AI Learning Hub

Getting Started with AI

What LLMs are, how they work, and how to get the most out of them — whether you use Claude, ChatGPT, or any other tool.

What is an LLM?

A highly capable colleague,
available 24/7.

A Large Language Model (LLM) is an AI tool you can have a conversation with — asking it to draft, summarise, analyse, explain or brainstorm. Tools like Claude and ChatGPT are LLMs. Think of one as a highly capable colleague available 24/7.

Unlike a search engine, an LLM understands context and handles multi-step instructions. It does not look things up — it reasons from what it has learned. The more specific you are, the better the output.

Draft & Write

Tenant letters, inspection notices, arrears reminders, owner updates — first drafts in seconds.

Analyse & Summarise

Lease terms, body corporate minutes, inspection reports — distilled into what matters.

Brainstorm & Plan

Retention strategies, fitout specs, marketing angles — a thinking partner available any time.

Research & Explain

Legislation terms, market conditions, lease clauses — explained clearly, without the jargon.

Be Specific About Context

LLMs perform dramatically better when they know who you are, what you need, and who the output is for. Treat your prompt like a brief to a smart colleague who knows nothing about your situation yet.

Weak prompt

Write an email to a tenant about rent.

Strong prompt

I'm a property manager in Auckland. Write a firm email to a commercial tenant 14 days overdue. Include a payment deadline and next steps.

Works in: Claude, ChatGPT, and any other LLM. The more context you give, the less it guesses.

Give It a Role

Assigning a role shapes the language, tone and depth of the response. A leasing agent answers differently to a compliance officer — and so should your LLM.

"Act as an experienced property manager…"
"You are a commercial leasing agent…"
"As a building compliance officer…"
"You're a landlord preparing for renewal…"

This works in every LLM. The role sets the frame for everything that follows.

Specify the Output Format

If you want a table, bullet list, letter or short paragraph — say so. LLMs default to long prose if you do not tell them otherwise. You are in control of the output shape.

"Provide this as a bulleted list"
"Format as a formal letter"
"Give me 3 options in one sentence each"
"Summarise in under 100 words"
"Use plain language, no jargon"

Also useful: "Under 150 words", "No preamble, just the output", "Include a subject line".

Iterate, Don't Restart

If the first response is not quite right, ask the LLM to refine it — do not start a new conversation. It remembers everything said in the current chat.

"Make this more formal"
"Shorten to 3 sentences"
"Add a section about access requirements"
"Rewrite for a residential tenant instead"
"Make the tone warmer"

Treating the LLM as a conversation — not a one-shot request — is what separates average from excellent outputs.

Is my tenant data safe if I paste it into an LLM?

Most paid LLM plans (Claude, ChatGPT and others) do not use your conversation data to train their models by default — but check the terms for your specific plan. You should be cautious about what you share: avoid pasting full tenant names, addresses, financial details, passport numbers, or highly personal information into any AI tool. Where possible, anonymise data before using it in a prompt — replace real names with placeholders like [Tenant Name] and specific addresses with [Property Address]. Your firm may also have its own AI usage policy that sets out what is and isn't acceptable, and you should follow that.

Will an LLM give me legal advice I can rely on?

No. An LLM like Claude or ChatGPT can explain legal concepts, summarise document language, and draft notices — but it is not a lawyer and its output should not be treated as legal advice. Always have a solicitor review anything with legal consequences. Think of it as a very capable first-draft assistant, not a legal professional.

What if the LLM gives me wrong information?

LLMs can produce confident-sounding but incorrect information, especially on specific facts, current rates, or jurisdiction-specific rules. This is called a hallucination. Always treat AI output as a first draft that needs human review. Use it to save time on structure and wording, not as a source of truth for facts you have not independently verified.

How do I get the team to actually use it?

The fastest path to adoption is social proof and commitment as a team. Get one enthusiastic team member to publicly share a time-saving win in the first week. The team will follow. Formal training sessions alone rarely drive sustained adoption — small wins shared openly do.

Can an LLM access our Re-Leased data?

Not directly — you would need to paste relevant information into the LLM manually, or use Re-Leased's built-in AI features which are integrated with your portfolio data. A general-purpose LLM like Claude or ChatGPT works best when you provide the context yourself in your prompt. Re-Leased's native AI features act on your actual data automatically.

What's the difference between Claude and ChatGPT?

Both are LLMs capable of the same core tasks. Claude (Anthropic) tends to excel at nuanced, longer-form text — ideal for professional communications and document analysis. ChatGPT (OpenAI) is widely used and equally capable for most property management tasks. The prompting techniques in this hub apply equally to both. Start with whichever your team already has access to.

What's the role of an LLM vs a Property Management Platform like Re-Leased?

They do fundamentally different jobs. An LLM like Claude or ChatGPT is a thinking and communication tool — it helps your people draft faster, think more clearly, and handle language-based tasks with less effort. Re-Leased is your operational backbone — it holds your leases, tenants, financials, maintenance records, and compliance history. That data is the ground truth of your portfolio. Re-Leased as an AI-first property management platform means AI is embedded directly into those workflows — automating tasks that require your actual portfolio data, like lease abstractions, arrears triggers, and inspection summaries. An LLM cannot do that without your data. The two work best together: Re-Leased runs your portfolio, LLMs help your team operate it better.

What LLMs Do Well — and Where to Be Careful

✓ LLMs excel at

✓

Drafting communications

Tenant letters, inspection notices, arrears reminders, welcome packs

✓

Summarising information

Lease terms, body corporate minutes, reports

✓

Creating templates

Standard notices, checklists, procedure documents

✓

Explaining concepts

Legislation terms, market conditions, lease clauses

✓

Brainstorming options

Retention strategies, fitout specs, marketing angles

⚠ Always verify before using

Specific legal advice

Always consult a solicitor for lease disputes, tribunal proceedings, compliance obligations

Exact figures & current data

Market rents, rates, valuations — verify against live sources

Tenant-specific details

LLMs don't know your portfolio — always review AI output for accuracy

Financial decisions

Acquisition analysis, loan structuring, tax treatment — use professionals

AI Practitioner certificate

Download your full certificate after completing all 5 weeks.

A structured 5-week program straight to your inbox, using the Hub information to bring your whole team up to speed with AI, from first prompt to embedded habit. Examples use Claude but the steps apply to any LLM.