The "Bolted-On" Trap
Arthur Mensch, CEO of Mistral AI, opens McKinsey's latest report with a stark diagnosis: "Today, AI is bolted on. But to deliver real impact, it must be integrated into core processes, becoming a catalyst for business transformation rather than a sidecar tool" [1]. This observation cuts to the heart of why 80% of companies report no material impact from their AI initiatives despite widespread deployment [2].
The "bolted-on" approach—adding AI capabilities to existing workflows without fundamental process redesign—has become the default strategy for most organizations. It's also the primary reason why AI transformation consistently fails to deliver promised value. McKinsey's research reveals that most deployments use AI "in a shallow way—as an assistant that sits alongside existing workflows and processes— rather than as a deeply integrated, engaged, and powerful agent of transformation" [3].
This shallow integration creates what we call the "sidecar syndrome": AI systems that operate parallel to core business processes without fundamentally changing how work gets done. The result is incremental efficiency gains that are "spread thinly across employees" and "not easily visible in terms of top- or bottom-line results" [4]. For organizations that have invested millions in AI transformation, this represents one of the most expensive strategic mistakes in modern business history.
The bolted-on approach dominates enterprise AI deployment because it appears to offer the path of least resistance. Organizations can implement horizontal use cases like Microsoft 365 Copilot or Google Workspace AI with minimal disruption to existing operations. Nearly 70% of Fortune 500 companies have adopted Microsoft 365 Copilot [5], largely because "enabling Microsoft Copilot is as simple as activating an extension to an existing Office 365 contract, requiring no redesign of workflows or major change management efforts" [5].
This ease of implementation creates the illusion of AI progress while avoiding the difficult work of business transformation. IT departments can demonstrate AI deployment. Employees can point to AI tools in their daily workflows. Executives can report AI adoption metrics to boards and investors. Yet none of this activity translates to meaningful business value because the underlying processes remain unchanged.
The bolted-on approach also appeals to risk-averse organizations that want to "experiment" with AI without committing to fundamental change. By treating AI as an add-on rather than a transformation catalyst, companies believe they can capture AI benefits while maintaining operational stability. McKinsey's data proves this assumption false: organizations pursuing bolt-on strategies consistently fail to achieve material business impact.
McKinsey's research identifies specific technological constraints that make bolted-on AI inherently limited. First-generation large language models (LLMs) "faced limitations that significantly constrained their deployment at enterprise scale" [5]. These limitations become particularly problematic when AI operates as a sidecar to existing processes rather than being integrated into core workflows.
Accuracy and Reliability Issues: LLMs can produce inaccurate outputs, making them "difficult to trust in environments where precision and repeatability are essential". When AI operates as a bolt-on assistant, these accuracy issues create additional verification overhead rather than process improvement.
Passive Operation Mode: LLMs are "fundamentally passive; they do not act unless prompted and cannot integrate seamlessly with existing business systems" [5]. This passivity means bolted-on AI requires constant human intervention, limiting its ability to automate or optimize business processes.
Integration Complexity: Bolt-on AI systems struggle with the data integration and system connectivity required for meaningful business impact. Without deep integration into core processes, AI tools operate with incomplete information and limited context.
Scalability Constraints: Sidecar AI implementations create technical debt that makes scaling prohibitively expensive. Each bolt-on system requires separate maintenance, governance, and optimization, creating what McKinsey calls "mounting technical debt and new classes of risk".
The solution to the bolted-on trap requires recognizing that effective AI implementation demands process redesign from the ground up. McKinsey emphasizes that "unlocking the full potential of agentic AI requires more than plugging agents into existing workflows. It calls for reimagining those workflows from the ground up—with agents at the core".
This process-centric approach represents a fundamental shift from traditional technology implementation. Instead of asking "How do we add AI to our existing processes?" organizations must ask "How do we redesign our processes around AI capabilities?" This shift requires understanding AI not as a tool that enhances existing work but as a capability that enables entirely new ways of working.
Decision-Making Transformation: AI-integrated processes change how decisions are made, moving from human-centric decision trees to AI-augmented decision frameworks that operate at machine speed with human oversight.
Workflow Automation: Instead of AI assisting with individual tasks, integrated AI automates entire workflows, eliminating handoffs and reducing cycle times.
Data-Driven Operations: AI-integrated processes operate on real-time data analysis rather than periodic reporting, enabling continuous optimization and predictive management.
Adaptive Capabilities: Integrated AI systems learn and improve from process execution, creating self-optimizing operations that become more effective over time.
At Aidols, we've designed our methodology around the fundamental principle that AI transformation requires process transformation. Our AI-native approach doesn't bolt
AI onto existing workflows—we redesign workflows around AI capabilities from the ground up.
Process-First Design: We begin every engagement by analyzing existing business processes to identify transformation opportunities rather than automation targets. This ensures AI integration creates new value rather than marginal efficiency gains.
Integrated Architecture: Our AI systems are designed as integral components of business processes, not external assistants. This deep integration enables the kind of transformation that creates competitive advantage.
Continuous Optimization: Unlike bolt-on systems that require manual optimization, our AI-integrated processes continuously improve through machine learning and real time feedback loops.
Measurable Transformation: By redesigning processes around AI capabilities, we deliver measurable business transformation rather than diffuse productivity improvements.
McKinsey's research reveals that organizations must choose between two fundamentally different approaches to AI: bolt-on implementations that deliver minimal value or integrated transformations that create competitive advantage. This choice has profound implications for competitive positioning in an AI-driven economy.
Organizations pursuing bolt-on strategies remain fundamentally unchanged by AI. They may use AI tools, but they operate with the same decision-making speed, process efficiency, and innovation capabilities as pre-AI organizations. Meanwhile, competitors pursuing AI integration develop fundamentally superior operational capabilities.
The competitive gap is accelerating. As McKinsey notes, "the moment has come to bring the gen AI experimentation chapter to a close". Organizations that continue bolting AI onto existing processes will find themselves competing against AI-native companies that operate with integrated AI capabilities.
The transition from bolt-on to integrated AI requires more than technical changes—it demands organizational transformation. McKinsey emphasizes that organizations must "reset their AI transformation approaches from scattered initiatives to strategic programs; from use cases to business processes; from siloed AI teams to cross functional transformation squads".
This reset requires abandoning the comfortable fiction that AI can deliver transformation without transformation. The organizations succeeding with AI aren't using better bolt-on strategies—they're pursuing fundamental process redesign that integrates AI capabilities into core operations.
At Aidols, we specialize in this transition. Our AI-native consulting methodology helps organizations move beyond bolt-on implementations to achieve true AI integration. We don't add AI to existing processes—we redesign processes around AI capabilities to create sustainable competitive advantages.
The bolted-on trap isn't a technical problem—it's a strategic choice. Organizations can continue pursuing marginal improvements through AI assistants, or they can embrace the process transformation required for AI integration. The data is clear: only one approach delivers material business value.
Ready to move beyond bolt-on AI? Aidols redesigns business processes around AI capabilities to deliver integrated transformation in 3-6 months. No sidecars. No assistants. No bolt-ons. Learn more at www.aidolsgroup.com.
References
[1] Arthur Mensch, CEO of Mistral AI. Foreword in "Seizing the agentic AI advantage: A CEO playbook to solve the gen AI paradox and unlock scalable impact with AI agents." McKinsey & Company, June 2025.
[2] McKinsey & Company. "The state of AI: How organizations are rewiring to capture value." March 12, 2025.
[3] McKinsey & Company. "Seizing the agentic AI advantage: A CEO playbook to solve the gen AI paradox and unlock scalable impact with AI agents." June 2025.
[4] Satya Nadella. "Microsoft Fiscal Year 2025 First Quarter Earnings Conference Call." Microsoft, October 30, 2024.
[5] McKinsey & Company. "Seizing the agentic AI advantage: A CEO playbook to solve the gen AI paradox and unlock scalable impact with AI agents." June 2025.
