Menu
Log in



Log in

10 Implementation

Improving Practices and Enabling a Decision-Support System


Introduction

1.0 Implementing Strategic Predesign Improvements

1.1 Connect Strategy, Capability, and Decision Support

2.0 Improve the Organization

2.1 Align the Organizational Design Dimensions

2.2 Make Change Management Part of the Work

3.0 Competitive Advantage or Public Value

4.0 Build an Integrated Decision-Support System

4.1 Use AI: Interpretation and Draft Assistance

4.2 Software: Integration, Automation, and Validated Analysis

4.3 Data: Authoritative Records and Controlled Evidence

4.4 People: Judgment, Accountability, and Authorization

4.5 Apply Govern, Guide, Guard

5.0 Principles for Implementation

5.1 Start with the Decision

5.2 Integrate the Process, Systems, and Data

5.3 Establish Trusted, Traceable Evidence

5.4 Govern Internal and External Information Differently

5.5 Use Agentic Capabilities Selectively and Within Boundaries

5.6 Preserve Human Authority and Accountability

5.7 Compare Alternatives and Reveal Uncertainty

5.8 Preserve Information Continuity Across the Lifecycle

5.9 Capture Reasoning and Learn from Outcomes

5.10 Redesign the Work Process with Its Practitioners

5.11 Pilot Narrowly and Measure Decision Value

5.12 Requirements to Define and Validate through Implementation


6.0 Apply Improvements through the Predesign Phases

6.1 Strategic Planning: Establish Direction and Compare Portfolio Responses

6.2 Real Estate: Test Fit and Control Commitments

6.3 Project Definition: Reconcile Requirements and Establish Readiness

7.0 Implementation Roadmap

7.1 Define the Decisions and Governance

7.2 Establish the Trusted Information Foundation

7.3 Design a Bounded Pilot

7.4 Configure, Test, and Train

7.5 Evaluate Decision Value

7.6 Expand and Connect Lifecycle Learning

7.7 Pilot Scorecard and Acceptance Rules

7.8 Roadmap RACI Matrix

8.0 Implementation Outcome

Introduction

Implementation turns strategic predesign into a repeatable organizational capability. It is a coordinated way to translate enterprise strategy into justified decisions and approved requirements before design begins or capital is committed. This chapter explains how to improve organizational arrangements, work processes, information, and technology together. The aim is better leadership decisions that achieve competitive advantage or public value.

1.0 Implementing Strategic Predesign Improvements

1.1 Connect Strategy, Capability, and Decision Support

Implementation defines the capability needed, redesign the work with its practitioners, configure the supporting information and tools, test the results, and retain approved improvements. Foundations provides the lifecycle and systems framework. Organization explains authority, participation, organizational design, and change management. Process defines the three predesign phases—Strategic Planning, Real Estate, and Project Definition—and the decisions and evidence developed within them. Procedures supply adaptable protocols and review questions. Techniques provide repeatable methods, tools, templates, and reviewable outputs., while Decision-Support System organizes evidence preparation, analysis, governance, and controlled records into four logical tiers. These chapters to define the work; use Implementation to assign ownership, configure the workflow, verify performance, and adopt tested improvements.

Implement a strategic predesign decision-support system as a redesigned operating model. Fragmented information, mismatched planning horizons, conflicting priorities, unclear ownership, and static business cases are problems of coordinated work as well as technology. Address their causes through a defined workflow, governed evidence, integrated software, proportionate AI assistance, and identifiable leadership judgment and accountability.

2.0 Improve the Organization

Missing evidence creates uncertainty. Requiring the collect and validate the information needed for the decision.

Different interpretations create equivocality. Necessitating facilitated work sessions, concrete options, and feedback to establish operational meaning.

Competing priorities require explicit criteria and an authorized decision; more data alone cannot decide which priority should prevail.

Assess whether information demand exceeds the organization’s capacity to review and coordinate it.

Reduce unnecessary transfers and duplicate work, provide approved capacity or schedule allowances where justified, and strengthen information systems and lateral coordination where needed. Record the diagnosis, chosen response, responsible parties, and remain uncertainty before configuring automation.

Evaluate the organizational capability the strategy requires. Examples include comparing portfolio options consistently, adapting capacity as demand changes, or translating business-unit needs into design-ready requirements. For each capability, identify the decisions it supports, the current performance gap, the changes needed, the accountable owner, and the evidence that will demonstrate improvement.

2.1 Align the Organizational Design Dimensions

Apply the organization-design framework. Strategy establishes direction; required capabilities connect that direction to choices about structure, processes, people, and rewards. A new tool cannot resolve conflicting authority, inadequate staffing, or incentives that favor isolated departmental outcomes.

  • ·       Strategy: Define the organizational outcome and the criteria for judging alternatives. Determine whether changing work, technology, or use of existing assets can address the need before assuming a property transaction or new building.
  • ·       Structure: Confirm sponsor authority, phase leadership, decision rights, and escalation. Establish a cross-functional team with the participation needed from business units, Real Estate, HR, IT, Finance, Legal, Procurement, and operations. Do not create a parallel approval hierarchy for the technology pilot.
  • ·       Processes: Map how requests, evidence, recommendations, reviews, approvals, and handoffs move between functions. Remove duplicate entry and unnecessary transfers, clarify acceptance criteria, and assign unresolved issues. Use lateral coordination to investigate exceptions while retaining mandatory reviews and delegated limits.
  • ·       People: Allocate time for stakeholder participation, evidence checking, training, support, and onboarding. Pair incoming practitioners with experienced internal leads or advisers during actual work. Preserve specialist responsibilities rather than treating AI as a substitute for professional judgment.
  • ·       Rewards and measures: Align recognition and shared process measures with evidence quality, useful decisions, reliable handoffs, and organizational outcomes. Avoid incentives that reward speed, tool usage, or space reduction while concealing review effort, unresolved requirements, or operational consequences.

2.2 Make Change Management Part of the Work

Start engagement during Strategic Planning and continue through Real Estate and Project Definition. Use vision sessions, user interviews, cross-functional workshops, and repeated option reviews to establish shared meaning and test apparent agreement.

Explain terms such as flexibility, collaboration, and efficiency in the context of actual work. Record dissent, operational constraints, and unresolved questions; silence is not evidence of support.

Prepare a communication plan that identifies each audience, the reason for change, the proposed future state, effects on roles, the responsible sender, timing, channels, and feedback routes. Distinguish proposals from authorized decisions.

Assign responsibility for answering concerns and reporting how feedback affected the outcome. Coordinate these activities with project milestones, funding, staffing, and technology readiness rather than adding them after the plan is approved.

Maintain a controlled playbook containing the workflow, role assignments, review questions, templates, evidence rules, escalation routes, and lessons from tested changes. Use it for onboarding and future projects.

Project-specific approved requirements belong in the Program of Requirements (POR) and supporting records; related commissioning expectations may be recorded in an Owner’s Project Requirements (OPR) document.

3.0 Competitive Advantage or Public Value

Competitive advantage is the strategic purpose, not an automatic consequence of installing software or producing analysis faster. The Process chapter connects facilities choices to the organization’s competitive context; Organization connects strategy to the capabilities needed to execute it.

Use those connections to explain why the proposed improvement matters, how it changes the organization’s work, and which outcomes would demonstrate value. For mission-driven and public-sector organizations, apply the same discipline to service effectiveness, continuity, transparency, and responsible use of resources.

Record a traceable value chain for each proposed improvement: strategic need → required capability → organizational and technology changes → better decision or approved requirement → intended outcome → review evidence. Assign an owner for monitoring the intended benefit. Separate expected benefits from demonstrated results and examine implementation cost, dependencies, and trade-offs in the business case.

  • ·       Synergy: Combine portfolio evidence and coordinate business-unit demand to test shared capacity, consolidation, or common services. Review lifecycle cost and functional fit without assuming that consolidation preserves service quality.
  • ·       Differentiation: Translate distinctive customer, service, research, or operating requirements into location, infrastructure, workplace, and performance criteria. Use technology to organize and test these requirements, not to imitate competitors’ amenities.
  • ·       Focus: Prioritize the units, locations, and capabilities that support the selected strategy. Compare expansion, optimization, contraction, and staged responses against common criteria; a merger or acquisition may require several responses at once.
  • ·       Skills: Build cross-functional planning and review capability through shared work, mentoring, and documented playbooks. Retain reasoning and methods so subsequent teams can use organizational knowledge without relying on informal recollection.
  • ·       Technology: Connect authorized evidence, validated calculations, and human-reviewed synthesis to the decisions that matter. Test whether the combination improves option quality and adaptation while reducing avoidable reconciliation, errors, and rework at an acceptable total cost.

Assess these responses as alternatives. An apparent saving may shift costs to fit-out, relocation, disruption, maintenance, or end-of-term obligations.

A faster recommendation may require more verification or omit a critical requirement. Compare the whole commitment and retain flexibility where uncertainty justifies staged action.

4.0 Build an Integrated Decision-Support System

Configure the four logical tiers described in Decision-Support System.

  • ·       The Ingestion & Data Foundation Tier prepares permitted evidence with shared identifiers, definitions, provenance, and validation status.
  • ·       The Processing & Intelligence Tier (AI, Software & Analytics) applies documented methods to scenarios and comparisons, with specialists checking assumptions, calculations, and material AI claims.
  • ·       The Governance & Decision-Support Tier routes reviews, mandatory concurrence, exceptions, and recommendations under confirmed decision rights and records authorization separately.
  • ·       The Output, Records & Lifecycle Continuity Tier retains the living business case, requirements, versions, decisions, conditions, and controlled handoffs, with later evidence informing future planning.

Components describe what the system uses; tiers describe responsibilities and information handoffs. Neither is a set of additional predesign phases.

Integration and automation operate across the tiers, and human verification applies at every consequential handoff. Assign an owner and acceptance criteria for each interface rather than requiring four separate applications or one centralized database.

Use the four pillars in Foundations to select the simplest combination that supports the redesigned workflow. Prepare a system brief stating the decision and users served, approved sources, record owners, information exchanges, tools, permissions, validation methods, review points, support responsibilities, and limits.

Distinguish the technology that supports predesign analysis from the IT, AV, network, and building-system requirements of the proposed facility; coordinate both, but do not confuse their scopes or approvals.

4.1 Use AI: Interpretation and Draft Assistance

Use AI where interpreting supplied material helps a defined task: finding relevant evidence, drafting themes from interviews, identifying candidate requirements or conflicts, and preparing option narratives. Define the permitted sources, expected contribution, reviewer, and acceptance checks before use.

Retrieval-augmented generation (RAG) retrieves reference material to support an AI response, but does not guarantee complete retrieval, accurate interpretation, or valid citations. Preserve retrieved source references, versions, and qualifications; check material claims against the underlying records and report missing or conflicting evidence.

A practitioner checks the interpretation and confirms requirements before they enter an approved baseline. Add multi-step agents only when the task requires them and simpler assistance is insufficient; begin with read-only access, draft outputs, explicit stop conditions, and human review.

4.2 Software: Integration, Automation, and Validated Analysis

Use software to connect compatible records and interfaces, execute approved routine actions, manage review status, and perform spatial, financial, or environmental analysis. Build each reusable decision model from separately identified observed data, reference benchmarks, adjustable assumptions, calculation rules, and evaluation criteria.

Version these layers and retain the inputs, method, and results for each material scenario. Run defined calculations in the analytical service rather than relying on a language model to generate numerical answers; AI may propose questions or explain checked outputs.

Validate definitions, mappings, calculation methods, units, assumptions, and exceptions, and document uncertainty for statistical forecasts or optimization methods. Use rule-based automation for predictable tasks such as missing information flags and review reminders.

Use spatial tools or dashboards only where the decision needs them. Repeatability, integration, or an attractive display does not establish correctness: reviewers must check both the inputs and the method.

4.3 Data: Authoritative Records and Controlled Evidence

Begin by assigning the governing record and source owner for each decision use, not by choosing a new storage platform. Identify which enterprise and lifecycle systems hold the authoritative records for strategy, workforce assumptions, leases, costs, utilization, property constraints, and prior outcomes.

Agree consistent identifiers and definitions from portfolio to asset level. Distinguish workforce headcount, attendance, occupancy, utilization, and required capacity; align their periods, locations, populations, and assumptions before comparison.

Keep source records separate from analytical outputs, and record owner, date, version, permissions, and validation status. An existing enterprise data lake may provide curated evidence but is optional and does not make its content authoritative.

A common data environment controls information status, review, sharing, and versions; it is not synonymous with a data lake or necessarily one application. Use only necessary authorized personnel information.

Check external evidence for authority, currency, jurisdiction, suitability, and reuse terms; public availability does not establish accuracy or permission to reuse.

4.4 People: Judgment, Accountability, and Authorization

Assign people to frame the problem, establish priorities, interpret uncertainty, validate evidence and calculations, reconcile stakeholder differences, recommend alternatives, and make authorized decisions. Preserve the recommendation separately from the decision record.

Where leadership departs from the recommendation, retain the original analysis and document the accountable authority, reasons, accepted risks, safeguards, monitoring indicators, and reassessment conditions.

4.5 Apply Govern, Guide, Guard

  • ·       Govern: Establish the purpose, accountable system and process owners, risk classification, procurement and authorization requirements, monitoring, and oversight.
  • ·       Guide: Define approved and prohibited uses, information protection, documentation, human review, training, and support. Explain what users must verify and where to escalate uncertainty.
  • ·       Guard: Enforce access and action limits through permissions, testing, verification, logging, security, stop conditions, and mandatory human intervention for consequential decisions. Test the controls before use and after material changes.

Choose methods that fit the assignment’s scale, complexity, uncertainty, and proposed commitment rather than deploying the entire catalogue.

  • ·       In Strategic Planning, connect existing-condition assessment and gap analysis to scenario planning, options appraisal, lifecycle value, and the living business case.
  • ·       In Real Estate, connect search criteria, market assessment, specialist due diligence, and property appraisal to transaction readiness and controlled handoff.
  • ·       In Project Definition, connect interviews, surveys, observation, functional relationship analysis, requirements prioritization and traceability, and scope–budget–site–schedule reconciliation to the POR and readiness package.
  • ·       Apply change-management planning, shared-meaning workshops, apparent-agreement checks, and coordinated communication alongside these methods.

Use conventional methods as the baseline and add AI assistance where its contribution provides a substantive benefit.

5.0 Principles for Implementation

5.1 Start with the Decision

Define the decisions before choosing tools. Identify the organizational goal, the alternatives, the limits, and the person authorized to decide. Agree which measures matter, such as space efficiency, adjacency, lifecycle cost, carbon impact, and occupant flow.

Strategic predesign encompasses Strategic Planning, Real Estate, and Project Definition, upstream of design and construction. Late recognition of space needs can allow market availability to drive decisions instead of business requirements. The system should help determine whether—and how—an initiative should proceed, rather than assume that a new building or a real estate transaction is the answer.

Implementation implication: For each decision, record its owner, evidence, criteria, alternatives, and output. Assess uncertainty early: use automated extraction for clear, documented requirements and workshops where needs or priorities are unclear.

5.2 Integrate the Process, Systems, and Data

Treat AI as one component of an operating model. A conversational interface alone does not establish a reliable decision-support system.

Growth projections, lease obligations, occupancy indicators from sensors and badge records, and financial metrics can remain in disconnected systems. Manual reconciliation can leave baseline assessments incomplete. An integrated model connects these sources to the decision workflow: data supplies evidence, software applies validated rules and calculations, automation executes defined tasks, AI supports retrieval and synthesis, and people interpret results and authorize decisions.

Implementation implication: Map the existing predesign workflow, information exchanges, and software interfaces first. Define which tasks require AI interpretation, rule-based automation, or validated calculation engines. Integrate data into existing transactional and operational activities, including broker correspondence, tour feedback, and proposal-term summaries, to build portfolio intelligence without unnecessary administrative work.

5.3 Establish Trusted, Traceable Evidence

Ground analysis is approved, authoritative information, and make the basis of each consequential conclusion inspectable.

Use shared definitions and identifiers for buildings, spaces, leases, costs, and authorized personnel data. Organize historical project benchmarks, space ratios, costs, post-occupancy findings, site information, and applicable regulatory requirements. Record stakeholder needs, culture, and mission requirements in searchable categories. Connected, governed sources can serve this purpose; one central database is not required. Source links and consistent definitions help reviewers check evidence but do not guarantee accuracy.

Implementation implication: Record source, date, version, ownership, and verification status. Distinguish documented facts, assumptions, calculated results, and AI-generated interpretations. If evidence is missing or contradictory, the system should identify the gap rather than manufacture certainty.

5.4 Govern Internal and External Information Differently

Combine protected enterprise knowledge with selected public information, but do not treat either category as automatically reliable.

Unmanaged use of AI outside approved organizational controls, often called “shadow AI,” can expose private information and intellectual property and weaken the audit trail. Combine a governed enterprise information environment with selected external sources for market, demographic, regulatory, and benchmark evidence. Protected internal records and public information require different access and verification controls; neither is automatically reliable.

Implementation implication: Establish access permissions and rules for admitting external evidence. Confidential information should remain protected, while public information should be evaluated for authority, relevance, and currency.

5.5 Use Agentic Capabilities Selectively and Within Boundaries

Apply multi-step AI workflows where they improve a defined process—not simply because autonomy is available.

Use AI to find information, extract draft requirements from interviews and briefs, and compare evidence. People must check the extracted requirements before they become approved inputs. Use rule-based automation for routine tasks and validated software for spatial, financial, and environmental calculations. Where needed, use parametric test-fitting tools or portfolio models to compare massing, adjacency, and layout options. Choose capabilities for the decision at hand; do not deploy them all by default.

Implementation implication: Start with a bound task and clear review points. Specify what the system may access or change, when it must stop, and who reviews exceptions. Make calculation methods, inputs, and limits visible.

5.6 Preserve Human Authority and Accountability

Keep responsibility for strategic choices with identifiable people.

“AI acts within authorized boundaries; people decide” applies to organizational priorities, stakeholder values, investment, risk, and long-term commitments. Executive approval alone does not ensure operational readiness: conflicting business-unit priorities, misaligned incentives, inadequate capability development, and unclear ownership can delay execution or reduce value. Human governance must connect strategic approval with practitioner validation, operational responsibilities, and stakeholder deliberation.

Implementation implication: Assign decision owners and approval gates. AI may assemble evidence and compare alternatives; people should establish priorities, resolve trade-offs, and authorize commitments.

5.7 Compare Alternatives and Reveal Uncertainty

Design the system to support deliberation rather than merely producing a persuasive recommendation.

Business cases become outdated as economic conditions, work policies, staffing, or operations change. Keep decision models reviewable and compare options using current evidence—for example, lease exit costs, consolidation savings, and commute impacts. A comparison view can show how changed assumptions affect results; an interactive dashboard is optional. Faster analysis must not bypass approval or public deliberation.

Implementation implication: Apply the same criteria to each option, disclose assumptions, and test whether changed conditions alter the conclusion. Include no-project or defer options where appropriate. Compare short-term budgets and staffing changes with actual lease, fit-out, and facility commitments rather than relying on generic planning horizons.

5.8 Preserve Information Continuity Across the Lifecycle

Create reusable, structured predesign information rather than an isolated final report.

Information is often rebuilt at phase transitions, making institutional knowledge difficult to retain. Carry predesign evidence and reasoning into design, construction, operations, and future portfolio planning. Continuously updated portfolio models can connect space-planning rules, lease benchmarks, and headcount assumptions across the asset lifecycle while retaining dated records of the evidence used for each approved decision.

Implementation implication: Carry forward approved requirements, assumptions, constraints, decision criteria, and decision history using consistent identifiers and version control. Continuity need not mean one monolithic database; it means that information remains connected and interpretable.

5.9 Capture Reasoning and Learn from Outcomes

Retain why a decision was made—not just its final answer.

A searchable record of prior reasoning helps teams reuse knowledge rather than rediscover it. Capture work processes, assumptions, criteria, alternatives, stakeholder reasoning, lessons learned, and outcomes. For portfolio planning, connect supply and financial commitments with observed occupancy and utilization and qualitative employee experience. This combined evidence can improve demand assessments beyond predetermined space-entitlement ratios alone.

Implementation implication: Maintain a decision record linking evidence, alternatives, rationale, approval, and subsequent results. Use later outcomes to improve future predesign analysis.

5.10 Redesign the Work Process with Its Practitioners

Implementation is an organizational change effort, not merely software deployment.

Consistent data protocols, practitioner involvement, small pilots, gradual expansion, and cost-benefit monitoring support process change. Address conflicting priorities and operational readiness as well as software adoption. For public asset delivery, incorporate early community engagement, applicable legislative and zoning requirements, and transparent scenario evaluation to support public trust and continuity through multi-year leadership transitions.

Implementation implication: Assign a process owner and involve practitioners, executives, and affected stakeholders. Use review workshops to test options against operational and cultural needs, resolve conflicts, and record agreed changes. Train users to check evidence and assumptions, not simply accept fluent outputs.

5.11 Pilot Narrowly and Measure Decision Value

Start with a manageable decision-support use case and scale only after demonstrating value.

Test whether the system improves evidence retrieval and option preparation without increasing errors or review effort. Measure quality, total effort, cost as well as speed. A short pilot can show immediate value, but later project and operating results are needed to assess long-term benefits.

Implementation implication: Measure more than usage and time saved. Evaluate evidence completeness, traceability, errors, rework, usefulness of alternatives, and downstream changes caused by missed requirements. Establish a baseline before the pilot.

5.12 Requirements to Define and Validate through Implementation

  • A decision map and approval structure for the three predesign phases.
  • ·       A shared information model, software interfaces, and rules for source verification, refresh, and calculation validation.
  • Agent permissions, review gates, and exception handling.
  • A method for comparing alternatives and expressing uncertainty.
  • Pilot acceptance criteria and measures of decision quality.
  • Ownership of lifecycle information and feedback from actual outcomes.
  • ·       Required organizational capabilities, cross-functional handoffs, staffing, training, change communication, and benefit-monitoring owners.
  • ·       A controlled playbook and living business case with review intervals, material-change triggers, and a record distinguishing recommendations from authorized decisions.

The six-stage implementation roadmap below assigns work, responsibilities, and improvement gates for defining, testing, and maintaining these requirements. These implementation stages are distinct from the three predesign phases and their investment and project authorizations.

6.0 Apply Improvements through the Predesign Phases

The implementation stages below improve how strategic predesign is performed; they do not replace its three phases or create permission to commit funds, transact property, award contracts, or begin design. Apply the common Procedures groups throughout: problem framing; participation and coordination; evidence and information governance; risk; alternatives and business case; and authorization, change control, and learning.

Adapt the inquiry to uncertainty and the proposed commitment.

6.1 Strategic Planning: Establish Direction and Compare Portfolio Responses

The strategic planner leads Executive Vision, Portfolio Analysis, and preparation of the Business Case. Use the improved workflow to connect mission, competitive or service context, business-unit demand, asset capacity, and financial commitments.

Compare work-process change, technology, use of existing resources, and expansion, optimization, or contraction. Record criteria, scenarios, assumptions, resources, intended benefits, and conditions for sponsor direction.

Maintain the business case as a living decision instrument rather than a one-time persuasive report.

6.2 Real Estate: Test Fit and Control Commitments

The portfolio and asset manager leads location strategy and site-selection criteria, market survey and technical due diligence, and transaction readiness and lease negotiation. Use common definitions and evaluation criteria to compare functional fit, availability, infrastructure, lifecycle cost, timing, and risks.

Qualified advisers validate material technical, environmental, legal, and commercial findings. Negotiation authority is not commitment authority.

Transfer authorized direction, evidence, requirements, commercial terms, risks, and outstanding actions to the receiving project team.

6.3 Project Definition: Reconcile Requirements and Establish Readiness

The project manager coordinates the project management team, translating business priorities into a Program of Requirements (POR), and alignment of delivery and procurement with project objectives. Apply the five programming steps—establish goals, collect and analyze facts, test and evaluate concepts, determine needs, and state the problem—across function, form, economy, and time.

Validate stakeholder and technical requirements, reconcile scope, budget, schedule, and risk, and document the delivery strategy. Assemble the charter, POR, updated business case, and readiness findings for the designated owner’s authorization decision.

Findings may require a return to earlier assumptions or approvals. Where no real estate decision is needed, proceed from Strategic Planning to Project Definition within the authorized mandate.

Where critical uncertainty affects a property commitment, authorize the needed definition work before making that commitment. A readiness review establishes evidence; the authorized owner decision specifies whether and under what conditions work may proceed.

Transaction, procurement, design, and later construction permissions remain separate.

7.0 Implementation Roadmap

Establish a human-governed strategic predesign capability for Strategic Planning, Real Estate, and Project Definition. The operating model combines AI + Software + Data + People, with integration and automation supporting the workflow.

The roadmap addresses organizational alignment, information fragmentation, planning-horizon mismatches, static business cases, and governance risks. Begin with one decision workflow; expand only when organizational readiness and tested results justify it.

Use the principles as operating rules, the six implementation stages as the improvement delivery plan, and the RACI matrix to assign task responsibility. Within the predesign decisions served by the pilot, apply the Organization chapter’s RAPID framework and established phase leadership.

Planning assumption: approximately 24 weeks to establish and evaluate a bounded pilot, followed by controlled expansion. Timing depends on data readiness, security review, integration complexity, and available staff. Advance through approval gates, not simply because a scheduled date has arrived.

7.1 Define the Decisions and Governance

  • ·       Accountable role: Executive sponsor. Delivery lead: Predesign process owner.
  • ·       Actions: Map decisions across Strategic Planning, Real Estate, and Project Definition. Agree goals, alternatives, constraints, decision rights, and measures such as space efficiency, adjacency, lifecycle cost, carbon impact, and occupant flow. Identify the required capabilities and gaps in structure, processes, people, and rewards. Map information handoffs, duplicate work, conflicting priorities, planning-horizon mismatches, and readiness issues with practitioners and stakeholders. Assess uncertainty to decide where document extraction is suitable and where workshops are needed. For public assets, include community engagement and applicable legislative and zoning reviews.
  • ·       Governance: Confirm the executive sponsor, process owner, and existing predesign phase leaders. Assign data stewardship, technical delivery, security review, professional review, and change-management responsibilities. Identify the person authorized to decide each predesign matter and required concurrence under RAPID. Reserve strategic choices and commitments for authorized people. Define escalation when evidence is missing, contradictory, or outside the system’s scope.
  • ·       Deliverables: Implementation charter; decision and capability map with agreed measures and uncertainty assessment; current-workflow and handoff map; responsibility and approval matrix; organizational change and communication responsibilities; initial risk register; baseline measurement plan with benefit owners.
  • ·       Gate 1: Sponsor approves scope, responsibilities, resource commitment, and decision boundaries before system configuration.
  • ·       Principles implemented: 1, 2, 6, and 10.

7.2 Establish the Trusted Information Foundation

  • ·       Accountable role: Predesign process owner. Delivery leads: Data steward and technical lead, with security review.
  • ·       Actions: Inventory strategy documents, portfolio records, historical project benchmarks, space ratios, costs, and post-occupancy evaluations. Include relevant growth projections, leases, financial commitments, occupancy, utilization, and employee experience where authorized. Identify site and geographic data and applicable zoning, building-code, and environmental requirements. Organize stakeholder needs, culture, and mission requirements into searchable categories. Record authoritative versions, gaps, sensitivity, refresh needs, and system interfaces.
  • ·       Evidence rules: Record source, date, version, owner, and verification status. Separate facts, assumptions, calculations, and AI interpretations. Define criteria for admitting external information and reporting unavailable evidence.
  • ·       Deliverables: Source register; shared definitions and identifiers; system-interface inventory; approved pilot sources; access matrix; evidence-checking and refresh rules. Link decisions, options, requirements, assumptions, evidence, approvals, and outcomes. Connect existing authoritative systems where practical; a single database is not required.
  • ·       Gate 2: Data steward and security reviewer confirm that pilot sources are sufficiently complete, traceable, current, and authorized. Resolve critical gaps or narrow the pilot before proceeding.
  • ·       Principles implemented: 3, 4, 8, and 9.

7.3 Design a Bounded Pilot

  • ·       Accountable role: Predesign process owner. Delivery leads: Practitioner pilot team and technical lead.
  • ·       Recommended starting use case: assemble evidence and compare alternatives for one bounded, repeatable Project Definition decision, such as reconciling a departmental capacity requirement with available space and budget before POR approval. Name the decision authority and select the relevant methods—checked demand and facts, needs analysis, requirements traceability, and scope–budget–site–schedule reconciliation—rather than testing an unrestricted assistant.
  • o   Define the minimum evidence, expected comparison output, specialist reviewers, and stopping criteria. Use a historical case with known results and a current case suitable for supervised testing.
  • o   Confirm case availability rather than assuming it. Keep pilot outputs separate from approved requirements and commitments until the applicable review and authorization have occurred.
  • ·       Actions: Define inputs, retrieval, calculations, output templates, review points, and exception handling. Include extraction of draft requirements from interviews, client briefs, and strategy documents, with human confirmation before use. Select only the spatial analysis needed for the case: parametric test-fitting or portfolio models may compare massing, adjacency, and layouts. Compare options against approved measures, include defer or no-project options where appropriate, and show assumptions and sensitivity. Start with a simple comparison view; specify an interactive dashboard only if changing constraints is important to the pilot.
  • ·       Agent boundaries: Begin with read-only information access and draft outputs. No autonomous approvals, external commitments, or source-record changes. Define stop conditions and escalation for inadequate evidence or conflicting requirements.
  • ·       Deliverables: Redesigned pilot workflow; system brief; reviewed requirements template; alternatives-comparison template; tool and calculation specifications; agent permissions; test cases; decision-record template; baseline results; proposed acceptance thresholds; staffing, training, support, and communication plan.
  • ·       Gate 3: The authorized decision owner approves test scope, review requirements, and measurable acceptance thresholds before testing. Confirm practitioner availability, specialist review, change support, and the intended value to be evaluated. This is approval of a supervised test, not approval of its eventual recommendation or an investment commitment.
  • ·       Principles implemented: 1, 5, 6, 7, and 11.

7.4 Configure, Test, and Train

  • ·       Accountable role: Predesign process owner. Delivery lead: Technical lead, supported by data stewardship, security, and practitioner reviewers.
  • ·       Actions: Configure source-linked retrieval, access controls, draft requirement extraction, decision logs, and the selected calculation or spatial tools. Connect only the systems needed for the pilot. Use automation for routine tasks and AI for evidence synthesis and draft comparisons. Add multi-step agents only if simpler workflows are insufficient. If selected, test a dashboard that shows how changed constraints affect options, including its limits and response time. Test routine data capture to avoid duplicate administration.
  • ·       Testing: follow a representative decision from source evidence through analysis, recommendation, authorization, and controlled handoff. Check retrieval, source links, extracted requirements, shared definitions, interfaces, calculation methods, spatial constraints, option comparisons, permissions, and repeatability.
  • o   Reproduce material calculations and compare outputs with source records and independently reviewed baseline cases. Test missing and conflicting evidence, failed retrieval, incorrect calculations, changed inputs, and attempted out-of-scope actions.
  • o   Verify that dependency notifications flag affected working outputs without overwriting approved baselines; that versions survive; and that review routing, mandatory clearances, approval thresholds, and signature limits remain distinct. Check that controls and incident procedures address unauthorized AI use and data access, including misleading instructions in retrieved material.
  • o   Confirm that the handoff contains the actual authorization, conditions, responsible parties, and checkpoints. Record defects, corrections, and retest results.
  • o   Repeatable results are not necessarily correct, and a completed dashboard or checklist does not establish readiness or approval.
  • ·       Review and training: use multidisciplinary workshops to test options against operational, cultural, and stakeholder needs and to check apparent agreement. Programmers lead requirements analysis; the project manager coordinates decisions, issues, actions, and formal change control.
  • o   The assigned change manager supports participation, shared understanding, practical concerns, and readiness with HR and Communications, recording contributions in existing project records rather than a parallel system. Resolve conflicts through the agreed decision process and record the rationale, dissent, and outstanding actions.
  • o   Adjust scoring weights, constraints, rules, or prompts only after review; version the changes and rerun affected tests. Train users to check evidence, interpret uncertainty, record overrides, and escalate exceptions.
  • o   Pair incoming practitioners with experienced leads and test handoffs. Communications coordinates approved messages with the change manager and project manager; use the communication plan to explain changing roles and feedback routes, distinguish proposals from decisions, and check stakeholder understanding.
  • ·       Deliverables: Working pilot; test report and corrected defects; workshop decisions and versioned rule changes; controlled workflow playbook and user guidance; support, fallback, and incident procedures; trained pilot team; confirmed staffing and operational readiness.
  • ·       Gate 4: Process owner, technical lead, and security reviewer authorize supervised live use after required tests pass and critical defects are resolved.
  • ·       Principles implemented: 2, 3, 4, 5, 6, and 10.

7.5 Evaluate Decision Value

  • ·       Accountable role: Executive sponsor. Delivery lead: Predesign process owner, supported by decision owners and reviewers.
  • ·       Actions: Run the bounded workflow under human supervision. Compare results with the baseline using equivalent cases and criteria. Record review effort, corrections, conflicts, exceptions, overrides, and final decision rationale. Use stakeholder feedback to identify improvements but revalidate changes before relying on revised outputs.
  • ·       Evaluate: Evidence completeness and traceability; unsupported claims and errors; quality of alternatives and uncertainty analysis; preparation time plus human review time; rework; usefulness to decision makers; operating and support cost. Review cross-functional handoffs, unresolved concerns, training and support adequacy, and whether the workflow supports the required capability. Benefit owners distinguish observed pilot results from projected financial, operational, or strategic outcomes. Do not equate adoption or fluent output with decision quality.
  • ·       Deliverables: Evaluation report; cost-benefit assessment; updated risk register; prioritized improvements; recommendation to expand, revise, pause, or stop.
  • ·       Gate 5: Sponsor approves expansion only if pre-agreed quality and control thresholds are met and the pilot demonstrates useful decision support at an acceptable cost. Short-term results cannot establish long-term project benefits; track those later.
  • ·       Principles implemented: 6, 7, 9, and 11.

7.6 Expand and Connect Lifecycle Learning

  • ·       Accountable role: Executive sponsor for expansion; predesign process owner for continuing operation. Delivery leads: Data steward, technical lead, and receiving lifecycle teams.
  • ·       Actions: Add one workflow or source at a time and recheck quality, access, and support capacity after material changes. Connect project-management, BIM, and enterprise platforms through approved interfaces where needed. Extend from project definition into real estate and strategy as evidence and governance permit. Maintain portfolio models linking supply, commitments, occupancy, employee experience, and business outcomes. Set refresh intervals and review triggers for assumptions, benchmarks, market changes, applicable codes, and technical standards. Retain approved briefs and versioned decisions rather than overwriting them.
  • ·       Lifecycle handover: Transfer approved requirements, assumptions, constraints, alternatives considered, evidence, approvals, and rationale into downstream information systems with consistent identifiers and versions. Identify who receives and maintains each record.
  • ·       Learning: Record differences between predicted and actual results as design, construction, and operations provide evidence. Review the causes and curate lessons for reuse. Audit evidence, methods, controls, and usefulness regularly; approve and test updates to data, rules, tools, or models before release. Do not treat automatic model adaptation or unreviewed AI output as trusted knowledge.
  • ·       Deliverables: Expansion backlog; repeatable deployment package and controlled playbook; approved integration and lifecycle handover rules; maintained decision repository; outcome-feedback log linked to benefit owners; operating ownership, onboarding materials, and review schedule.
  • ·       Gate 6: Receiving teams confirm usable handover and continuing ownership. Further releases require quality checks and explicit authorization; scaling is not automatic.
  • ·       Principles implemented: 2, 8, 9, 10, and 11.

7.7 Pilot Scorecard and Acceptance Rules

  • ·       Evidence quality: Track required evidence available, source traceability, unsupported claims, and factual or calculation errors. Proposed minimum control: every consequential factual claim must have a verified source or an explicit unresolved-evidence flag.
  • ·       Decision usefulness and strategic fit: use an agreed rubric to assess whether the selected technique answered the decision-relevant question and whether its inputs and method were suitable for the commitment under review. Review approved requirements, complete alternatives, consistent criteria, assumptions, sensitivity, and fit with stakeholder needs.
  • o   Trace the recommendation to the organizational need and required capability, and check that decision makers understand material trade-offs, limitations, and unresolved conditions. Record unresolved conflicts, dissent, and how authorized people resolve them; preserve any departure from the recommendation and its rationale.
  • o   Report method and quality failures alongside benefits rather than treating more scenarios, fluent explanations, or a completed template as evidence of better judgment.
  • ·       Efficiency and cost: Compare total elapsed and staff time, including verification and correction, with the baseline. Include configuration, operation, training, and support costs; agree improvement targets at Gate 3.
  • ·       Governance: Check approval records, permission enforcement, incident handling, and overrides. Proposed minimum controls: every final decision has an authorized human approval record; no unresolved critical access or confidentiality defects before live use or expansion.
  • ·       Downstream results: Benefit owners compare actual costs, timing, operational performance, employee or service experience, and requirement-related rework with the approved business case when later phases provide evidence. Record changed conditions and causes of differences before attributing results to the improvement. Feed reviewed findings into benchmarks, rules, and playbooks; do not claim long-term advantage from a short pilot alone.
  • ·       Organizational readiness: Confirm assigned role holders, practitioner and specialist capacity, tested handoffs, training completion, support and fallback arrangements, and functioning feedback and escalation routes. Set case-appropriate acceptance criteria at Gate 3; tool access or training attendance alone does not demonstrate readiness.
  • ·       Review cadence: Hold weekly pilot issue reviews, formal approval at each gate, and a quarterly operating review after deployment. Adjust cadence to the initiative’s risk and scale.

7.8 Roadmap RACI Matrix

This proposed role-based matrix assigns tasks and implementation Gates 1–6. Confirm role holders and delegated authority at Gate 1.

R — Responsible: performs the work. A — Accountable: owns completion or the specified implementation decision; each row has exactly one A. C — Consulted: provides input before completion. I — Informed: receives status or results. A/R: owns and performs the activity. —: no routine assignment.

Required specialist signoffs remain mandatory even where the role is marked C; accountability does not permit bypassing them.

Role key: ES = Executive sponsor; PO = Predesign process owner; DO = Authorized decision owner for the relevant Strategic Planning, Real Estate, or Project Definition decision; DS = Data steward; TL = Technical lead; SR = Security reviewer; PT = Practitioner pilot team, including the relevant phase leader, professional reviewers, stakeholder facilitation, and assigned HR, Communications, Finance, Legal, Procurement, and operations contributors; LT = Receiving lifecycle team, represented by a designated receiving owner.

Assign an individual role holder for each A, including where a team is represented. Identify change-management and benefit-monitoring owners within these assignments rather than leaving them implicit. AI/agents support responsible people; they hold no accountability or approval authority.

RACI Matrix

Stage / activity or approval

ES

PO

DO

DS

TL

SR

PT

LT

1. Charter, goals, measures, uncertainty, roles, decision map, baseline plan

A

R

C

C

C

C

R

C

Gate 1. Approve scope, resources, and decision boundaries

A

R

C

I

C

C

I

I

2. Source register, benchmarks, site and regulatory evidence, stakeholder needs, data model

I

A

C

R

R

C

C

C

2. Access rules and authorization controls

I

A

C

R

R

C

C

C

Gate 2. Confirm evidence readiness and authorized access

I

A

C

R

C

R

C

I

3. Pilot cases, reviewed requirements, comparison tools, test design, baseline

I

A

C

C

R

C

R

C

3. Agent permissions, stop conditions, escalation rules

I

A

C

C

R

C

R

I

Gate 3. Approve test scope, review rules, acceptance thresholds

I

R

A

C

C

C

R

I

4. Configure retrieval, extraction, selected tools, integrations, controls, logs

I

A

I

R

R

C

C

I

4. Validate evidence, calculations, permissions; resolve defects

I

A

C

R

R

R

R

I

4. Review workshops, conflict resolution, rule changes, training and support

I

A

C

C

R

C

R

I

Gate 4. Authorize supervised live use after required sign-offs

I

A/R

C

C

R

R

C

I

5. Supervised pilot, scorecard, cost-benefit report, recommendation

A

R

C

R

R

C

R

C

5. Approve each strategic choice and record final rationale

I

C

A/R

C

C

C

R

I

Gate 5. Decide expand, revise, pause, or stop

A

R

C

C

C

C

C

C

6. Expansion backlog, approved integrations, deployment package, funding proposal

A

R

C

R

R

C

C

C

6. Prepare lifecycle handover and continuing ownership records

I

A

C

R

R

C

C

R

Gate 6. Accept usable handover and receiving ownership

I

R

C

R

R

C

C

A

6 / ongoing. Authorize funded releases after quality revalidation

A

R

C

R

R

C

C

C

Ongoing. Source refresh, support, incidents, weekly issue reviews

I

A

C

R

R

R

R

C

Ongoing. Compare predicted and actual outcomes, curate lessons, maintain repository

I

A

C

R

C

C

R

R

Ongoing. Quarterly review of value, evidence, rules, tools, and controls

C

A/R

C

R

R

C

C

C

8.0 Implementation Outcome

The intended result is improved organizational capability: accountable people using repeatable procedures, connected evidence, and appropriate tools to make better decisions. Competitive advantage or public value depends on how well the organization puts these decisions into practice and learns from outcomes.

Prepared by Steven Parshall, with the use of AI agents. © Copyright of the Strategic Predesign Association


Feedback on this section

Click on POST to share your thoughts, ideas, comments, and constructive critiques about this section. Tell us what is unclear, what could be improved, and what additional topics we should address.

© 2026 Strategic Predesign Association

Domain: pred-int.org

Powered by Wild Apricot Membership Software