Looking Glass 2026
Most enterprises still treat AI as a set of isolated experiments. That mindset is already outdated. In 2026 the question isn’t “how many models can we deploy?” but “how fast can we rewire our business so agents can operate across it?” The companies that win won’t be the ones building more proofs of concept. They’ll be the ones that rebuild their core workflows so intelligence can move freely, act autonomously and deliver outcomes without waiting on human bottlenecks.
The shift is simple but uncomfortable: operationalizing AI is no longer about scaling models. It’s about designing architectures where agents can execute work with transparency, guardrails and continuous improvement. If your environment can’t support that, you’re not ready for AI that works, because you’re still in experimentation mode. And as the saying goes, you need to be able to walk before you can run.
When AI moves out of the lab and into production, the organization changes shape. Processes become adaptive. Decision loops compress. Governance becomes embedded instead of documented. And human roles shift from performing tasks to overseeing the behavior of intelligent systems.
This is a profound cultural shift for many enterprises — and such step changes require planning and energy to be done successfully.
Key trends
- Agentic workflows and autonomous operations. AI agents are managing complex, multi-step business workflows, reducing latency between insight and action, and allowing organizations to scale responsiveness and productivity without increasing the burden on humans.
- Embedded AI governance. New governance models are integrating AI explainability and transparency directly into operations. This trend matters because trust and accountability will determine the pace of enterprise AI adoption as global regulations tighten.
- AI-enhanced decision systems. Businesses are moving from descriptive analytics to adaptive ‘co-decision’ systems where AI continuously augments human judgment, a major potential boost to agility and precision in fast-moving markets.
- Data fabrics and synthetic data ecosystems. Synthetic data and privacy-preserving techniques such as federated learning are unlocking value from sensitive or limited datasets, addressing one of AI’s biggest bottlenecks: access to high-quality, compliant data.
- Real-time translation and global AI collaboration. Advances from Apple, Google and others in real-time multimodal translation are enabling borderless collaboration across languages and cultures. This will effectively expand the reach of AI-driven businesses, connecting global teams and customers seamlessly.
Signals of this shift include
- Growth in AI pilots moving to full production. Some research, like MIT’s widely-reported and alarm-inducing 95% study, points to a high proportion of AI projects failing or remaining stuck in the experimentation phase. But these setbacks have to be seen as learning experiences that will inform future successes, and there’s equal evidence of businesses extending AI deployments. In one recent poll 58% of enterprises reported AI is now embedded within strategies organization wide, more than double the previous year’s figure.
- The rise of agentic workflows. AI agents are now capable of orchestrating multi-step business processes autonomously, from customer support to DevOps. The growing confidence in agents and the expansion of their responsibilities is evident in enterprise implementations of orchestration frameworks like those offered by OpenAI and Anthropic. The overall orchestration market is forecast to nearly triple to over $30 billion by 2030.
- Integrated AI governance. As AI highlights the vital importance of data governance to digital leaders, more enterprises are adopting end-to-end governance systems for explainability, compliance and bias detection, transforming responsible AI from policy to platform.
- Streamlined data-to-value pipelines. Advances in synthetic data, federated learning and automated labeling are accelerating the data lifecycle from creation to consumption, making AI deployment faster and more secure.
- The emergence of human-AI co-decision systems. Businesses are beginning to grasp that AI’s real value rests not in automation but its potential to enhance decision-making. More organizations are integrating AI into decision loops where human judgment and machine prediction work together to make informed and impactful choices.
- The proliferation of AI-native platforms. Formerly monolithic enterprise systems such as ERP, CRM, HR and finance are evolving into AI-first architectures that embed intelligence into every interaction, boosting operational efficiency and the accuracy of decisions.
Trends to watch
Adopt
Analyze
-
AI assistance across the entire software development lifecycle: requirements, design, coding, testing and deployment.
-
Securing AI systems and governing their use, including prompt injection prevention and model security.
-
Enterprise software with built-in AI capabilities; 40% of apps will have AI agents by 2026.
-
Using ML/AI to identify threats, detect anomalies and automate security operations.
-
Vehicles with advanced driver assistance, progressing toward higher automation levels.
-
An ongoing assessment and prioritization of vulnerabilities based on real exploitability and business impact.
-
Treating datasets as products with defined SLAs, documentation and consumer focus.
-
Connected data flow from design through production and service for traceability.
-
AI models fine-tuned for specific industries or functions, delivering higher accuracy than general-purpose models.
-
Integration patterns using events for loose coupling between enterprise applications.
-
AI assistants powered by generative models for natural conversation and task completion.
-
Decoupled commerce architecture separating frontend presentation from backend functionality.
-
Cloud solutions tailored for specific industries with pre-built compliance, workflows and integrations.
-
Networks configured through business intent rather than device-level commands.
Anticipate
-
Engineered materials with properties not found in nature, enabling new applications.
-
Autonomous AI systems performing data analysis, generating insights and executing data tasks.
-
Using AI to predict infrastructure needs and optimize resource allocation.
-
Using AI/ML to automate cloud resource optimization, cost management and performance tuning.
-
AI detecting and responding to human emotions through facial, voice and behavioral analysis.
-
Brain-inspired computing architectures for energy-efficient AI processing.
-
Cryptographic algorithms resistant to quantum computer attacks; NIST finalized standards in August 2024.
-
Unified platforms that address customer, employee and user experience holistically.
Adopt
Analyze
Anticipate
-
Next-generation wireless research targeting 2030 deployment with terahertz frequencies.
-
AI systems matching human-level intelligence across all cognitive tasks, capable of reasoning, learning and adapting without task-specific programming.
-
Vehicles capable of self-driving in any conditions without human oversight, enabling true driverless transportation.
-
Human-shaped robots capable of diverse physical tasks in unstructured environments, working alongside humans in homes and workplaces.
-
Persistent, photorealistic digital twins of factories, cities and infrastructure enabling real-time simulation, optimization and remote collaboration.
-
AI agents acting as autonomous economic actors, purchasing goods and services on behalf of humans or organizations without human intervention.
-
Digital currencies with embedded rules enabling automated compliance, conditional payments and machine-to-machine transactions.
The opportunities
By getting ahead of the curve on this lens, organizations can:
Operate more effectively
While not every effort to scale AI throughout the enterprise proves successful, a substantial and growing body of research points to AI adopters achieving tangible and sustained productivity and efficiency gains.
Choose courses of action with confidence
AI can never be a direct replacement for human judgment, particularly when choices could have far-reaching consequences. But it can play a powerful supporting role by providing data analysis to inform decisions or increase visibility into their possible outcomes. In areas like product pricing AI’s contributions to decision-making are directly enhancing revenues and profitability.
Develop more proactive approaches to risk
Embedding AI through operations and processes opens the possibility of assigning agents to constantly monitor enterprise resources, flagging changes in the environment or sudden signs of distress. Leading companies like BMW are using AI-supported systems to identify potential faults before they become bigger problems, avoiding potentially costly disruption and downtime.
Better understand and deliver on customer demands
As AI agents become more integrated into a broader spectrum of customer interactions, the data and insights they generate will generate more granular data, and market and segment-specific insights, enabling marketing and experiences to be tailored with greater precision.
Increase capacity to innovate
AI can augment the enterprise’s ability to experiment with and see new concepts through to production both directly — by for example shedding light on areas of opportunity or modeling how a product might function in the field — and indirectly, by allowing team members to carve out more time for creative and collaborative activities.
What we’ve done
A global pharmaceutical company recognized that traditional engagement models with healthcare professionals (HCPs) were limited by generic tools and static dashboards. In response, the organization partnered with Thoughtworks to rewire its engagement approach using data-driven AI tools that support smarter, more contextual interactions.
At the core of the solution is an AI-powered next-best-action recommendation engine built on a unified data platform that integrates diverse sources and delivers tailored suggestions to go-to market representatives. Complementing this, a contextual AI assistant provides real-time product insights, content discovery and patient case information to support representative workflows.
These capabilities have boosted effectiveness and trust: representatives access relevant insights faster, address HCP questions with confidence and tailor communications to individual preferences. By embedding intelligence into daily workflows, the organization is transforming engagement from reactive outreach to adaptive, insight-driven interaction, setting the stage for more autonomous support systems in the future.
Actionable advice
Things to do (Adopt)
- Make AI part of talent strategies. For AI to be integrated and impactful throughout the enterprise, teams will need to embrace changes in their workflows and understand the broader goals the technology is designed to serve. By some estimates, companies miss out on up to 40% of the potential productivity gains of AI due to skill gaps and anxiety, pointing to the need for ongoing training and support.
- Define and track AI ROI. It’s also critical to employee buy-in and the wider success of AI initiatives that definitions of value are established and progress measured regularly. Many of the organizations implementing ambitious AI-led transformations still lack clear end-goals or the ability to measure what they achieve.
- Ensure transparency, and human oversight, in ‘co-decision’ systems. As agents take on more initiative within workflows, their reasoning, evidence and data sources must remain inspectable. Techniques like retrieval-augmented generation (RAG) help reinforce explainability so agents operate within verifiable boundaries.
Things to consider (Analyze)
- Multi-agent systems. As more companies deploy agents, multi-agent systems, where a number of agents collaborate to perform more complex tasks, are becoming more mainstream and producing positive results in fields like agriculture and supply chain management. Businesses considering agentic AI should ensure they have sufficiently robust data resources to support agentic networks and remain on guard against ‘agent-washing.’
- Tighter AI regulation. From US to the UK and India, concerns about cybersecurity and societal impact are fueling plans for tougher rules around the use of AI in multiple contexts. Companies will need to monitor regulatory developments closely and factor a general trend of increased scrutiny and restrictions into their AI plans.
Things to watch for (Anticipate)
- Adaptive AI ecosystems. By 2030, enterprises will function as systems of systems where intelligence flows across every process, platform and product. The challenge will move from scaling AI to governing it sustainably.
- The rise of AI reliability engineering. With concerns about the (mis)use of AI growing, more leaders will invest in ethical design frameworks and agentic orchestration to ensure resilience and trust. The most successful organizations will not just use AI to optimize decisions; they will reimagine how intelligence itself is operationalized within the enterprise and leverage exemplary AI governance as a differentiator.