---
"title": "How do I roll out AI across an asset management team?"
"description": "A step-by-step playbook for rolling out AI across a real estate
  or investment asset management team, from mapping reporting work and setting
  data rules to training, review steps, measurement and deciding what to
  automate."
"section": "Playbooks"
"sidebarTitle": "Roll out AI in asset management"
"published": "2026-11-11T14:15:20.742Z"
"updated": "2026-11-11T14:15:20.742Z"
---

## How do I roll out AI across an asset management team?

It depends on where your asset managers spend their week. If most of the time goes into reading and writing, such as lease and contract review, property or portfolio reports, investor updates and meeting notes, a phased rollout of an approved AI tool plus hands-on training usually pays off within one or two reporting periods. If the time goes into chasing and reconciling data from property managers, operators or portfolio companies, fix that data flow first, or build a connected workflow, because AI on inconsistent inputs mainly produces faster errors. Large teams should pilot with one group first.

## Who this is for

Heads of asset management, COOs and portfolio leads at real estate owners, developers and investment firms who want their team using AI every week on real work, while keeping reports accurate and data under control.

## The short version

- Map one reporting period task by task and pick three to five tasks to start.
- Set approved tools and data rules before anyone uploads a lease or a rent roll.
- Train on last period's real material, then use AI on the current period.
- Keep numbers in your models and systems, and review every AI-assisted draft.
- Measure time and rework per task, then decide what to standardize or automate.

## Step 1: Map the reporting period

Write out one full month or quarter of asset management work, with owners and rough time per task. For most teams the list includes:

- Collecting monthly or quarterly reports from property managers, operators or portfolio companies
- Reviewing financials against budget and business plan
- Reading leases, amendments, service contracts and loan documents
- Writing asset or portfolio reports and variance explanations
- Preparing investor and lender updates
- Tracking capital projects, leasing activity and key dates
- Taking and sharing notes from operator and leasing calls

Mark each task as mostly reading, mostly writing, mostly calculation or mostly moving data. Our guide on [finding which workflows to automate with AI](/docs/find-workflows-to-automate-with-ai) has a simple scoring method.

## Step 2: Choose where to start

Good first tasks are frequent, document heavy and easy to check. Common choices:

- **Lease and contract summaries:** key terms, renewal options, escalations and notice dates, checked against the document.
- **Report narratives:** first drafts of the "what happened and why" section of asset reports.
- **Call notes:** turning operator or leasing call notes into action lists.
- **Investor update drafts:** a first version built from the approved report.

Avoid starting with anything where AI calculates figures that go straight into a report or a valuation. Keep that in your models and systems, where it can be audited.

## Step 3: Set tools and data rules

Asset management teams handle tenant data, loan terms, investor information and figures that may be confidential under agreements with partners or lenders. Before rollout, decide:

- Which AI tools are approved, on business plans that do not use your data for training
- Which documents may be uploaded and which may not, for example anything covered by a confidentiality agreement that forbids third-party processing
- Who reviews AI-assisted output before it goes to investors, lenders or partners

See [how to write an AI acceptable use policy](/docs/write-an-ai-acceptable-use-policy) and [can we use AI on NDA documents?](/docs/can-we-use-ai-on-nda-documents).

## Step 4: Train on last period's work

Run two short hands-on sessions using real material from the most recent period. Have each asset manager summarize a lease they know well and compare the result to what they know is in it. Have them draft last quarter's report narrative and compare it to what was sent. This shows both what AI does well and where it slips, such as misread dates or missed amendments. Our page on [AI training for real estate development teams](/docs/ai-training-for-real-estate-development-teams) covers similar content for the development side.

## Step 5: Use it on the live period

In the next period, each person uses AI on the chosen tasks, with a named reviewer. Hold a 30-minute check-in each week for the first month. Save prompts that work into a shared library, and drop uses that do not help. See [creating prompting standards](/docs/create-prompting-standards).

## Step 6: Build in review steps

Keep controls simple:

- Every number in a report comes from your model or system, not from AI text.
- Key lease dates and terms pulled by AI are checked against the source before they go into a tracker.
- The person who signs off on reports today still signs off. AI does not change accountability.
- Keep the source files used for anything sent to investors or lenders.

## Step 7: Measure and decide

Before you start, record how long each chosen task takes and how many review rounds it needs. After two periods, compare. Gains are usually uneven: summaries and narratives improve first, while tasks that depend on late or messy operator data may not move. Then decide task by task whether to standardize, automate or drop it. See [measuring whether AI training worked](/docs/measure-ai-training-results).

## When to move from training to engineering

If the team still spends days each period collecting reports in different formats and retyping figures, training will not fix it. A connected workflow that collects operator reports, extracts key figures into a standard template and flags variances for review can remove much of that work, with a person approving the output. Searching across years of leases and past reports is another common build. See [custom AI vs off-the-shelf tools](/docs/custom-ai-vs-off-the-shelf-tools) and [company brain vs shared drive search](/docs/company-brain-vs-shared-drive-search).

## Common mistakes

- **Letting AI produce figures** for reports or valuations.
- **Uploading restricted documents** before checking confidentiality terms.
- **Starting during quarter end** with no time to learn.
- **No baseline,** so nobody can show whether it helped.
- **One session and done.** Habits fade without follow-up. See [why AI adoption stalls](/docs/why-ai-adoption-stalls).

## How Ruby helps

- **AI training:** hands-on sessions built on your own leases, reports and reporting calendar, with follow-up during the live period.
- **AI strategy and consulting:** mapping the work, choosing first tasks, setting data rules and measuring results.
- **AI engineering:** building connected workflows for report collection, lease abstraction and document search where training alone is not enough.

If you want a rollout plan built around your asset management calendar, [book a call with Ruby](https://rubyadvisory.com/book-a-call).

## Related pages

- [How do I roll out AI across an FP&A team?](/docs/roll-out-ai-across-fpa)
- [How to roll out AI across fund operations](/docs/roll-out-ai-across-fund-operations)
- [AI for real estate development underwriting](/docs/ai-for-real-estate-development-underwriting)
- [How do I run an AI pilot that doesn't stall?](/docs/run-an-ai-pilot-that-does-not-stall)
