model. Here, humans do not just use AI as a tool but provide rigorous oversight to ensure the architectural integrity and security of AI-generated outputs.
Adaptive Systems: Overall, interactions are becoming more adaptive, focusing on the user’s intent rather than requiring the user to navigate complex manual interfaces.
Changes in Data Ecosystems
To support agentic AI, legacy data structures are being replaced by more dynamic, federated architectures. These changes include:
Transition to Data Mesh 2.0: Traditional data lakes are increasingly viewed as insufficient. Organizations are moving toward AI-ready data ecosystems—specifically Data Mesh 2.0—which are product-centric, federated environments designed to provide trustworthy, real-time data to both humans and intelligent agents.
Elimination of Data Silos: For AI to be effective, data must flow freely across the enterprise. This requires breaking down data silos so that agents can access information from anywhere within the organization without being hindered by human bottlenecks.
Computational Governance: As mentioned in our previous discussion regarding system rewiring, data ecosystems are integrating “computational governance.” This means that safety, privacy, and security requirements are no longer just aspirational goals but are codified directly into the architecture and data delivery pipelines.
By converging platforms, data, and AI, technology is no longer just a set of tools but a reconfigured foundation that creates value by anticipating needs and providing seamless access to intelligence,.
SOURCE Thoughtworks