What problem does it solve?
An advisor responsible for a large book of clients cannot watch every account every day. The signals are there (cash building up after a sale, a deposit about to mature, a portfolio drifting away from its mandate, a client who has stopped logging in), but they sit in different systems and surface too late, often when the client has already moved money or called a competitor.
Dashboards do not solve it: they show everything and prioritize nothing. What advisors need is a short list each morning of the few clients worth calling, why, and what to say, with the freedom to ignore a prompt that does not fit what they know about the client.
How does it work?
- Collect signals. Holdings, cash flows, maturities, product usage, service contacts, portfolio alignment and permitted external data are gathered per client.
- Score and rank. Predictive models and business rules score opportunities and risks (for example propensity to invest idle cash, risk of attrition), and a ranking step picks the few that matter most for each advisor.
- Explain. A language model turns each prompt into a short rationale and suggested talking points, citing the data behind it and, where relevant, the house view or an approved product.
- Advisor decides. The advisor acts, snoozes or dismisses the prompt, and the feedback is used to improve the ranking.
- Controls on the way out. Any product recommendation that follows still goes through the firm's suitability and product governance checks.
- Audience
- Employee facing
- Autonomy
- Assist
- Adoption
- Early adopters
- Channels
- Internal tools, Email
What is it worth?
Benchmarks are computed from the public deployments below: one data point per organization per KPI, with who made each claim.
| KPI | Median | Reported range | Data points | Claimed by |
|---|---|---|---|---|
| Employee adoption | Too few to pool | 80% | 1 | 1 organization |
Value drivers: Revenue growth, Employee productivity, Customer experience.
Indicative value
A wealth manager with 50,000 advised clients
USD 300,000 to USD 5 million
Additional annual revenue from acted prompts per year
How this is calculated
Formula: clients * promptsActed * conversion * revenuePerWin. The low scenario uses every low input, the high scenario every high input.
| Input | Low | High | Basis |
|---|---|---|---|
| Advised clients clients, clients | 50,000 | 50,000 | The reference firm. |
| Prompts acted on per client per year promptsActed, prompts per client per year | 0.2 | 0.5 | Editorial assumption, replace with your own advisor capacity and pilot data. |
| Share of acted prompts that lead to new business conversion, fraction of acted prompts | 0.1 | 0.2 | Editorial assumption. No public source on this page states a conversion rate for advisor prompts. |
| Annual revenue per converted prompt revenuePerWin, USD per conversion per year | 300 | 1,000 | Editorial assumption, for example fees on newly invested cash. Replace with your own margins. |
What it leaves out: Gross revenue before cannibalization, costs and the business that advisors would have won anyway. Measure it with a control group of advisors or clients before relying on it.
Who already uses it?
5 public deployments, strongest evidence first. Grades: A regulator or audit, B the organization itself, C vendor case study, D anonymous or estimate.
Citi
United States · Wealth and asset management · 2025
Citi Wealth launched two AI tools built by its Data, Analytics and Innovation team. AskWealth is a generative AI assistant that gives service teams, advisors and managers answers across the wealth business, so that advisors can reach market insights and research when clients ask questions; after a launch in Asia it became available to Citi Wealth colleagues worldwide. Advisor Insights is a dashboard of timely messages about market moves, portfolios and events, including Chief Investment Office insights, piloted with Citigold and Citi Private Client advisors in North America with a wider rollout planned for Q4 2025 and Q1 2026. Citi says the tools will save hours of time but published no figures.
No outcome disclosed.
UBS
United States · Wealth and asset management · 2025
UBS's US wealth management business runs STAAT Insights, a machine learning engine from its Smart Technologies and Advanced Analytics Team (STAAT) that surfaces client opportunities and alerts to financial advisors, such as shifting liquidity needs from a maturing CD, a concentrated stock position or a life event, and sends pre meeting client briefings with suggested talking points. In a December 2025 interview its chief data and analytics officer said 80% of US advisors actively use the engine. Two time saving figures UBS has published are about its AI tools in general, not STAAT Insights alone, so they are not recorded as metrics here: an advisor recruiting page says some advisors using UBS AI have saved three to four hours per client meeting, and the same executive conservatively estimated that US advisors save 10,000 hours a month by using AI to prepare for client meetings.
- Employee adoption: 80%, US advisors actively using STAAT Insights
"Let me give you some stats here: 80% of them are actively using that STAAT Insights engine."
Claimed by: organization
Morgan Stanley
United States · Wealth and asset management · 2023
Morgan Stanley Wealth Management built Next Best Action, an internal AI based engine that delivers timely, customized messages to clients and prospects, guided by the financial advisor. In March 2023, when it announced a strategic initiative with OpenAI to create a bespoke solution that its financial advisors would use, the firm listed it among its recent AI projects, alongside its Genome capability that uses data analytics and machine learning to personalize client communication. No outcome figures are published in that release.
No outcome disclosed.
CIMB Niaga
Indonesia · Banking · 2026
CIMB Niaga, one of Indonesia's largest banks, built purpose built AI agents with its AI Center of Excellence and Artefact on Google Cloud. The agents help bank staff offer tailored advice and proactive guidance matched to a customer's financial goals and life stage. No outcome figures were published.
No outcome disclosed.
JPMorgan Chase
United States · Wealth and asset management · 2025
J.P. Morgan's private client advisers use an internal generative AI tool, Coach AI, to find research and content for client conversations more quickly. The firm's asset and wealth management chief executive credited its AI tools, which pull clients' trading patterns and anticipate their questions, with helping advisers respond to clients during the April 2025 market sell off. Its asset and wealth management chief information officer said advisers find the right information up to 95% faster. The firm's aim of growing adviser client books by 50% over three to five years is a target, not a result.
- Search time reduction: up to 95%, time to find information for a client conversation
"Our advisers are finding the right information up to 95% faster - which means they spend less time searching and more time engaging in meaningful conversations with clients"
Claimed by: organization
How do you implement it?
A model agnostic playbook: what to prepare, the order to build in, and what goes wrong.
Data you need
- Client holdings, transactions and cash flows across accounts
- Product maturities, mandates and model portfolios
- CRM activity, service contacts and prior prompt outcomes
- Consent and marketing preference data per client
Systems to integrate
- Portfolio management and core banking systems
- CRM and advisor desktop
- Data platform for features and model scoring
- Suitability and product governance engine for any resulting recommendation
Complexity: High
The explanation layer is easy; the signals are not. It needs a clean client data model across banking, investment and CRM systems, models that are validated and monitored, fairness testing, and adoption work so advisors trust the prompts.
- 1
Start with a handful of high value signals
Pick three to five prompts with clear value and simple logic (idle cash above a threshold, maturing deposits, large inflows) before building propensity models.
- 2
Put the reason on every prompt
Each prompt shows the data that triggered it and a suggested talking point. Advisors ignore prompts they cannot explain to a client.
- 3
Measure against a control group
Hold out a random group of clients or advisors and compare outcomes, so the revenue story survives scrutiny from finance and risk.
- 4
Close the feedback loop
Capture act, snooze and dismiss with a reason, and retrain or retune rankings on that feedback every cycle.
- 5
Test for fairness and conduct risk
Check that prompts do not systematically favor higher fee products or neglect client segments, and review any prompt type that pushes a product.
Guardrails
- Prompts inform the advisor; nothing is sent to a client automatically
- Any product recommendation passes the suitability and product governance checks
- Contact respects marketing consent and preferences
- Fairness and conflict of interest review of prompt types and rankings
- Logged rationale for every prompt shown, with the data used
- Prompt and action data are not used to rate individual advisors
KPIs to instrument
- Prompt action rate and dismiss reasons per prompt type
- Conversion and revenue against a control group
- Client attrition in treated versus control books
- Weekly active advisors using the prompts
- Complaints or suitability exceptions linked to acted prompts
Human in the loop
The advisor decides whether and how to act on each prompt and owns the resulting advice. Business and risk owners approve new prompt types, and model risk validates the scoring models.
Common failure modes
- Prompt fatigue
- Too many low value prompts and advisors stop looking. Cap the list and retire prompt types with low action rates.
- Product push dressed as insight
- Rankings optimize for revenue and drift toward high margin products. Add conduct review and suitability checks.
- Unexplainable scores
- Advisors cannot tell a client why they called. Show the triggering data with every prompt.
- Credit shown as credit decision
- A prompt suggests a loan based on a score that is really a creditworthiness assessment, which brings high risk obligations. Keep credit decisions in the regulated credit process.
What are the risks and rules?
EU AI Act
Depends on design
Ranking investment and service prompts for an advisor is not listed in Annex III. It becomes high risk if the system evaluates the creditworthiness of natural persons, for example to decide which clients are offered lending (Annex III point 5(b)), so keep credit decisions out of the prompt engine. It is also high risk if the system itself is used to monitor or evaluate advisors' performance and behaviour, for example by scoring or ranking advisors on how they act on prompts (Annex III point 4(b)), so keep adoption reporting separate from performance management.
Rules that apply
Guidance
- ESMA public statement on the use of AI in the provision of retail investment services (European Securities and Markets Authority, Europe). Expects firms to act in the client's best interest when AI shapes recommendations, and to govern algorithmic bias and data quality.
- Principles to Promote Fairness, Ethics, Accountability and Transparency (FEAT) (Monetary Authority of Singapore, Asia Pacific). Fairness principles apply to models that decide which customers receive which prompts or offers.
- Consumer Duty (Financial Conduct Authority, Europe). Sets high standards of consumer protection across financial services and requires firms to put their customers' needs first, which applies to advisor prompts that lead to a sale to retail customers.
Controls to put in place
- Model inventory entries and validation for scoring models
- Fairness and conflict of interest testing of rankings per segment
- Logging of every prompt, its rationale and the advisor's action
- Suitability check on any recommendation that results from a prompt
- Periodic review of prompt types by business, risk and compliance
Frequently asked questions
- Does next best action for advisors actually get used?
- Where it is built into the daily workflow, it can be. In a December 2025 Financial Planning interview about UBS's US wealth management unit, UBS's chief data and analytics officer said 80% of advisors were actively using the STAAT Insights engine. In March 2023 Morgan Stanley listed its Next Best Action engine among its recent AI projects, describing it as an internally built engine that delivers timely, customized messages to clients and prospects, guided by the financial advisor.
- How do you prove the revenue effect?
- With a control group. Compare treated and untreated advisors or clients over the same period, because the clients an engine flags are often the ones advisors would have called anyway.
- Is this a high risk AI system?
- Not for investment and service prompts. It becomes high risk under the EU AI Act if it assesses the creditworthiness of individuals, or if the system itself is used to monitor or evaluate advisors' performance and behaviour, so credit decisions should stay in the regulated credit process and prompt data out of performance reviews.
How to cite this page
Blits.ai AI Use Case Library, "AI next best action prompts for wealth advisors", last verified 27 September 2026, https://www.blits.ai/ai-use-cases/next-best-action-for-advisors. Licensed under CC BY 4.0. Method: how we verify use cases.
Changelog
- 27 September 2026: First published