Anticipation was great among the TOGAF and EA communities for the long-awaited rollout of the latest edition of the TOGAF Standard, the 10^th Edition, after over a dozen years of TOGAF 9. The expectation was that a new edition would come out by the summer of 2020, but updating it proved so complicated that the 10^th Edition didn’t come out until late 2022. In the process, a major decision was made to keep most of the core of TOGAF 9.2 in the TOGAF 10 and to allow selected TOGAF Series Guides to be incorporated into the new edition. In the process, the size of the latest version tripled over that of TOGAF 9.2.

One would assume that TOGAF 10 would be so much better because of the time that had been needed to deliver it and its expansion to include some more specialized guidance on topics such as Security, Risk, and Agile.

Because of its unwieldly size, though, major decisions were made to downplay the names of work products (deliverables and artifacts) and techniques. To apply an implicit principle of minimum jargon, something dominant in TOGAF 9 EA training, and on ADM project cycles rather than the overall EA practice, the concept of Essential Knowledge and Outcomes emerged for the ADM cycle phases.

In addition, the whole approach to the purpose of EA (identifying gaps and building a consensus among stakeholders on how to fill them) has become very simplistic and undermines the critical expertise of architects to help paint a comprehensive picture of the shifting EA landscape for optimum impact and what if analyses.

In short, TOGAF EA training/certification is now focused on more abstract notions and nuanced considerations kowtowing to a narrowed definition and oversized role of stakeholders than mastery of the various building blocks that had been the heart of TOGAF EA training/certification for over a decade.

Consequently, there are now huge holes in TOGAF 10 that can undermine effective training and the ultimate value of TOGAF for realistic use cases. The holes are critical in terms of the nearly absent Preliminary Phase and no emphasis on traditional TOGAF building blocks, such as work products and techniques, or even steps for tackling the ADM Phases.

As a result, the selection of a highly experienced trainer is more critical than ever if TOGAF will be valuable in real-world situations, whether to set up/mature/sustain an EA practice or to relay key TOGAF ADM building blocks, such as viewpoints/artifacts, deliverables, and techniques that are part and parcel of doing the steps of each phase.

In conclusion, highly experienced and capable training is more critical than over to patch the holes in TOGAF 10, largely by still emphasizing key learning points of TOGAF 9.2 and by heavy inclusion of meaningful use cases to better illustrate how effective TOGAF can be in its ultimate sweet spot: problem/opportunity identification and associated roadmapping to targeted, valued business outcomes.

To best illustrate the type of scenario possible to tackle with proper patches by expert training, consider the one below (for streamlining, specific TOGAF deliverables and artifacts are not called out and only selected ADM Phase steps are included):

Modernizing Nirvana County’s 911 System: A Secure and Resilient Future with GenAI

Introduction

Nirvana County is committed to protecting its citizens with a modern and reliable emergency response system. Modernizing its legacy 911 system to a Next-Generation 911 (NG911) with the potential integration of Generative AI (GenAI) will significantly improve response times, enhance public safety, and ensure our future preparedness.

It will leverage the TOGAF framework and integrate specific techniques to address key concerns, while exploring the feasibility and potential benefits of GenAI:

Phase 1: Architecture Practice and Project Vision (Preliminary & Architecture Vision)

• EA Maturity and Skills Assessment in the Preliminary Phase:  Assess current state vs. necessary maturity and skills needed to define a roadmap to achieve them in organizational context, also considering the culture and related methods.
• Stakeholder Identification and Management for the ADM Cycle planned in the Architecture Vision Phase (Phase A). 
• Identify the targeted, valued business outcomes.
• Define project-aligned core principles and potential GenAI integration in the ADM Cycle.
• Interoperability Requirements Technique (IRT): Identify overarching interoperability requirements (further refined in the Domain Architectures and Transition Planning).

Phase 2: Business Architecture (Value Stream Analysis, Capabilities and Risks Assessment)

• Value Stream Analysis: Mapping workflows and identifying bottlenecks that need to be addressed to achieve the targeted, valued business outcomes.
• Capabilities and Risks Assessment: Analyzing NG911 capabilities and associated risks.

Note: These first 2 items are also key to setting up the cycle in Phase A.

• Leverage ArchiMate and BPMN to analyze and plan business processes.
• Define Candidate Target Domain Architectures, respective gaps, and candidate building blocks to fill the gaps.

Phase 3: Information Systems Architectures  

• Leveraging ArchiMate and UML: Design the Candidate Target Application and Data Architectures, gaps, and Candidate Building Blocks to fill the gaps.

Phase 4: Technology Architecture 

• Baseline Technology Architecture: Develop a Baseline Description of the existing technology architecture, identifying potential integration points for GenAI.
• Target Technology Architecture: Design the Candidate Target Technology Architecture considering scalability, security, and integration with GenAI components. Utilize ArchiMate diagrams to visualize the architecture and data flows.
• Technology Standards and Governance: Establish technology standards and governance policies to ensure consistent and secure GenAI integration.

Phase 5: Transition Planning & Implementation 

• Architecture Roadmap: Define a roadmap for transitioning from the baseline to the target architecture, including for the Domain Architectures and GenAI integration milestones.
• Mitigation Strategies: Proactive strategies using GenAI to address risks identified in earlier phases.

• Communication Plans: Engaging stakeholders, vendors, and community.
  • Develop a detailed rollout plan that phases in GenAI functionalities for controlled evaluation and feedback.
  • Establish robust training programs for emergency responders and public on using the new GenAI-powered features.
  • Define clear data governance policies and privacy measures for GenAI data collection and usage.
• Conduct Architecture Review Boards.
• Change Management: Implement a comprehensive change management plan to prepare stakeholders for the transition and address potential concerns.

Phase 6: Requirements Management & Governance 

• Monitoring and Evaluation: Continuously monitor GenAI performance, identify areas for improvement, and adapt strategies as needed.
• Risk Management: Maintain a proactive risk management approach to mitigate potential issues with GenAI integration.
• Ethical Oversight: Establish an independent ethical review board to assess and guide the responsible use of GenAI in the NG911 system.

Phase 7: Benefits & Outcomes 

• Highlight positive impacts: Quantify improvements in response times, resource allocation, and decision-making through GenAI integration.
• Showcase community benefits: Demonstrate how GenAI enhances accessibility, inclusivity, and public trust in the 911 system.
• Share best practices: Document lessons learned and share successful practices with other communities considering GenAI adoption.

In conclusion, the above sample scenario illustrates how architecture expertise is essential to emphasize in the application of TOGAF for varied use cases. In other words, TOGAF 10’s overemphasis on abstracted ADM Phase Outcomes and Knowledge is not nearly actionable enough and dwells far too much on stakeholders as driving the work, as if the expertise of the architecture team/architects is only of tangential importance. Effective training is necessary to patch the approach of TOGAF 10 in order to move closer to making practical use of it.

Authored by Dr. Steve Else, Chief Architect & Principal Instructor