What has changed across the platform.

After reviewing thousands of tenant fit-out submissions across retail, commercial and mixed-use assets, one pattern is remarkably consistent. The submission that looks complete is rarely the one that gets approved first time. The submissions that close comments quickly are the ones that make it easy for the reviewer to connect every design decision back to a requirement.

Landlords are often accused of being pedantic or slow. In our experience, the real problem is information asymmetry. The tenant's design team knows why a detail was drawn a certain way. The landlord's reviewer can only see what was submitted. If the reasoning isn't visible, the reviewer has two choices: reject the submission or make an assumption. Neither is good risk management.

This is why a compliant-looking drawing set can still fail review.

The drawing may show the right ceiling height, the right fire rating and the right structural opening. But if the reviewer cannot see the specification clause, the test evidence, the authority approval or the lease restriction that justifies the detail, the comment will come back. And when it comes back, it costs both sides money.

So what does a landlord actually need? Not more drawings. Better evidence.

Here are the six things that consistently separate a first-time approval from a revision cycle.

One: a clear requirement map. Every design criterion in the landlord's design criteria manual, lease, work letter or authority condition needs to be traceable to a submission item. The reviewer should not have to guess which drawing answers which requirement. A simple matrix is enough, provided it is accurate and version-controlled.

Two: the right evidence for the right requirement. Drawings prove geometry. Specifications and schedules prove performance. Certificates, test reports and manufacturer datasheets prove compliance with codes and landlord standards. Correspondence proves agreement. One document cannot do all of these jobs.

Three: consistency across the package. We frequently see a drawing that says one thing, the specification that says another, and the landlord's own criteria that imply a third. When the documents contradict each other, the reviewer has no basis for approval. The submission needs a single source of truth, or at least a clear hierarchy of documents with contradictions explained.

Four: authority and specialist inputs captured early. Fire strategy, acoustic treatment, structural loading, MEP capacity and accessibility compliance are not details to be added later. If a submission arrives without evidence that these matters have been signed off by the relevant specialist, the reviewer cannot treat it as compliant. The evidence should be in the package, not promised in an email.

Five: lease and work-letter alignment. Tenant teams often design to a brief that does not match the lease obligations. Landlord consents, permitted alterations, hours of access, signage restrictions, landlord contributions and make-good obligations all affect what can be approved. The submission should show awareness of these constraints, not just design intent.

Six: a clear response to previous comments. Re-submissions that simply reissue drawings without a marked-up response sheet waste everyone's time. The reviewer needs to see what changed, why it now complies, and where the supporting evidence sits. Without that, the comment gets raised again.

The common thread is traceability.

A reviewer is not looking for perfection. They are looking for confidence that the proposal can be demonstrated to comply with the requirements that matter. The more traceable the evidence, the faster the approval.

This matters beyond speed. Revision cycles have a cost that is easy to underestimate.

Each round of comments delays the contractor, the shopfitter, the ordering of long-lead items and the authority submissions downstream. Programme float disappears. Mobilisation dates slip. Opening dates move. And the relationship between landlord and tenant becomes strained before work has even started on site.

We have seen projects where a single missing fire-stopping certificate caused a three-week delay. We have seen retail openings put at risk because a specification clause was updated in one document but not in the package submitted for consent. These are not design failures. They are evidence-handling failures.

The good news is that this is fixable.

The discipline is not complicated. It is simply unfamiliar to teams used to producing design packages rather than consent packages. A consent package is a structured argument. Every drawing, every specification extract, every certificate and every cross-reference is a piece of evidence in that argument.

At StradHolm, this is precisely what UnifyReview is designed to support.

UnifyReview converts landlord design criteria, lease obligations, work letters and authority requirements into governed rule packs. Each submission is reviewed against those rules, with findings linked directly to the evidence required to close them. The reviewer still makes the decision. The AI does the tedious work of checking that the evidence is present and consistent.

The result is fewer revision cycles, faster approvals, and a clear audit trail of what was checked, what was missing, and what was accepted.

If your team is spending more time chasing submission completeness than reviewing design quality, the problem is probably not the design. It is the evidence structure.

Start there and the rest becomes much easier.

Every landlord and developer has a library of design criteria.

There are the technical manuals. The tenant design criteria. The lease schedules. The work letters. The authority conditions. The asset-specific addendums. The emails from the last fit-out that changed one clause but never made it back into the master document.

In practice, these documents rarely behave like a single standard. They behave like a pile of standards.

And that pile is exactly what makes design review so expensive.

A reviewer has to hold all those sources in their head, compare them against the submission, remember which version applies to which unit, and decide whether the evidence on the table actually satisfies the requirement. It is slow, error-prone and impossible to scale consistently across a portfolio.

The real-world impact is bigger than a few extra review hours.

When the same requirement is interpreted differently by different reviewers, tenants get inconsistent decisions. Consent conditions drift from one asset to another. Approved details from one project are forgotten by the next. Valuable precedent is buried in old PDFs and inbox threads.

The problem is not a lack of rules. It is that the rules are not governable.

A governed rule pack changes that.

A rule pack is not just a digitised manual. It is a structured, versioned set of criteria that can be applied, traced, updated and audited. Each rule has a source, a scope, a test and a required evidence type. When the source document changes, the rule is updated. When a project uses a rule, the decision is recorded.

This is the foundation of the UnifyOS approach.

Start with the sources.

The first step is to accept that design criteria come from many places. The landlord's technical manual. The lease. The work letter. Authority conditions. Asset-specific restrictions from a refurbishment. Even lessons learned from previous projects. Each of these becomes a source document with a version and an owner.

Next, extract the rules.

A rule is a single, testable statement. It is not a chapter. It is not a guideline. It is something a reviewer can confirm or challenge. For example: "Ceiling finish in retail units must achieve a fire classification of Class 0 or equivalent, supported by test evidence." That is a rule. It has a subject, a requirement, and a required evidence type.

Then define scope.

The same rule does not apply everywhere. A rule about mall-front glazing may apply to a new mall fit-out but not to a back-of-house alteration. A rule about acoustic separation may apply to cinemas but not to fashion units. Scope prevents rules from being over-applied or under-applied.

Then attach evidence requirements.

A rule without an evidence requirement is just an opinion. For each rule, the reviewer needs to know what would prove compliance. Drawing? Specification? Test certificate? Manufacturer datasheet? Authority approval? The rule pack should say so upfront.

Then version and govern.

When a rule changes, the old version remains traceable. Every submission reviewed against a rule is linked to the version that applied at the time. If a tenant challenges a consent decision years later, the organisation can show exactly which criteria were active and when.

The effect on review is immediate.

A submission arrives. The system loads the rule pack that applies to that asset, that unit type and that project phase. Each rule is checked against the evidence provided. Missing evidence, contradictions and deviations are flagged with references back to the source document.

The reviewer is no longer hunting through folders and memories.

They are reviewing findings. They are applying judgement. They are making decisions, not reconstructing the standard.

The real-world impact shows up in several places.

Speed. First-pass reviews become faster because the criteria are already structured. The time saved is not from skipping rigour; it is from removing the administrative work of assembling the standard each time.

Consistency. The same rule is applied the same way across reviewers, assets and projects. That is especially valuable for portfolios where consistency is part of brand protection and tenant experience.

Auditability. Every decision is traceable to a rule version and an evidence item. That matters for disputes, insurance, handover and operational handoff.

Knowledge retention. When an experienced reviewer leaves, the rule pack does not leave with them. The organisation's accumulated criteria remain available and improvable.

Continuous improvement. Every comment, every rejection, every approved exception becomes feedback. If a rule is repeatedly misinterpreted, it can be clarified. If a new asset type exposes a gap, the rule pack is extended.

This is how scattered documents become reusable intelligence.

We have worked with organisations where the design criteria lived in five or six documents, plus unwritten team knowledge, plus project-specific emails. Converting that into a governed rule pack is not a trivial exercise. But once it exists, the payback is compounding.

UnifyReview is built around this principle.

It treats landlord design criteria, lease obligations, work letters and authority conditions as structured rule packs. Submissions are reviewed against them. Findings are linked to the specific rule and the evidence required to satisfy it. Approvals carry the rule version. Changes to the source documents flow through to future reviews without rewriting history.

The result is a review process that is faster, more consistent and easier to defend.

If your team is still reviewing each submission against a folder of documents and a set of individual memories, the inconsistency you are experiencing is not a people problem. It is a structure problem.

Govern the rules. The rest follows.

The real-estate industry is entering an important phase in its adoption of artificial intelligence.

The first wave demonstrated that AI could read documents, extract information, generate reports and assist professionals with individual tasks.

The next wave is considerably more ambitious.

Agentic AI has the potential to change how the work itself gets done.

Recent analysis from McKinsey on agentic AI in real estate describes a shift away from isolated AI applications towards operating models where people and specialised AI agents work together across entire workflows.

At StradHolm Technologies, we believe that distinction is critical.

It is also why we are building something broader than an AI design-review platform.

We are building UnifyOS — Agentic Real Estate Operating Intelligence.

Start with a real problem. Then connect the lifecycle.

There is a temptation with emerging technology to attempt to solve everything at once.

We are taking the opposite approach.

Start with a defined problem where AI can deliver measurable value. Prove it. Govern it. Then connect it to the next workflow.

For StradHolm, that starting point is UnifyReview.

Design and technical review remains one of the most fragmented workflows within commercial real estate.

A single tenant submission can require teams to move between drawings and specifications; landlord design criteria; lease obligations and work letters; MEP requirements; authority regulations; previous comments and approvals; project correspondence; and multiple internal and external stakeholders.

Much of that work remains manual.

AI can change that.

UnifyReview — Design & Compliance Intelligence is being developed to provide an intelligent first-pass review of those submissions while maintaining traceability to the requirements against which decisions are made.

But there is a much more important question.

What happens to that intelligence after the drawing is approved?

In most traditional workflows, surprisingly little.

The project moves forward. Another system takes over. Information gets exported, emailed, uploaded, re-entered or archived.

Context is lost.

The next team effectively starts again.

We think that model needs to change.

Introducing the Unify ecosystem

Our architecture is therefore being developed around a master intelligence and orchestration layer rather than a collection of disconnected applications.

UnifyOS — Agentic Real Estate Operating Intelligence. UnifyOS is the master layer. Its role is to connect information, specialist intelligence, workflows, decisions and ultimately actions across the asset lifecycle.

Beneath it sit specialist capabilities.

UnifyReview — Design & Compliance Intelligence. The entry point into the ecosystem. AI-assisted review of drawings and supporting documentation against landlord standards, leases, technical criteria, authority requirements and asset-specific rules. The objective is not simply faster drawing review. It is to convert traditionally unstructured review activity into structured, traceable asset intelligence.

UnifyProjectVision — Programme & Project Intelligence. Once a project moves into delivery, a different set of questions emerges. Is the programme actually achievable? Where is slippage developing? Which activities are affecting the critical path? What is the likely commercial consequence? Which decisions require intervention now rather than reporting next month? UnifyProjectVision is being developed to use AI to interrogate programme and project information and identify emerging schedule, cost, risk and delivery issues earlier. Not another dashboard telling the project team what already happened. Intelligence focused on what is likely to happen next.

IntelliBridge — The Integration Layer. Perhaps the most important principle behind UnifyOS is that we don't believe organisations should have to replace every existing platform to benefit from AI. Developers, landlords and operators have already invested heavily in systems including Aconex, Procore, property-management platforms, ERPs, document-management systems and specialist solutions such as RealCube. Those systems remain important systems of record and workflow. IntelliBridge is designed to connect them to the intelligence layer. Through APIs, webhooks and controlled data exchange, existing systems can continue performing the functions they perform well while UnifyOS provides intelligence and orchestration across them. The objective isn't another technology silo. It is an intelligence layer across the existing technology estate.

From lease to operations. This creates a fundamentally different proposition.

Imagine a new commercial lease being executed.

Instead of that lease becoming another PDF stored in a document-management system, relevant obligations can become structured requirements.

Lease. The system understands landlord obligations, tenant obligations, permitted works, technical restrictions, dates and dependencies.

Design. Those requirements inform the tenant's design criteria and submission requirements.

Review. UnifyReview assesses drawings and supporting information against the relevant requirements and identifies potential non-compliance.

Approval. Approved decisions become part of the project's structured record rather than disappearing into correspondence.

Delivery. UnifyProjectVision evaluates programme, milestones, dependencies, changes and emerging risks. A design change can potentially be understood not simply as a revised drawing, but as something with compliance, programme, cost and operational consequences.

Handover. Approved design and construction information follows the project into handover. The next team doesn't need to reconstruct what happened.

Operations. Relevant information becomes available to FM and asset teams. What was approved? What equipment was installed? Where is it? What warranties apply? What landlord obligations remain? What maintenance requirements were established?

Asset Intelligence. The asset progressively develops a usable operational history. And that intelligence can inform the next lease, refurbishment, capex decision or tenant change.

Lease → Design → Review → Approval → Delivery → Handover → Operations → Asset Intelligence

That is the journey we believe real-estate technology should enable.

The intelligence should follow the asset

This may ultimately be one of the most important principles behind UnifyOS.

Information should not stop because one project stage has finished.

Today, enormous amounts of valuable project intelligence are effectively lost at every organisational handover.

Leasing hands to design. Design hands to delivery. Delivery hands to operations. Operations eventually hand back to leasing or development when the asset changes again.

Each transition introduces friction.

Each transition creates another opportunity for information to become disconnected.

Agentic systems potentially allow us to create continuity.

The lease clause that influenced a design decision can remain connected to that decision.

The design decision can remain connected to the approved drawing.

The drawing can remain connected to the delivered asset.

The delivered asset can remain connected to its operational history.

That creates something far more valuable than document automation.

It begins to create an institutional memory for the asset.

Specialist agents, not one AI trying to do everything

We also don't believe the answer is one enormous AI model attempting to manage an entire property portfolio.

The stronger architecture is likely to involve specialist agents with clearly defined responsibilities.

A Lease Intelligence agent should understand leases. A Design & Compliance agent should understand drawings and technical requirements. A Programme agent should understand schedules, dependencies and critical paths. A Cost agent should understand commercial consequences. A Risk agent should identify developing exposure. A Handover agent should understand completion requirements. An FM agent should understand operational obligations.

Each performs bounded tasks.

UnifyOS orchestrates how those capabilities work together.

And critically, humans remain responsible for the decisions requiring professional judgement, commercial authority or accountability.

The objective isn't to remove experienced professionals.

It is to stop experienced professionals spending disproportionate amounts of their time performing work that machines can increasingly undertake.

Review is the entry point — not the destination

This is why we see UnifyReview as strategically important beyond design review itself.

It provides organisations with a practical way to begin adopting agentic workflows without attempting a wholesale transformation on day one.

Start with review.

Measure the reduction in review time.

Measure consistency.

Measure turnaround.

Measure compliance.

Measure the effect on approvals and project mobilisation.

Establish the governance.

Then connect the next domain.

Review → Project Intelligence → Handover → Operations.

The organisation moves towards agentic operations progressively rather than through a high-risk "big bang" technology transformation.

From systems of record to systems of intelligence

Real estate has spent decades digitising its records.

We have databases. Document-management systems. Project-management platforms. Property-management platforms. ERPs. Dashboards. Digital twins.

The industry does not necessarily need another place to store information.

It needs technology capable of understanding the information already distributed across those systems and helping professionals act on it.

That is the opportunity we see.

And it is why our proposition at StradHolm is deliberately bigger than: "AI can review your drawings."

Our proposition is: Start with a workflow where AI can demonstrate measurable value today — and use that foundation to build an intelligent operating layer across the real-estate lifecycle.

That is what UnifyReview starts.

That is what UnifyProjectVision extends.

That is what IntelliBridge connects.

And that is what UnifyOS is ultimately intended to orchestrate.

One asset. Multiple specialist intelligence layers. One connected lifecycle. From lease to operations.

That's the future we're building at StradHolm Technologies.

McKinsey's recent paper on Agentic AI in real estate caught my attention — not because it validates another use case for AI, but because it describes something much bigger. The transition from AI as a tool to AI as part of the real-estate operating model.

McKinsey makes the point that meaningful value won't come from simply adding more AI tools to existing processes. It comes from redesigning entire domains so that people and AI agents work together across connected workflows.

That distinction is important. And it closely reflects the architecture we are building at StradHolm Technologies.

We deliberately start with UnifyReview. Why? Because design and compliance review is an ideal entry point.

It is document-heavy, repetitive and fragmented across drawings, leases, landlord criteria, technical requirements, authority requirements, comments and approvals. There is enormous potential to automate those steps, while keeping experienced professionals responsible for the judgement.

But we don't believe the opportunity ends when a drawing receives an approval. Quite the opposite. The approval should become data for everything that happens next.

That is where UnifyOS comes in. Our longer-term architecture connects the real-estate lifecycle: Lease → Design → Compliance → Approval → Delivery → Handover → Operations → Asset Intelligence.

Consider what that potentially means. A lease obligation becomes a structured requirement. That requirement informs the design review. An approved design establishes part of the project's delivery baseline. A design change can trigger compliance, programme and cost implications. Project information flows into handover. Handover information becomes operational asset intelligence. And operational performance feeds the next decision.

The information doesn't die at each handover. The intelligence follows the asset.

That is the difference between applying AI to a task and building an AI-native operating model.

McKinsey describes five important layers for agentic real estate: factual, orchestration, action, control and reusable agent capabilities.

That architecture is particularly interesting because the future is unlikely to be one enormous AI agent trying to run a property. It will be specialist agents operating within governed workflows, connected to the systems where the work already happens and escalating decisions when human judgement is required.

That's the direction we see for UnifyOS.

UnifyReview is the entry point. UnifyOS is the orchestration layer. The asset lifecycle is the opportunity.

And importantly, this doesn't require developers or landlords to throw away the systems they already use. The intelligence layer can connect into existing property, project and operational platforms through APIs and integrations — progressively extending automation across the asset lifecycle.

McKinsey summarises the transition particularly well: from understanding what needs to happen to actually helping get it done.

Real estate has spent years digitising documents. The next phase is about connecting decisions, workflows and actions.

At StradHolm, we're building for that phase now. We're not trying to predict the future of real-estate operations. We're accelerating towards it.

#AgenticAI #RealEstate #PropTech #CRETech #ArtificialIntelligence #DigitalTransformation #AssetManagement #ProjectManagement #StradHolm #UnifyReview #UnifyOS

UnifyRetail OS has been extended so that a single governed record follows a unit from lease negotiation through to trading. Each lifecycle stage carries its own evidence set, rule pack and reviewer accountability.

The change matters most to portfolio teams running new and existing malls side by side, where the same design criteria must be applied consistently across very different asset ages and tenant mixes.

AI remains assistive throughout. Findings are evidence-linked and every approval stays with an authorised decision-maker.

UnifyReviewLite delivers the same engine as UnifyReview through an embedded route. Submissions can now be pushed in and findings returned over signed webhook payloads, so the host platform stays the system of record.

Every payload is HMAC-signed and written to an auditable event log. Setup prerequisites and event types are documented on the integrations pages.

RealCube, Aconex and Procore are named as integration targets rather than partnerships.

UnifyProjectVision is a standalone project-intelligence platform, separate from the UnifyReview compliance family. Its latest release improves how early warnings are detected and explained.

Signals are traced back to the schedule and cost evidence that produced them, so a project controls lead can challenge a finding rather than accept it.

Financial outputs remain forecasts intended to support judgement, not certified valuations.