IPS updates are a workflow problem

Most financial advisory firms don’t struggle to write an investment policy statement from scratch. The problem is keeping that statement current after a client’s life changes.

A client retires six months earlier than planned. Another sells a business. A couple’s risk tolerance changes after a sharp market drawdown. A child starts university. A client moves from accumulation to pension drawdown. Each event may affect objectives, liquidity needs, time horizon, asset allocation, or withdrawal assumptions.

The adviser often hears about the change in a review meeting, an email, or a call. Then the work fragments.

Someone writes file notes. Someone checks the existing IPS. The adviser decides what needs to change. A paraplanner drafts revised language. Compliance reviews the document. The client receives it for approval. The CRM, portfolio platform, and document management system need updates.

None of those tasks is mysterious. They are just easy to delay when a firm has 80, 200, or 800 active client relationships.

That delay matters. An outdated IPS can create a gap between what the client says they need and what the firm has documented. It also creates a poor client experience. The client has told you something significant, yet their paperwork may still describe a life that no longer exists.

AI can help by monitoring the inputs that indicate a changed circumstance, preparing a structured review pack, and drafting a revision for adviser and compliance approval. It should not make investment recommendations or change a client’s investment settings on its own. The point is to move the repetitive preparation work out of email inboxes and into a controlled process.

For a broader view of where this work sits across the firm, see Omni for financial advisory firms.

What triggers an IPS review

Firms often rely on annual or biennial reviews to refresh investment policy statements. That cadence is sensible, but it isn’t enough on its own. A material change can happen two weeks after the review.

The best automation starts with a clear list of review triggers. These should reflect your advice process, product set, jurisdiction, and compliance policy. They are not generic alerts that fire every time a client mentions the word “retirement.”

Common triggers include:

• A stated retirement date moves by more than a defined threshold
• A client changes employment status, sells a business, or receives a large equity payout
• A marriage, divorce, death, or new dependant changes the household position
• A new property purchase, health event, or education expense changes cash flow needs
• A withdrawal request changes the expected income drawdown pattern
• A portfolio risk assessment is completed with a meaningfully different result
• A client reports a change to income, debt, assets, tax position, or liquidity preference
• A portfolio drifts beyond the bands your firm has defined for review
• A recurring review is overdue and the client has had new contact since the last IPS sign-off

The word “material” needs to be defined by your firm. It can’t be left to an AI model to interpret without guardrails.

For example, an adviser may choose to review an IPS when a client’s intended retirement date changes by 12 months or more. Another rule may trigger a review when the client plans to withdraw more than a set percentage of their investable assets over the next 24 months. A risk tolerance result that moves one category may require an adviser assessment, but not always a full IPS revision.

Those rules turn a vague compliance concern into an operating system.

Where the manual work builds up

The work usually begins before anyone touches the IPS document.

An adviser is preparing for a quarterly or annual review. They need to find the last IPS, check recent portfolio performance, scan CRM notes, review emails, and remember what the client said last time. If a team has 10 advisers spending five to 10 hours a week on meeting preparation and post-meeting documentation, that is a large block of capacity tied up in searching, summarising, and rewriting.

After the meeting, the adviser or associate needs to turn a conversation into a defensible record. The detail that matters may be spread across a meeting transcript, handwritten notes, a client email, and updated fact-find documents.

Then comes the document process. Many firms use a combination of templates, Word files, PDFs, compliance checklists, and manual data entry. It is common for advice documentation to pass back and forth several times before the language is ready for issue. Paraplanner costs per advice document can land in the $3,000 to $8,000 range, depending on complexity and review requirements. An IPS update is often smaller than a full advice document, but it still competes for the same people and review queue.

The firm also has to avoid a different problem, an AI-generated document that sounds polished but contains a factual error, an unsupported assumption, or language that doesn’t match the approved template.

That’s why this use case isn’t “ask AI to write an IPS.” It is a controlled workflow with clear source data, clear escalation rules, and named owners at each decision point.

What an AI IPS update agent actually does

A useful AI agent works in stages. It gathers evidence, identifies possible changes, prepares a draft, and routes the work to people who hold advice and compliance responsibility.

Here is a practical end-to-end design.

1. Monitor approved client signals

The agent connects to the systems your firm already uses, subject to your security and permission requirements. That may include the CRM, review meeting transcripts, client portal submissions, fact-find platform, risk profiling system, document repository, portfolio reporting platform, and approved email channels.

It does not need unrestricted access to every message ever sent. Start with specific sources and events that are useful to the advice process.

A simple rule might look like this:

When a client review transcript, adviser note, or fact-find update contains a change to retirement timing, income needs, risk tolerance, household circumstances, or liquidity requirements, create an IPS review candidate.

The agent records the source, date, client, and relevant excerpt. It can then compare that evidence against the current IPS.

If a client says, “I now plan to stop work at 58 rather than 62,” the system should flag the retirement horizon as potentially changed. It should not assume the client needs a more conservative portfolio. That remains an adviser decision.

2. Compare the new information with the current IPS

This is where a structured IPS template matters.

The AI agent extracts the current documented values, such as investment objectives, time horizon, risk profile, liquidity requirements, target allocation, rebalancing approach, withdrawal assumptions, and known constraints.

It compares those fields with new evidence from the client interaction. The output should be a concise difference report, not a long narrative.

For example:

• Current IPS retirement date: June 2034
• New client statement: intends to retire by June 2030
• Current income requirement: no planned portfolio withdrawals before retirement
• New client statement: expects to draw $120,000 annually from portfolio assets from 2030
• Action: adviser review required before any IPS revision

That report makes it easier to separate a real trigger from normal client conversation. It also means the adviser does not have to search through 40 pages of history to understand why the alert exists.

The Meeting Prep Agent can support this stage by assembling portfolio data, recent communications, goal progress, and current advice documents into a one-page meeting brief. The adviser enters the meeting with the relevant changes already visible.

3. Ask for missing information

A good system doesn’t pretend it has all the answers.

A client might mention selling a business without stating the expected proceeds, timing, tax treatment, or intended use of funds. A risk questionnaire might indicate a lower appetite for volatility, while the adviser knows the client has a long enough horizon to require a deeper conversation.

The agent can prepare a short list of missing facts for the adviser or client service team. It can also send a secure, approved request to the client through the client portal, if that fits your process.

Questions should be specific:

• Has the expected retirement date changed, and is this a firm decision or an early assumption?
• What annual net income do you expect from your investment portfolio after retirement?
• Do you expect a major withdrawal in the next three years?
• Has your ability to tolerate investment losses changed, or has your comfort with volatility changed?
• Are there new dependants, property commitments, business interests, or debts to consider?

The Client Onboarding Agent is useful here as well. It can run a guided fact-find, collect supporting documents, and prepare a clean information pack. The same capability that prevents KYC from dragging out for 30 to 60 days can be applied to an existing client whose circumstances need to be refreshed.

4. Draft the IPS revision from approved language

Once the adviser has enough information, the agent can draft the revision.

It should work from your firm’s approved IPS template, clause library, terminology, and disclosure requirements. It should not use a generic web template. The draft should preserve unchanged sections and clearly mark the proposed changes.

A strong draft includes:

• The reason for review
• Source records supporting the update
• Revised goals, objectives, or time horizon
• Revised liquidity or withdrawal requirements where approved
• Any adviser-selected updates to investment approach or allocation parameters
• Items that were considered but did not change
• Required disclosures and client acknowledgement wording
• A version history showing who reviewed and approved each stage

The agent can also create a comparison view. Compliance and the adviser should be able to see the old wording beside the proposed wording, along with the client evidence that prompted the change.

The Advice Document Agent is designed for this type of work. It drafts SOAs, ROAs, and file notes from meeting transcripts and your firm’s compliance template. For IPS updates, the same approach can prepare the draft and supporting file note while leaving advice decisions with the authorised adviser.

5. Route the work through human approval

This part cannot be optional.

The adviser reviews the facts, decides whether the proposed changes are appropriate, and edits the draft as needed. Compliance reviews according to your firm’s established controls. The client receives a clear explanation and approval request where required.

Only after the right approvals should the workflow update the final document repository, CRM record, task status, and any downstream systems.

The audit trail should capture:

• What triggered the review
• Which source documents the agent read
• What fields were identified as changed
• Which version of the template was used
• Who made advice decisions
• Who approved the final document
• When the client accepted or acknowledged the update

This is a major reason to build the workflow properly rather than give staff an unmanaged AI chat tool. Your firm needs repeatable controls, not just faster wording.

The commercial case for automating IPS updates

For firms in the $1 million to $25 million revenue range, the leakage often isn’t visible as one line item. It sits across adviser admin time, paraplanner queues, delayed reviews, document rework, and client service staff chasing information.

For financial advisory firms, we usually see annual leakage in the $70,000 to $200,000 band when core advice workflows remain heavily manual. IPS updates are not the only contributor, but they expose the same underlying issue. Skilled people are doing document retrieval, data comparison, note writing, follow-up, and version control by hand.

The better question isn’t, “Can AI write this document?”

Ask these instead:

• How many IPS reviews are triggered late because a change was buried in a meeting note?
• How long does it take from a material client event to a reviewed document?
• How much paraplanner capacity is spent gathering information that already exists in your systems?
• How often do advisers need to re-read a client history before they can decide what changed?
• How many reviews require rework because source information was incomplete?

If you reduce the preparation and drafting component while preserving approval controls, advisers can spend more time on advice conversations. Paraplanners can focus on exceptions and complex cases. Clients receive evidence that the firm is responding to their changing circumstances.

You don’t need to automate every IPS update on day one. Start with one client segment and two or three triggers. Retirements, risk profile changes, and planned large withdrawals are sensible candidates because the evidence is usually identifiable and the review requirement is clear.

How to implement this without creating a compliance headache

The first step is process mapping, not model selection.

Document the current process from client event through to final record. Identify where the information begins, who owns each handoff, what the approved template requires, and which decisions need an adviser or compliance sign-off.

Then define the boundaries.

The agent may identify relevant evidence, ask approved questions, draft sections, and create tasks. It must not make a recommendation, alter portfolio instructions, claim a client has agreed to something they haven’t approved, or send advice documentation without review.

A practical pilot has five parts:

1. Select one client segment and one IPS template.
2. Define three to five material change triggers.
3. Connect only the data sources required for those triggers.
4. Create a draft-and-approval workflow with full version history.
5. Measure time from trigger to adviser review, document completion, and client acknowledgement.

Build a list of exceptions from the first 20 to 30 cases. Those exceptions are useful. They show where your policy language is vague, where source data is incomplete, and where the workflow needs a human checkpoint.

If you want help identifying the right starting point, Book a 60-min Omni Audit. In 60 minutes, we map the workflow, identify the highest-value automation opportunities, and outline the controls required to make them usable. No slide deck.

A better operating model for client change

An IPS should reflect the client’s current circumstances, not simply the last annual review date.

AI gives advisory firms a way to turn client conversations into structured review signals before they become compliance gaps or delayed work. It can find the relevant evidence, compare it with the current document, request missing facts, and prepare an IPS draft that follows your approved wording.

The adviser still owns the advice. Compliance still owns the control framework. The AI agent handles the repeatable preparation work that otherwise pulls experienced people into administration.

You can see how this fits into the wider operating model in the AI audit for financial advisory firms, or review the broader Omni advisory approach for firms looking beyond a single workflow.

If IPS updates are taking too long, start with the triggers and evidence your team already has. Then build the draft, review, and recordkeeping process around them. Book my Omni Audit and we’ll identify where the manual work is costing your firm time, margin, and client confidence.