Menu
Log in



Log in

9 System

Decision-Support System = Data + Software + AI + Judgement


Introduction

1.0 Built Environment Information Lifecycle

2.0 Enterprise Data Lake

3.0 Organization as an Information-Processing Network

3.1 Task Uncertainty and Interdependent Requirements

3.2 Coordination, Decision Rights, and Escalation

4.0 Organizing to Reduce Uncertainty and Equivocality

4.1 Reducing Equivocality

4.2 Three Interrelated Organizing Processes

5.0 Sensemaking Framework

5.1 From Equivocality to Structured Shared Meaning

5.1.1 Applying Enactment

5.1.2 Applying Selection (Retrospective Sensemaking)

5.1.3 Applying Retention

5.2 Conflict Resolution and Decision-Making

5.2.1 Disentangling Uncertainty from Equivocality in Trade-offs

5.2.2 Retrospective Sensemaking for Shared Understanding

5.2.3 Balancing Retention with Adaptability

5.3 Identifying Decisions

5.3.1 Connecting Evidence to Decisions

6.0 Decision-Driven Workflow

6.1 Matching Information Demand with Processing Capacity

7.0 Information-Processing Capacity

7.1 The Role of a Decision-Support System

8.0 Decision-Support System

9.0 System Components

9.1 Software Services

9.2 Internal and External Data Sources

9.3 AI Applications

9.4 Human Judgment and Source-Verification Interfaces

10.0 System Architecture

10.1 Integration and Automation Across the Architecture

10.1.1 Integration Requires Shared Meaning and Controls

10.1.2 Automation Is Not Synonymous with AI

11.0 Tier Responsibilities and Information Handoffs

11.1 Ingestion & Data Foundation Tier

11.2 Processing & Intelligence Tier (AI, Software & Analytics)

11.3 Governance & Decision-Support Tier

11.4 Output, Records & Lifecycle Continuity Tier

11.5 From Source Evidence to Authorized Records

12.0 Lifecycle Interface and Digital Twins

13.0 AI Governance: Govern, Guide, Guard

Introduction

A strategic predesign decision-support system helps connect organizational needs to evidence, evaluated alternatives, and accountable decisions before facility commitments are made. This chapter explains its use in Strategic Planning, Real Estate, and Project Definition. It first examines how organizations process information and interpret requirements, then describes four logical tiers linking software, data, AI assistance, and human verification across the built-environment lifecycle. The implementation approach follows from these principles: begin with a consequential owner decision, redesign the workflow, test a bounded pilot, and measure decision quality as well as speed. The central principle is to match organizational information-processing capacity to decision needs.

1.0 Built Environment Information Lifecycle

The built-environment information ecosystem provides the context for the strategic predesign decision-support system. It spans three connected stages: strategy and planning establishes direction and requirements; design and construction delivers authorized work; and operations and management provides asset records and performance evidence. Information flows between these stages to support decisions and learning throughout the lifecycle.

Strategic predesign comprises three phases: Strategic Planning, Real Estate, and Project Definition. It draws on information across the ecosystem, including historical design and construction costs, schedules, building-system performance, occupancy, and maintenance. Predesign concludes with a documented decision following the Project Readiness Review and, where design is authorized, a controlled handoff. Information may remain in several authoritative enterprise applications, a single platform, or a shared information repository; the system does not require every record to be centralized.

2.0 Enterprise Data Lake

Where an enterprise maintains a data lake, the decision-support system may connect to it as an additional evidence source rather than create a separate store for every dataset. A data lake holds structured, semi-structured, and unstructured information from multiple systems, including raw records and prepared datasets. Relevant predesign evidence may include workforce forecasts, asset inventories, lease information, historical costs, occupancy observations, and supporting documents. Approved connectors, query services, or scheduled transfers make selected information available to the Ingestion & Data Foundation Tier.

The interface should favor curated datasets—information prepared with documented definitions and quality checks. Raw information requires preparation and validation before analysis. Shared assets, location, and project identifiers connect records across sources; source references, observation dates, refresh schedules, and transformation histories make findings traceable. Access follows enterprise permissions and data-protection rules, and retrieval is limited to the authorized purpose.

Presence in a data lake does not establish that a record is authoritative, current, or suitable for a decision. Responsible specialists verify those qualities and reconcile conflicting definitions. For example, workforce forecasts and occupancy observations can inform space-demand scenarios, but they remain distinct measures and require compatible periods and assumptions. Where permitted, selected analytical outputs and links to controlled decision records can be returned to the enterprise data platform for reporting and later performance comparison. Approved business cases, requirements, and authorizations remain governed records: new evidence may prompt a proposed revision but does not automatically change an approved baseline. The connection is optional, complements direct source-system interfaces, and does not add a fifth logical tier or require every record to be centralized.

Strategic predesign is an information-processing and decision-making process. Its purpose is to establish organizational needs, identify uncertainty, and reconcile interdependent requirements before they become design and construction commitments. Understanding how an organization acquires and communicates information, coordinates work, and resolves competing priorities provides the basis for using a decision-support system during predesign.

The next sections establish the organizational requirements that the architecture must serve. Information-processing theory explains how coordination capacity responds to uncertainty and interdependence; sensemaking explains how participants develop workable interpretations of requirements. Read together, they frame three questions for system design: what evidence is missing, what meanings differ, and who has authority to resolve competing priorities? The later components, tiers, and workflows translate these questions into operating responsibilities.

3.0 Organization as an Information-Processing Network

An enterprise processes information to coordinate people, resources, and work across interdependent groups. Direct communication may be sufficient for a simple task. As the number of participants and the degree of uncertainty increases, the organization needs additional mechanisms to integrate activities and maintain performance.

Rules and procedures coordinate routine work by defining expected responses to familiar situations. Like individual habits, they reduce the need to treat each situation as new and provide institutional memory for recurring tasks. Rules and procedures are less effective when situations cannot be anticipated. An unresolved exception is referred through the hierarchy until it reaches a manager with sufficient authority over the affected work. Each link has finite information-processing capacity. When referrals exceed that capacity, decisions are delayed and the response back through the hierarchy slows.

Delegating authority and responsibility allows more decisions to be made closer to the relevant information. Training and professional expertise support informed judgment, while shared goals and performance targets align otherwise independent groups. Managers supervise coordination and retain escalation responsibilities for issues beyond delegated authority. Delegation alone may not accommodate increasing uncertainty and information demand. When processing requirements exceed capacity, an organization must reduce the need for information processing, increase its capacity, or combine both approaches.

3.1 Task Uncertainty and Interdependent Requirements

The organizational information-processing model defines task uncertainty as the gap between the information required to perform a task and the information already available to the organization. Rules, standard operating procedures, budgets, schedules, and performance targets reduce case-by-case coordination in routine work; they do not eliminate communication. Novel or complex tasks, such as defining requirements for a new building, introduce choices that these mechanisms may not anticipate. Because specialized units are interdependent, a requirement established by one unit can affect costs, schedules, or performance elsewhere:

  • ·       A business unit leader may require more precise space-allocation criteria.
  • ·       A facilities manager may identify higher occupancy costs or schedule delays associated with meeting those criteria.

These perspectives create an exception when existing rules or delegated authority cannot reconcile their implications. An exception may involve missing information, competing priorities, or both. Additional evidence can clarify the consequences of available choices, but it cannot by itself determine which organizational priority should prevail. Resolving that conflict also requires agreed decision criteria and an accountable project sponsor or executive with the appropriate authority.

3.2 Coordination, Decision Rights, and Escalation

Resolve an issue where the relevant information exists, provided the decision falls within established decision rights—the authority assigned to make decisions. Lateral coordination helps affected units examine alternatives and reconcile requirements through direct contact, liaison roles, cross-functional task forces, integrating roles such as product managers, and matrix structures. These mechanisms supplement rather than bypass hierarchical authority. If an issue exceeds delegated authority, refer it to the lowest shared level with sufficient authority over the affected units and requirements: a shared manager, project sponsor, or executive authorized to decide. A conflict between business units may go to their immediate shared manager. A conflict spanning facilities and workplace operations, HR, IT/AV, and Legal may require a project sponsor, general manager, or other authorized executive, depending on delegated decision rights.

4.0 Organizing to Reduce Uncertainty and Equivocality

Information-processing capacity concerns how much information an organization can handle; sensemaking concerns how participants interpret it. These complementary frameworks address different problems: missing information creates uncertainty, while multiple plausible interpretations create equivocality. The sensemaking framework describes three interrelated organizing processes—enactment, selection, and retention—through which participants develop and carry forward workable interpretations.

Strategic predesign may involve both uncertainty and equivocality. Additional evidence can reduce uncertainty when relevant facts are missing. When stakeholders interpret the same information differently, dialogue and comparison of assumptions are also needed. Neither more data nor a shared interpretation necessarily resolves competing priorities; those conflicts still require decision criteria and appropriate authority.

The following concepts explain how sensemaking supports requirements development and conflict resolution during strategic predesign.

4.1 Reducing Equivocality

Uncertainty concerns insufficient information about a task or situation. Equivocality concerns multiple plausible interpretations of that information. More data may reduce uncertainty; reducing equivocality also requires participants to examine meanings and assumptions. Participants coordinate action by developing workable interpretations of ambiguous situations. The objective is sufficient shared understanding to act, not necessarily complete agreement or the elimination of all ambiguity.

4.2 Three Interrelated Organizing Processes

Enactment, selection, and retention describe interrelated organizing processes, not a rigid project sequence. Actions generate experiences that participants interpret; retain interpretations then influence subsequent action.

Enactment: Participants do not simply receive information from an independent environment. Through their actions, interactions, and attention, they shape the situations and cues they subsequently interpret. In predesign, exploring an option can reveal assumptions and consequences that abstract discussion alone leaves unclear.

Selection: Participants interpret what has occurred and develop a workable account of its meaning. This retrospective sensemaking draws on existing interpretive frameworks and communication cycles. A double-interact loop is an act, another participant’s response, and the first participant’s adjustment. Here, selection refers to interpreting experience, not simply choosing a design alternative.

Retention: Interpretations, practices, and rules that have proved useful are carried forward in organizational memory, including causal maps that represent perceived relationships between actions and outcomes. This memory supports future action, but reliance on earlier interpretations can obscure changed conditions.

5.0 Sensemaking Framework

Applied to strategic predesign, sensemaking helps teams develop requirements and examine the assumptions behind conflicts. It complements the information-processing and decision-authority mechanisms described above. Executives, facilities managers, end users, and financial controllers may attach different meanings to the same goal or requirement even when the underlying facts are available.

5.1From Equivocality to Structured Shared Meaning

Requirements development involves both gathering evidence and clarifying meaning. Staff surveys and space-utilization measurements can address missing information. They may not resolve differing interpretations of terms such as “collaborative workspace,” “flexibility,” or “sustainability.” Predesign teams therefore need to establish what these terms mean operationally before translating them into spatial criteria.

5.1.1 Applying Enactment

Stakeholders may find it difficult to articulate needs in abstract terms. Interactive visioning workshops, scenario exercises, spatial prototypes, and preliminary blocking options can make assumptions visible and generate experiences for joint interpretation. Concrete spatial or operational options give stakeholders something specific to examine. Their responses help the predesign team identify implicit assumptions and translate them into explicit feedback.

5.1.2 Applying Selection (Retrospective Sensemaking)

After stakeholders explore scenarios or options, the predesign team compares their responses and the assumptions behind them. Selection develops a workable interpretation of organizational goals and requirements; approval of an alternative remains a separate, accountable decision.

Requirements are refined through communication loops:

  1. Act: The predesign team presents a spatial or operational concept based on initial interviews.
  2. Response: Stakeholders critique the concept and highlight operational conflicts.
  • 3.      Adjustment: The team revises the proposed requirement in response to that feedback, recording any unresolved differences for further review or an authorized decision.

5.1.3 Applying Retention

Retain agreed interpretations, validated requirements, and decision rationale in the Program of Requirements (POR) and supporting space programs, criteria matrices, and design guides. Once approved by the designated owner authority, these records provide a controlled baseline for subsequent design. An Owner’s Project Requirements (OPR) document records owner expectations and performance requirements for commissioning where applicable; it is related to, but not interchangeable with, the POR. Preserve approval status and version history, and route proposed changes through authorized change control.

5.2 Conflict Resolution and Decision-Making

Predesign conflicts may concern cost limits and spatial ambitions, or individual office requests and an open-plan strategy. Sensemaking can clarify whether participants disagree about facts, meanings, or priorities. Clarifying meaning helps frame a trade-off, but resolving it still requires agreed criteria and authority.

5.2.1 Disentangling Uncertainty from Equivocality in Trade-offs

A dispute about space allocation may involve missing utilization data, different interpretations of operational needs, competing priorities, or all three. More square-footage data can clarify consequences but cannot by itself reconcile different visions of how teams should work or decide which priority takes precedence. The facilitator helps participants clarify operational meanings and strategic objectives while collecting any missing evidence. Alternatives are then compared against agreed criteria. Priority conflicts that remain unresolved within delegated authority go to the appropriate executive.

5.2.2 Retrospective Sensemaking for Shared Understanding

Reviewing successes and failures in previous facilities can provide evidence for examining current assumptions. Retrospective sensemaking helps stakeholders interpret that experience but does not guarantee consensus or establish that past conditions will recur. Examining previous organizational changes can help stakeholders formulate principles for the current project. For example, evidence that a previous facility could not accommodate remote work may support giving adaptable infrastructure greater priority than private office count. That priority remains a project-specific decision, not an automatic conclusion from past experience.

5.2.3 Balancing Retention with Adaptability

Retained interpretations support continuity, but rigid reliance on them can obscure new conditions. Predesign teams need to distinguish non-negotiable constraints from adaptable guidelines and revisit the assumptions behind both as design progresses. Changes to approved requirements remain subject to established decision rights.

5.3 Identifying Decisions

Begin by identifying the decisions required, the information each depends on, the units affected, and the authority responsible for resolving competing requirements. Distinguish missing evidence from differing interpretations and conflicting priorities to determine the next action: collect information, develop a shared understanding, coordinate across functions, or seek an authorized decision. These actions may be needed together rather than in a fixed sequence. Record these elements for each significant decision: the requirement or unresolved issue; evidence and assumptions; affected units and dependencies; alternatives and evaluation criteria; budget, schedule, and performance implications; and the responsible authority and escalation path. The resulting decision record connects organizational priorities to facility requirements.

5.3.1 Connecting Evidence to Decisions

In strategic predesign, the workflow begins with a defined business problem or opportunity, not a predetermined solution. The sponsor and planner establish objectives, boundaries, success criteria, timing, and decision authority. Interviews, surveys, observations, and assessments of existing conditions supply evidence. Controlled software services organize this evidence for gap analysis, scenario development, and alternatives evaluation. Feasible alternatives are compared on a common basis, specialists validate material findings, and the planning team recommends a course through the living business case. The designated authority then records the authorized action, any conditions, or the need for further work.

Each handoff preserves sources, definitions, assumptions, versions, review status, and responsible parties. A revised workforce forecast, for example, may affect the space-demand model, property suitability, and occupancy-cost estimate. Where dependencies and calculations have been configured, software flags affected records and recalculates working outputs. AI may summarize the implications, while HR, programming, real estate, IT, and finance validate the consequences. The appropriate authority reviews any proposed baseline change; refreshed data never silently becomes an approved requirement or contractual commitment.

Confirm two separate assignments before configuring workflows: who leads the work and who has authority to approve each consequential action. Strategic planners lead strategic evaluation; portfolio and asset managers lead Real Estate work with qualified advisers; project managers coordinate Project Definition; and architectural programmers and workplace consultants develop the Program of Requirements with users and specialists. Across all three phases, the project sponsor maintains alignment with intended business outcomes. Phase leadership does not confer approval rights. Acceptance of working outputs, phase transitions, property commitments, POR approval, and authorization to enter design may require different authorities. An appointed Owner’s Project Manager (OPM) performs only the services and exercises only the authority established by the appointment. Procurement, professional duties, and signature authority remain with the assigned parties. Access permissions, review routing, decision records, and escalation paths must implement these confirmed responsibilities.

6.0 Decision-Driven Workflow

Each predesign phase organizes evidence and stakeholder reviews around three recurring activities: inquiry and analysis establish the evidence; alternatives evaluation develops a recommendation; and authorization records the designated authority’s decision. These activities support, rather than add to, the phase-transition gates described under Governance & Decision-Support. At those gates, readiness is assessed and the appropriate authority decides whether and under what conditions work may proceed:

  • 1.      Inquiry and Analysis: Combine quantitative metrics, qualitative stakeholder needs, and leadership direction to define the problem, assumptions, constraints, and evidence gaps. Responsible specialists validate the working output before comparison or recommendation.
  • 2.      Alternatives Evaluation: Compare credible courses of action, including the status quo and, where appropriate, deferral, process improvement, technology change, existing-space optimization, leasing, sharing, capital investment, or staged combinations. Apply explicit criteria, common demand assumptions and planning horizons, documented calculations, and appropriate scenario and sensitivity analysis. Preserve costs, benefits, feasibility, uncertainty, and trade-offs in a traceable recommendation.
  • 3.      Authorization and Decision Recording: Submit the business case and review findings to the designated authority, who records whether to proceed, proceed with conditions, defer, or decline. Requests for revision or further analysis remain recorded review actions until an authorization decision is made. An Executive Override records an authorized departure from a recommendation while preserving the original evidence, rationale, accountable parties, and risk-monitoring conditions. It cannot bypass reserved approvals, mandatory requirements, or signature limits.

6.1 Matching Information Demand with Processing Capacity

The coordination mechanisms described above must be matched with the work. Greater task uncertainty, more varied outputs, technological change, stringent standards, and tight schedules can increase information demand. Repeated hierarchical referrals consume managerial time and attention; when demand exceeds available capacity, decisions may be delayed. The central design objective is to match information-processing requirements with organizational capacity. Two complementary approaches are available: reduce the need for information processing and increase the capacity to process it. Their application depends on task uncertainty, interdependencies, and the decisions that must be coordinated.

Deliberately provided slack resources, such as additional capacity or schedule allowances, can accommodate uncertainty and reduce the frequency of exceptions requiring immediate coordination. Performance targets may also be adjusted explicitly. Self-contained task units organize more of the required resources around a product, market, client, or geographic area, reducing dependencies and scheduling conflicts across units. Self-containment changes coordination boundaries; it does not necessarily eliminate specialization, all cross-unit dependencies, or the organization’s overall output diversity.

Vertical information systems strengthen formal decision channels through technology, supporting staff, and more frequent planning or replanning cycles. Lateral relations enable cross-functional problem-solving closer to the source of information while retaining escalation for decisions beyond delegated authority. Together, these mechanisms can improve coordination without requiring every exception to travel through the hierarchy.

When information demand exceeds coordination and decision-making capacity, the consequences may include delayed schedules, budget overruns, or degraded performance. These outcomes do not demonstrate that capacity and requirements have been adequately matched, nor are they equivalent to deliberately planned slack. Strategic predesign should make the mismatch visible early enough for the organization to choose coordination mechanisms and allowances consciously.

7.0 Information-Processing Capacity

A decision-support system can strengthen information-processing capacity by organizing evidence, supporting analysis, and routing reviews. Its contribution depends on the coordination arrangements and decision rights established by the organization. Vertical information systems support planning and reporting; lateral workflows connect functions; multidisciplinary project teams handle bounded assignments within shared standards; and contingency tracking makes approved buffers and responses visible. These are operating arrangements, not additional technical tiers. Training, incentives, and accountability support their use but do not replace evidence validation or formal decision authority.

Information-Processing Strategy

Supporting Capability

Owner Pain Point Addressed

Principles

Increase Capacity:

Vertical Information Systems

Data & Analytics Engine

Fragmented or unreliable data; slow metric tracking.

Evidence-Based: Provides a validated information view with source provenance; supports defined benchmarking and cost calculations and specialist interpretation of zoning constraints.

Increase Capacity:

Lateral Relations

AI-Assisted Collaboration Hub

Organizational misalignment and communication silos.

Collaborative & Integrated: Connects HR, IT, Finance, and architecture, engineering, and construction (AEC) teams to flag trade-offs and route validated changes for review.

Reduce the Need:

Self-Contained Tasks

Coordinated Project Teams

Inefficient resource sharing and tangled workstreams.

Lifecycle-Oriented: Groups the resources needed for bounded work within a defined mandate; shared standards and interfaces preserve coordination across units.

Reduce the Need:

Slack Resources

Contingency and Buffer Tracking

Rigid forecasts or budget estimates that fail under unexpected site or market shifts.

Anticipatory & Adaptive: Uses uncertainty models and contingencies where supported by assumptions, and tracks approved schedule and resource buffers for risk mitigation.

7.1 The Role of a Decision-Support System

Together, the organizational information-processing model and the sensemaking framework explain why decision support must address both the availability of information and its interpretation. During predesign, the system preserves evidence, alternative interpretations, stakeholder feedback, revised assumptions, requirements, and decision rationale. It supports shared understanding and accountable resolution without treating consensus as a substitute for an authorized decision.

These principles translate into three practical requirements for system configuration: make information gaps visible, preserve differing interpretations for review, and route competing requirements to the appropriate authority before facility commitments are made. Matching processing capacity to task uncertainty, supporting shared understanding, and establishing decision authority provide an integrated basis for design.

In practice, shared evidence and issue records support lateral coordination among business units, facilities, HR, IT/AV, and Legal. Decision summaries support escalation when an issue exceeds delegated authority. Recording the criteria, trade-offs, responsible authority, and rationale makes the relationship between organizational priorities and predesign requirements visible.

Change management and process improvement make these arrangements usable in practice. Engagement begins in Strategic Planning and continues through Real Estate and Project Definition. The system can organize scenario reviews, capture operational concerns, distinguish proposals from approved decisions, and communicate decisions and their rationale to affected groups.

Anonymous feedback is used only where anonymity can be maintained; confidential follow-up and facilitated discussions have different protections. Review findings can improve templates, coordination routines, onboarding materials, and playbooks without standardizing project-specific answers. Record and validate proposed improvements before wider adoption and change approved project baselines only through authorized change control.

These organizational requirements lead to the system components and architecture described next. Configuration must reflect the enterprise’s information needs, review responsibilities, decision rights, and escalation paths. Technology supports these arrangements; it is not a substitute for governance.

8.0 Decision-Support System

For strategic predesign, a decision-support system connects organizational purpose to evidence, alternatives, validated requirements, and authorized decisions. It helps establish why an initiative is needed, which course best addresses the problem, and whether a project is ready for design. Four connected components support this work: software services, internal and external data sources, AI applications, and human judgment and source-verification interfaces. Connections with delivery and operations preserve continuity and support future planning. The system does not make commitments or confer decision authority.

9.0 System Components

The components describe what the system needs to function. Software provides services, data provides evidence, AI provides selected assistive capabilities, and human interfaces connect analysis to verification and decision authority. AI applications are implemented through software, but their distinct validation and governance needs warrant explicit treatment. None of the components is sufficient alone.

9.1 Software Services

Software connects sources, checks data quality, applies shared identifiers and definitions, and maintains access and version controls. Analytical services perform forecasting, financial modeling, spatial evaluation, and schedule analysis using documented methods. Workflow services assign reviews, route recommendations and approvals, and track conditions and changes. Record services preserve the business case, Program of Requirements (POR), decision history, and authorized baselines. These services can use existing enterprise and specialist applications rather than replace them.

9.2 Internal and External Data Sources

Internal evidence includes enterprise strategy, financial limits, HR forecasts, asset capacity, lease obligations, facilities records, and stakeholder needs. External evidence includes property-market information, zoning and environmental requirements, demographic and labor-market data, and economic indicators. Source records can remain in their authoritative systems. The decision-support system creates a validated information view that retains provenance and distinguishes recorded facts, observations, forecasts, assumptions, and stakeholder preferences.

9.3 AI Applications

AI applications assist retrieval from permitted sources, synthesis of interviews and documents, extraction of candidate requirements, identification of issues, development of preliminary options, and drafting of business cases and communications. They can also help explain results produced by analytical software. If an AI model contributes to forecasting or optimization, its inputs, method, limitations, and validation must be documented alongside other analytical methods.

Retrieval-augmented generation (RAG) retrieves permitted reference material to support an AI response. Preserve the retrieved source references and versions, check material claims against the underlying records, and report missing or conflicting evidence. An AI agent may coordinate multiple tool-based steps, such as retrieving records, requesting a scenario calculation, and preparing a draft. Its permitted actions, stopping conditions, and human-review points must be defined before use.

9.4 Human Judgment and Source-Verification Interfaces

Human interfaces make the evidence behind a finding available for review through inquiry forms, evidence views, scenario comparisons, review queues, and decision records. Reviewers inspect the underlying source, its date and version, the transformation or calculation applied, and unresolved qualifications. Source verification checks origin, whether the information is current, applicability, definitions, and supporting evidence. Material AI-generated claims are checked against underlying documents or records. Financial and spatial results are checked against inputs and methods. Automated checks can flag missing fields, inconsistent units, stale records, or unusual values, but they do not establish source authority or suitability for a decision. Reviewers can confirm, correct, reject, or escalate findings and record the reason. Specialists establish technical validity, business leaders reconcile priorities, and designated authorities authorize action. Training and incentives support informed use but do not substitute for defined decision rights or review.

10.0 System Architecture

The three predesign phases describe when decisions are developed. The four system components describe what resources and interfaces support them. The four logical tiers organize responsibilities and information handoffs. Software spans all four tiers, internal and external evidence enters through the foundation, and AI supports approved activities. Tiers describe the flow from evidence preparation through analysis and governance to controlled records. Review and revision can return information to an earlier tier.

  1. The Ingestion & Data Foundation Tier: Collects, cleans, and standardizes internal and external inputs.
  2. The Processing & Intelligence Tier (AI, Software & Analytics): Applies algorithms, scenario engines, and generative tools to process data and evaluate alternatives.
  • 3.      The Governance & Decision-Support Tier: Routes review and phase-gate workflows under confirmed RAPID assignments, delegated authority, and change-control rules; records authorized departures from recommendations.
  • 4.      The Output, Records & Lifecycle Continuity Tier: Preserves the living business case, requirements, decision records, and controlled handoff; connects performance evidence to future planning

10.1 Integration and Automation Across the Architecture

Integration connects data sources, software services, and AI applications. Automation coordinates selected activities across the four logical tiers. Components describe what the system uses, tiers organize responsibilities and information handoffs, and integration and automation describe how the components work together.

Architectural Tier

Integration and Automation

Human Responsibility

Ingestion & Data Foundation

Approved connectors, queries, or transfers retrieve selected information from enterprise applications, document repositories, external sources, or an enterprise data lake. Software checks formats, identifiers, units, dates, and completeness. AI may extract candidate information from documents.

Confirm source authority, relevance, definitions, and suitability for the intended analysis.

Processing & Intelligence

Analytical software uses prepared information to run forecasts, financial models, spatial comparisons, and scenarios. AI assists retrieval, synthesis, preliminary option development, and explanation. Configured dependencies identify results affected by changed inputs.

Validate assumptions, calculations, constraints, and material AI-generated claims.

Governance & Decision-Support

Workflow software assigns reviews, routes recommendations under confirmed RAPID roles, tracks mandatory clearances, and escalates unresolved issues. Automation issues reminders or prevents incomplete submissions from advancing where such controls are configured.

Reconcile priorities, determine readiness, and authorize decisions within established delegations.

Output, Records & Lifecycle Continuity

Record services preserve versions, evidence links, approved requirements, decision rationale, and handoff packages. Defined interfaces connect later performance evidence to future planning.

Confirm approval status, control revisions, and ensure handoffs reflect the actual authorization.

10.1.1 Integration Requires Shared Meaning and Controls

A technical connection allows information to move. Meaningful integration requires compatible definitions and controlled responsibilities. Connecting an HR forecast to a space-planning model is not enough. The interface must establish the forecast’s date, planning horizon, location, population, and assumptions, and distinguish workforce headcount from attendance, occupancy, utilization, and required capacity. A shared data dictionary defines each measure. Where relationships are complex, a governed ontology—a model defining entities and their relationships—can connect buildings, spaces, leases, teams, costs, and observations.

10.1.2 Automation Is Not Synonymous with AI

Automation executes configured activities. AI provides selected assistive capabilities. Scheduled data refreshes, documented calculations, review notifications, and versioned records can be automated without AI. Conversely, an AI-generated summary may require manual review before it enters a recommendation. Automation should advance information and review tasks. Integration makes evidence available; software applies documented methods; AI assists interpretation; automation coordinates repeatable work; and people verify findings and authorize action. The tier responsibilities define how functions operate at each handoff.

11.0 Tier Responsibilities and Information Handoffs

11.1 Ingestion & Data Foundation Tier

The foundation connects approved sources and prepares evidence for use. Shared identifiers link assets, locations, projects, and requirements; common definitions align units, periods, and measures. Integration retains source ownership and provenance, while validation records whether the information is current, complete, and fit for the intended analysis.

Enterprise Operations & Strategy Data: Revenue projections, business line expansion goals, CapEx thresholds, and headcount forecasts (from Finance & HR).

Portfolio & IWMS Asset Data: Spatial occupancy metrics, physical capacity data, lease administration details (expiration dates, contraction rights), and maintenance logs from Integrated Workplace Management Systems.

External Market & Environmental Data: Real estate market surveys, zoning laws, demographic forecasts, travel-time information for relevant labor markets, and macroeconomic indicators.

Stakeholder & Qualitative Inputs: Occupant survey feedback scores, structured interview notes from business unit leads, and vision session outcomes.

Each interface specifies the source owner, permitted access, definitions, refresh expectations, and validation rules. Internal connectors may draw on enterprise resource planning (ERP), HR, finance, Integrated Workplace Management Systems (IWMS), lease systems, geographic information systems (GIS), customer relationship management (CRM), and document repositories. External evidence retains publication dates and references. Sensors and Internet of Things (IoT) devices supply observations, not approved requirements. Occupancy, utilization, and capacity require distinct definitions, denominators, observation periods, and sampling qualifications. AI-assisted search can discover possible evidence, but the underlying document or record remains the source. A shared information repository may support integration without centralizing every record.

11.2 Processing & Intelligence Tier (AI, Software & Analytics)

This tier uses validated foundation data to support forecasting, financial comparison, spatial analysis, and scenario evaluation. Software performs defined calculations; AI assists discovery, interpretation, and preliminary option development; specialists verify inputs and findings before decision review. The services support both vertical information processing and lateral coordination without acquiring independent approval authority.

AI-assisted evidence capture includes interview and meeting synthesis, candidate requirement extraction, benchmark discovery, and retrieval from permitted repositories. Analytical assistance includes explaining space and budget forecasts, identifying candidate risks, and comparing preliminary test fits or sequencing options. Drafting services prepare documents and stakeholder messages. In each case, the responsible specialist checks the evidence and output before it becomes part of a recommendation; calculations and constraint checks remain tied to documented analytical methods.

A reusable decision model separates observed data, reference benchmarks, adjustable assumptions, calculation rules, and evaluation criteria. Version each layer and retain the inputs, method, and results for every material scenario. Defined calculations should run in the analytical service rather than rely on a language model to generate numerical answers. Deterministic calculations are reproducible under the same inputs and rules but can still be wrong if those inputs or rules are wrong. Statistical forecasts and optimization methods require their own validation and uncertainty reporting. AI can propose questions or scenarios and explain outputs; specialists verify the model, while designated authorities approve consequential action.

Demand Forecasting and Scenario Analysis: Translates validated business forecasts into capacity and spatial requirements using documented sector-relevant measures, such as patient volume, enrollment, or production units. Workplace analysis connects portfolio obligations, team allocations and growth, observed attendance and utilization, and qualitative experience. Test weekly peaks, concurrent activities, local operating conditions, and policy assumptions rather than applying a universal desk-to-headcount ratio. Reviewers assess source coverage, sampling limitations, planning horizons, and uncertainty; forecast demand is not an observed fact, and low average utilization alone does not establish surplus capacity.

Financial and Total Cost of Occupancy Analysis: Compares lifecycle costs and cash flows using consistent demand assumptions, planning horizons, price bases, and treatment of costs and benefits. Discounted cash flow (DCF) and net present value (NPV) are used where appropriate, with documented discount assumptions. Include relevant rent, fit-out capital expenditure (CapEx), operating expenditure (OpEx), tenant-improvement allowances, relocation, disruption, contingencies, and end-of-term obligations. Finance validates input and methods, prevents double counting, and reviews sensitivity and affordability; projected returns do not guarantee strategic value or deliverability.

AI-Assisted Option Development and Spatial Analytics: Supports preliminary test fits, blocking and stacking studies, and spatial alternatives from validated requirements and the Program of Requirements (POR), as it develops. Qualified planning professionals use connected spatial-analysis tools to test quantities, area definitions, circulation, relationships, and known constraints. Specialists assess unresolved technical and regulatory questions. A test fit or generated option is not an approved design, compliance determination, or confirmation of building-system capacity.

AI Control Enforcement: Analytical services implement risk classifications, usage boundaries, access restrictions, and audit requirements established through Govern, Guide, Guard. Material outputs retain their evidence and review status and are routed for human verification before decision use. Policy ownership remains crosscutting rather than confined to the analytical tier.

11.3 Governance & Decision-Support Tier

This tier separates the recommendation process from authorization and implementation. Bain & Company’s RAPID assigns decision roles, not a task-responsibility matrix or a sequence of steps. Recommend gathers input, develops options, and prepares the proposal. Agree confirms specified mandatory requirements and is assigned sparingly. Input contributes evidence and advice without concurrence or approval authority. Decide identifies the designated decision authority. Perform implements the authorized decision; preparation and analysis belong to the recommendation process, not to Perform merely because they involve tasks. For each decision, confirm the named recommendation lead and decision authority, mandatory concurrence requirements, approval thresholds, signature limits, and escalation route. Record separate decisions when authorities differ, and record the governing decision rule where a collective body decides. Workflow permissions implement actual delegations. Acceptance of a working output does not authorize capital expenditure, contracts, or reserved owner decisions.

Strategic Planning Gate: Reviews organizational direction, baseline conditions, demand forecasts, gaps, scenarios, portfolio priorities, and the initial business case. The strategic planner submits the recommendation. The sponsor, within delegation, or the designated executive or owner authority approves planning direction and next-phase work. The gate record preserves financial assumptions, conditions, unresolved issues, responsible parties, and any higher-level authorization required. Approval of planning direction does not itself authorize a property transaction or entry to design.

Real Estate Gate: Reviews approved search criteria, market evidence, property alternatives, specialist due diligence, fit and lifecycle costs, relevant commercial terms, and the updated business case. Selection criteria are approved before candidate review. Binding property commitments require a separate decision by the designated commitment authority and execution by an authorized signatory. The phase-transition decision approves real estate direction and transfers the controlled baseline to the Project Manager, including evaluation evidence, separately authorized commitments, unresolved conditions, and assigned actions. Material mismatches revealed during programming require reassessment through the appropriate authority; a completed transaction does not authorize design.

The Project Definition Gate and Project Readiness Review maintain four related but distinct records. First, the Approved Plan establishes the charter, responsibilities, milestone schedule, specialist scopes, and controlled information environment. Second, the owner-approved Program of Requirements (POR) and supporting baselines reconcile requirements, success criteria, scope, quality, budget, schedule, technology, and risk. Third, the Project Readiness Review findings and updated business case establish the evidence and recommendation for the gate decision; the review does not itself authorize design. Fourth, the designated owner authority records whether to proceed, proceed with conditions, defer, or decline. Readiness and authorization may share a decision maker only under explicit delegation. Procurement, Legal, and technical advisers assess delivery and sourcing options against project needs and applicable rules; strategy approval does not authorize contract awards. Record each condition, responsible party, deadline, and checkpoint, and transfer the controlled package in accordance with the actual authorization.

Executive Override and Risk Register: Preserves the original recommendation and analytical record when an authorized decision departs from them. The decision record captures rationale, authority, approval conditions, accountable parties, and ongoing monitoring. The workflow escalates departures beyond delegated limits and does not permit mandatory requirements to be waived without applicable authority.

11.4 Output, Records & Lifecycle Continuity Tier

This tier preserves records and their connections over time, linking evidence, recommendations, authorizations, and subsequent outcomes. The living business case is a single controlled, versioned case maintained through Strategic Planning, Real Estate, and Project Definition—not a new case at each gate or a report that updates approved commitments automatically. Working evidence, reviewed analysis, recommendations, approved baselines, and proposed revisions remain distinguishable. The business case, decision record, and readiness report remain separate deliverables. Controlled handoffs identify requirements, assumptions, success criteria, accepted conditions, responsible parties, supporting evidence, and unresolved actions.

The Living Business Case Repository: Maintains the current case and its version history from Strategic Planning through Real Estate and Project Definition. Assign an owner, a review cadence, and material-change triggers for workforce forecasts, market availability, costs, occupancy patterns, policy, or other decision-critical assumptions. Relate annual budget cycles to relevant lease dates, delivery lead times, and longer asset horizons without assuming all projects share the same periods. Each review identifies what changed, which alternatives are affected, and whether further analysis or authorization is required. Updated calculations produce a proposed revision; contributors validate impacts and route material changes to the appropriate authority rather than amend an approved commitment automatically.

Project Charter & Decision Record Store: Preserves the formal mandate, approved scopes, dissenting views, conditions of approval, and accountable parties.

Operational Feedback Loop: Carries approved requirements, intended outcomes, and decision records into design, construction, and operations. Subsequent actual construction costs, post-occupancy performance data, and utilization metrics return to enterprise systems to inform future strategic planning. These later activities belong to the broader built-environment lifecycle; they do not extend strategic predesign beyond its authorization and handoff to design.

11.5 From Source Evidence to Authorized Records

System Tier

Core Data Inputs

Processing Logic & Techniques

Primary System Outputs

1. Foundation

Financial metrics, HR headcount, IWMS asset data, market surveys

Source verification, data-quality checks, shared definitions, and access control

Validated enterprise information view with retained source provenance

2. Intelligence

Validated baseline data and traceable qualitative stakeholder inputs

Documented financial and spatial calculations, scenario analysis, AI assistance, and specialist validation

Comparative alternatives analysis and risk profiles supported by documented assumptions

3. Governance

Comparative findings, risk registers, stakeholder constraints

Human evaluation, confirmed RAPID routing, mandatory clearances, and authorization checks

Separate working-output acceptance, readiness findings, authorization decisions, conditions, and override audit trails

4. Records & Lifecycle Continuity

Approved POR and supporting baselines, current business case, charter, readiness findings, and authorization records

Controlled versioning, evidence links, handoff records, and defined interfaces for later performance feedback

Versioned business case, controlled POR and decision history, authorized handoff package, and assigned follow-up

12.0 Lifecycle Interface and Digital Twins

Where design is authorized, predesign concludes with a controlled handoff, but its information remains useful throughout delivery and operations. Shared asset and project identifiers connect approved requirements with actual construction costs, occupancy, maintenance, and performance measures. Later observations inform future business cases and forecasts without retroactively changing the original decision record.

Digital twins can provide connected digital representations of portfolio, campus, or building assets for monitoring, simulation, or decision support when relevant models and observations are maintained over time. Occupancy, sustainability, and energy analyses can relate measured performance to approved objectives and assumptions. A dashboard or static model alone does not establish the connected modeling capability described here. Digital twins are optional lifecycle interfaces, not prerequisites for predesign. Their updates enter through the same source-verification and change-control processes as other evidence.

13.0 AI Governance: Govern, Guide, Guard

In this chapter, applicable controls connect accountable ownership, permitted activities, data protection, testing, human verification, monitoring, and policy updates across all four tiers. These controls must be implemented and tested in the operating workflow, not left solely in a checklist.

Govern: Establish an accountable governance body or equivalent authority, process and data owners, risk classifications, approved uses, and monitoring and incident responsibilities. Maintain an inventory of approved tools, agents, and their uses; review changes in models, permissions, vendors, and workflows. Identify and address unapproved use through clear policies and supported alternatives. Permission to use an AI service is distinct from authority to approve a business decision.

Guide: Define approved tools and organizational accounts, permitted data classes, prohibited disclosures, retention rules, intellectual-property requirements, and escalation routes. Review vendor handling of submitted data before use; approved access does not make every record suitable for every task. Minimize personal information and use aggregate workplace evidence where appropriate, avoiding individual performance assessment. Train users and reviewers to verify sources, calculations, unsupported claims, and limitations before decision use or dissemination. Spatial option development, drafting, and cost sensitivity analysis remain subject to specialist review and documented methods.

Guard combines preventive controls, boundary testing, and incident response. Limit each agent to the tools, records, and actions necessary for its approved task, using read-only access where sufficient. Enforce permissions and approval checks in the application and workflow rather than rely on instructions to the AI alone. Treat retrieved documents and external content as evidence, not authority to change permissions or execute actions. Require human authorization for consequential commitments, approved-baseline changes, and external release. Test unauthorized access, misleading retrieved instructions, fabricated references, excessive actions, and failure handling. Log permitted tool actions, outputs, and approvals while protecting sensitive information. At defined exceptions, stop and escalate; provide means to suspend the workflow and preserve records for incident review. These safeguards reduce risk but do not prove factual correctness or replace human authority.

The test of this decision-support system is not the volume of data collected or the speed of AI output. It is whether the organization can trace a consequential decision from its evidence and assumptions through specialist review, stakeholder interpretation, authorization, and controlled handoff. Information-processing capacity, shared meaning, and decision rights must work together. Begin with a bounded workflow, verify its quality and safeguards, and retain improvements without losing the record of what was known and approved at the time.

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