A Progress Framework for AI Adoption in your Organisation
Artificial intelligence is transforming industries, yet many organisations struggle to determine how AI fits within their business. Instead of chasing hype-driven solutions, organisations need a structured approach to AI discovery–one that prioritises real business needs over technology trends and unlocks value throughout the journey.
Most business operations rely on processes and systems with predictable outcomes based on known inputs (deterministic). In contrast, Large Language Models (LLMs) and other AI-driven solutions produce non-deterministic outputs, introducing uncertainty. It’s little wonder that combining these two systems has created angst, caution and resistance due to the potential disappointments and risks involved. The continual and rapid development of AI and Agentic AI creates additional reservations around future-proofing your technology decisions.
To approach AI strategically, consider these two key questions:
- How can we take this really powerful tool of LLMs, with their inherent non-deterministic behaviour, and build a deterministic (structured, reliable) system around it?
- How can we create an AI strategy and associated architecture that fosters innovation in the short term while adapting to industry advancements over time?
Successfully adopting AI requires a balance between traditional, well-proven engineering disciplines in addition to a new mindset for solution design, particularly if your strategy involves the adoption of Agentic AI.
This blog outlines a 5-step framework to help your organisation progress AI adoption from initial business stakeholder engagement through to ongoing value assessment.
1. Business event Mapping through Event Storming
A business-led AI discovery process ensures AI investments align with real business needs. By focusing on business problems rather than technology trends, simplifying AI concepts and leveraging methodologies like event storming can help identify high-value AI opportunities.
Event Storming facilitates AI adoption by:
- Identifying decision points in business processes where AI can provide the most value.
- Mapping information flows to understand where data is generated, transformed, and consumed.
- Recognising event cascades that could benefit from real-time AI monitoring and response.
- Separating deterministic processes from areas where non-deterministic AI can add value.
By visualising the entire business domain through events, organisations can pinpoint specific moments where AI capabilities can enhance workflows without disrupting existing deterministic systems. This creates a ‘wrapper architecture’, containing LLM-based unpredictability within well-defined boundaries while leveraging its strengths.
Short-term Action: Create event maps for high-value business domains, highlighting decision points that require contextual intelligence or pattern recognition.
Long-term Vision: Establish an event-driven architecture that enables future AI agents to autonomously observe, reason, and act on business events.
2. Bounded Contexts and Domain Driven Design for AI Integration
Domain-Driven Design (DDD) helps manage AI complexity by defining clear boundaries for AI capabilities, ensuring AI solutions integrate seamlessly with existing systems.
Key DDD Principles for AI adoption:
- Bounded contexts prevent AI scope creep and clarify system responsibilities.
- Ubiquitous language ensures shared understanding between business stakeholders and technical teams.
- Context mapping outlines how AI components interact with existing ecosystems.
- Strategic domain patterns distinguish where AI should augment vs. replace existing processes.
This approach enables organisations to implement AI in manageable, business-aligned increments while maintaining a cohesive overall architecture.
Short-term Action: Define bounded contexts for initial AI implementations, with clear interfaces to existing systems and explicit handling of non-deterministic outputs.
Long-term Vision: Create a domain model that accommodates increasingly autonomous AI agents with well-defined responsibilities and interaction patterns.
3. Capability-Driven Discovery: From Tasks to Agentic Functions
Rather than focusing on specific AI technologies, organisations should identify the capabilities they need to enhance business outcomes.
Capability-Driven Discovery enables organisations to:
- Identify capacity gaps where current human-driven processes cannot scale.
- Map augmentation opportunities where AI can enhance human decision-making.
- Discover new capabilities that were previously infeasible without AI.
- Define capability tiers from basic automation to advanced cognition.
This framework prioritises AI investments based on business impact while creating a roadmap that evolves with technology.
Short-term Action: Inventory current manual processes and decision points that could benefit from existing AI capabilities like classification, prediction, or natural language understanding.
Long-term Vision: Define higher-order capabilities that would become possible with agentic AI, such as autonomous planning, negotiation, or complex problem-solving.
4. Agent-Oriented Architecture: Building Blocks for Future AI
An agent-oriented approach provides a powerful paradigm for scalable, evolvable AI systems.
Key elements of Agent-Oriented AI Design:
- Single-responsibility AI agents that excel at specific tasks.
- Composable architectures that enable AI agents to collaborate for complex workflows.
- Observation-oriented design where agents monitor events and trigger responses.
- Progressive autonomy frameworks that gradually increase agent independence.
This structure delivers immediate value through focused AI implementations while providing a foundation for more advanced agentic AI solutions.
Short-term Action: Deploy simple AI ‘agents’ for targeted tasks (e.g., document classification, anomaly detection, routine response generation) with human oversight.
Long-term Vision: Develop multi-agent architectures where specialized AI agents collaborate, negotiate, and solve complex problems with minimal human intervention.
5. Continuous Value Assessment and Capability Expansion
AI discovery and adoption is not a one-time exercise, but an evolving process that requires continuous evaluation.
Sustaining AI value involves:
- Value-complexity mapping to reassess AI opportunities over time.
- Capability maturity assessments to track evolving AI technologies.
- Technical feasibility monitoring to identify when emerging AI approaches become viable.
- Feedback loops to capture insights from early implementations
This creates a dynamic AI roadmap that evolves with both business needs and technological advancements.
Short-term Action: Establish a framework to assess AI implementations against business outcomes, with clear success metrics.
Long-term Vision: Develop a capability expansion framework that systematically identifies opportunities for increased AI autonomy and agentic behavior.
Putting It All Together: The Progressive AI Adoption Framework
These five steps work together to create a structured, comprehensive approach to AI discovery, balancing immediate business value with long-term innovation:
- Event Storming identifies where AI can add value in business processes.
- Domain-Driven Design creates clear boundaries and interfaces for AI components.
- Capability-Driven Discovery prioritizes based on business impact.
- Agent-Oriented Architecture provides a scalable implementation approach.
- Continuous Value Assessment ensures ongoing alignment with both business needs and technological evolution.
By following this framework, organisations can strategically implement AI, delivering value at each stage while building towards a more sophisticated agentic future.
Are you ready to explore the potential of AI Agents in your organisation?
Let’s talk about how AI solutions can transform the way you work.
