From Prompts to Playbooks

How retail landlords and tenants can turn AI into reliable legal and commercial workflows

Every week, property teams receive the same breathless pitch: AI will transform your practice. What they hear less often is how — in concrete, repeatable steps — to make that actually happen.

This is not about the hype. It is about building AI workflows that hold up under pressure — ones a paralegal can run on Monday morning and a partner can trust on Friday afternoon. The discipline applies equally whether you are an institutional landlord managing a shopping centre portfolio or a retail brand operating across multiple markets.

The Prompt Is the Product

The biggest mistake property professionals make when adopting AI is treating it like a search engine. Type a question, get an answer, move on. This produces results that look plausible and read like they were written by a committee.

The better mental model: the prompt is the product. A well-crafted prompt is a specification. It tells the AI who it is, what the task is, what the output should look like, and what the constraints are. Done properly, it functions like a detailed instruction to a capable junior colleague.

Every strong prompt has four building blocks:

• Role — sets expertise and perspective. "You are a senior commercial real estate attorney advising a US institutional landlord."

• Context — provides the relevant background. "This is a triple-net lease on an inline retail unit in a regional shopping center. The tenant is a national apparel brand."

• Task — states precisely what is required. "Summarize the key repair obligations, rent escalation mechanism, and assignment provisions."

• Format — defines the output structure. "Return bullet points under three headings. Flag anything that departs from ICSC standard market practice."

A prompt you can reuse is worth ten times one that only works once. Build for repeatability from day one.

Build Brick by Brick

Complex workflows collapse when you try to do too much in a single prompt. Asking AI to review a lease, identify issues, compare to market standard, suggest redlines, and draft a cover letter in one go will produce something comprehensive-looking and analytically shallow.

The fix is simple: decompose the task. Break the workflow into discrete steps, run each as its own prompt, and let the output of one feed into the next.

A lease review workflow might run like this:

1. Extract commercial terms — parties, premises, term, base rent, percentage rent, co-tenancy and kick-out clauses

2. Identify landlord and tenant obligations separately

3. Flag deviations from your house standard or local market norms

4. Draft a risk-rated summary suitable for an investment committee memo

5. Draft the cover email to the client

By the time you reach step five, the AI has full context and you have already validated the substance. Junior team members can own individual steps. Quality checks happen at each stage, not just at the end.

Beyond Leases

Leases are the obvious starting point — long, repetitive, full of standard provisions that invite systematic analysis. But retail property generates an enormous range of documents, and AI adds value across all of them.

Document analysis and reporting — reciprocal easement agreements, operating covenants, estoppel certificates, SNDAs, construction side letters, guaranty agreements, and letter agreements. Letter agreements in particular are notorious for being filed and forgotten. AI can surface their commercial impact fast — a buried rent abatement right or co-tenancy cure period discovered before a lease renewal negotiation can be worth real money.

ESG and sustainability — extract green lease provisions across a portfolio, identify leases with no framework for sharing utility data, flag assets at risk as energy benchmarking and disclosure laws expand across states, and draft ESG narratives for investor reports. As sustainability reporting requirements tighten — from EPA Energy Star to state-level benchmarking mandates — teams with structured AI workflows will move faster and report more accurately than those doing this by hand.

Precedent comparison and base document selection — one of the most time-consuming tasks in any property practice is selecting the right starting document for a new deal. Upload your last five leases for a similar asset class, ask the AI to produce a comparison table across fifteen key provisions, and use it to select your precedent and justify the choice in a file note. Landlord's counsel can identify their most favorable precedent; tenant's counsel can find the most favorable assignment regime from the client's executed lease portfolio.

Research — CAM audit rights, co-tenancy clause enforceability by state, percentage rent calculation disputes, force majeure scope after COVID litigation, radius restriction enforceability. Set the tone explicitly: "This is for internal research only. Where you are uncertain, say so. Do not cite cases unless you are confident they are real." That one instruction prevents most of the failure modes.

Set the Tone, Every Time

Tone calibration is where most first-time prompt builders cut corners — and where the most failures occur. The same underlying content requires a different register depending on who is reading it.

A letter from outside counsel to opposing attorneys reads differently from a memo to a tenant's VP of Real Estate, which reads differently again from a report going to an investment committee. Include the audience, the relationship, and the intent in your prompt every time. "Direct and professional, but not combative. We are seeking to progress the transaction, not re-open commercial terms."

This matters beyond emails. A research note for a first-year associate needs different calibration from a briefing for a board member. A CAM reconciliation narrative for a tenant reads differently from one going into an investor package. AI will match whatever tone you specify — the failure is usually in not specifying one at all.

Template First, Always

The most durable AI workflows are built on templates. A template is a prompt that has been tested, refined, and documented so that anyone on the team can run it without knowing how it was originally designed. Think of it as the difference between a recipe and a meal.

A retail property playbook might include templates for: lease abstracts, letter of intent review, SNDA and estoppel certificate analysis, CAM reconciliation review, co-tenancy and kick-out clause tracking, assignment and subletting consent letters, construction allowance disbursement memos, and precedent comparison tables. Each template should include role, context, task, format, and constraints — especially constraints, which is where most drafts go wrong.

Assign ownership. Someone should be accountable for each template, updating it as they discover what works and discarding what does not. Review the full playbook quarterly. The teams that treat this as infrastructure — not a novelty — are the ones that compound their advantage over time.

The Shift That Matters

The real transformation will not come from individual attorneys using AI to work faster. It will come from teams that have built structured, repeatable workflows — playbooks that encode institutional knowledge and make consistent quality accessible to everyone.

The landlord who can generate a precedent comparison in ten minutes rather than ten days has a genuine edge in a deal. The tenant whose in-house team can surface a buried rent abatement right before entering renewal negotiations is better positioned than one who finds it six months later.

That advantage is not about the technology. It is about the discipline of prompt engineering, the patience to build step by step, and the commitment to turn good individual practice into shared playbooks.

Start with one template. Test it. Improve it. Share it. Then build the next one.