Dubai Robotaxi's 4-Million-Kilometre Milestone: A Practical Guide for Financial Advisors in UAE

Direct answer: Dubai's robotaxi program crossing 4 million kilometres driven with a 97% rider satisfaction rate is not really a transport story — it is proof that UAE residents will hand over a high-stakes, safety-critical decision to an autonomous system once trust is engineered correctly. For financial advisors, the practical takeaway is that client-facing AI (planning tools, portfolio monitoring agents, onboarding flows) can win the same trust if it is built with the same discipline: visible reliability, transparent limits, and consistent performance over volume, not a flashy demo.

According to UAE mobility reporting from August 2026, Dubai's autonomous robotaxi fleet has now driven more than 4 million kilometres on public roads with a 97% rider satisfaction rate. That is a meaningful volume threshold — not a pilot with a few hundred careful test rides, but millions of kilometres of ordinary traffic, unpredictable pedestrians, and everyday routes across a major city, with the overwhelming majority of riders reporting they were satisfied with the experience. We do not have a public breakdown of what specifically drove that satisfaction score, and we are not going to guess at sub-metrics that were not disclosed. What the headline figure does tell us, reliably, is that a large population of UAE residents has now had direct, repeated experience trusting an autonomous system with something as personal as their physical safety in traffic — and reported being satisfied with it at a rate few new technologies achieve this early. For an industry built entirely on trust, that is worth financial advisors' attention regardless of whether they ever touch a car.

What the Robotaxi Milestone Actually Signals

Four million kilometres is the kind of number that matters more for what it represents about adoption behavior than for the transport statistic itself. A single successful robotaxi ride proves a system can work under one set of conditions. Millions of kilometres of accumulated rides, spread across a real city's traffic patterns, weather, and driver behavior, proves something different: that the system holds up under volume, and that riders kept choosing to use it again after the novelty wore off. Satisfaction scores on a first ride are easy — everyone is curious, everyone is a little forgiving. A 97% satisfaction rate sustained across millions of kilometres is a signal about repeat trust, which is a much harder thing to earn and a much more useful thing to study.

This distinction matters because it maps directly onto a problem financial advisory firms in the UAE are already living with. Clients are increasingly comfortable using AI-driven tools somewhere in their financial life — a robo-advisor tier, an AI chat assistant on a banking app, algorithmic portfolio rebalancing — but comfort on the first interaction and comfort on the fiftieth interaction are different thresholds. The robotaxi program did not earn a 97% satisfaction rate by being impressive once. It earned it by being consistently reliable enough, ride after ride, that riders stopped treating each trip as an experiment and started treating it as routine. That is the exact trust curve a financial advisory practice needs any client-facing automation to travel, and it does not happen by accident — it happens because the underlying system was engineered for consistency, not just for a good demo.

Why "Autonomous" and "Trusted" Are Not the Same Milestone

It is worth being precise here rather than reaching for a bigger claim than the data supports. The 4-million-kilometre figure does not tell us that UAE consumers are now broadly comfortable with autonomy in every context, or that financial services clients specifically are ready to hand a robo-advisor full discretionary control over their portfolio. Financial decisions and physical transport decisions are not interchangeable categories of trust — one is a five-minute low-frequency safety judgment, the other is a long-horizon judgment about money that carries different emotional weight and different regulatory expectations. What the milestone does establish, honestly, is a regional pattern: when an autonomous system is deployed with visible safety guardrails, consistent performance, and enough real-world volume to prove itself, UAE residents are willing to extend meaningful trust to it. That pattern is directly relevant to financial advisors even though the specific domain is different, because the psychological mechanics of building that trust — transparency about what the system does and does not do, consistent reliability, and a visible track record — transfer across domains even when the trust threshold for the outcome itself does not.

Why This Matters Specifically to Financial Advisors in the UAE

Financial advisors occupy an unusual position in this conversation. They are not building autonomous vehicles, but they are increasingly building or adopting AI agents that sit close to the same trust category: systems that act on a client's behalf, make recommendations with real financial consequences, and need to earn continued use rather than a one-time sign-up. A financial planning chatbot that mishandles a client's risk tolerance question, or a portfolio-monitoring agent that sends a confusing or wrong alert, does not just lose that one interaction — it resets the client's trust calculus for every future AI touchpoint the firm offers, the same way a single bad robotaxi ride would have set back public trust in the entire program regardless of the 4 million kilometres that came before it.

The UAE market makes this dynamic sharper than in most regions. Dubai and the wider UAE have positioned themselves deliberately as an environment where advanced automation is visible, government-endorsed, and part of daily life — the robotaxi program itself is emblematic of that positioning, not an outlier. UAE clients, particularly the wealth management and family office segments concentrated in Dubai and Abu Dhabi, are more likely than clients in many other markets to have already formed an opinion about AI-driven services from firsthand experience with autonomous systems elsewhere in their daily life. That cuts both ways for financial advisors. On one hand, it means the market is primed to be receptive to well-built AI tools rather than skeptical of the concept itself — the robotaxi milestone is evidence the appetite for good automation exists here. On the other hand, it raises the bar: a UAE client who has ridden in a robotaxi with a 97% satisfaction track record has a real, lived reference point for what "well-engineered AI system" feels like, and a clunky, poorly-explained AI feature bolted onto an advisory practice's client portal will read as noticeably behind that bar, not ahead of it.

The Regional Trust Gap Advisors Cannot Ignore

There is a specific gap worth naming plainly: transport-sector AI in Dubai has now had a multi-year, highly public runway to build trust through visible, government-backed rollout and steady volume. Financial services AI adoption, by contrast, tends to happen more quietly, client by client, tool by tool, without the same kind of cumulative public proof point. That means individual advisory firms are mostly on their own to build the trust that the robotaxi program built collectively and publicly. A firm cannot borrow the robotaxi program's 97% satisfaction score, but it can borrow the underlying lesson: trust in an autonomous or AI-assisted system is built through demonstrated reliability at volume, transparent boundaries around what the system will and will not decide, and a track record clients can actually observe rather than a marketing claim they are asked to take on faith.

What Changes in Practice for an Advisory Firm's Website and Client Tools

For an advisory practice, this trend translates into concrete product and communication decisions rather than an abstract shift in "AI sentiment." A few things change in practice.

First, any AI-driven feature on a firm's website or client portal — a planning calculator, a document intake assistant, a query-answering chat feature — needs to be explicit about its boundaries. The robotaxi program did not succeed by pretending to be a human driver; it succeeded by being a well-defined autonomous system operating within known parameters, with humans and oversight systems clearly present around it. A financial AI tool that is vague about whether it is giving general information versus personalized advice, or that does not make clear when a human advisor is looped in, creates exactly the ambiguity that erodes the trust this milestone shows is achievable when handled correctly.

Second, consistency at volume matters more than a single polished feature. A firm that launches one impressive AI demo — say, an eye-catching portfolio simulator — but leaves the rest of its digital experience clunky, slow, or inconsistent is working against the same trust mechanics the robotaxi program leaned into. Clients build confidence from the sum of repeated, reliable interactions, not from a single showcase moment. That argues for treating AI-assisted client tools as an ongoing product to maintain and refine, not a one-off launch.

Third, and most directly relevant to service selection, the tools worth investing in are the ones that behave like genuine agents doing bounded, well-defined work reliably — scheduling, document collection, meeting prep summaries, routine client-status updates — rather than tools that try to simulate open-ended human judgment and inevitably disappoint when they hit their limits. This is exactly the design philosophy behind AI Agents & Automation work: build automation that is honest about its scope, reliable within that scope, and layered with clear human handoff points, the same architecture that let a robotaxi program earn a 97% satisfaction score without ever claiming to be more capable than it actually is.

What to Do About It: A Practical Path for Advisory Firms

Turning this into action does not require an overhaul of a firm's entire client experience. It requires a deliberate, staged approach to where automation appears and how its reliability gets proven before it is trusted with more.

Start by auditing where AI or automation already touches the client journey — intake forms, scheduling, document requests, portfolio update emails — and identify which of those touchpoints are inconsistent or unclear about what is automated versus human-reviewed. Inconsistency at this stage is where trust erodes fastest, and it is usually the cheapest thing to fix. From there, prioritize automating narrow, well-bounded, repetitive tasks first — the equivalent of a fixed route rather than an unpredictable one — before layering in anything resembling advisory judgment. A meeting-prep agent that reliably pulls a client's recent portfolio activity and flags anything notable ahead of a call is a strong first build: bounded, useful, and easy to evaluate for reliability. An agent that tries to draft investment recommendations without a defined review step is a much harder trust problem to solve early, and attempting it before the basics are solid risks the same kind of setback a single bad robotaxi incident would have caused this program.

Alongside the automation itself, firms should invest in the surrounding content and communication that makes the system's reliability visible to clients — much the way public reporting on the robotaxi program's kilometres and satisfaction rate became part of how trust was built collectively. That might mean a clear explainer page on how the firm uses AI tools, what a client can expect from them, and where a human advisor remains the decision-maker. It also touches how a firm talks about its own capabilities more broadly: a firm explaining new technology-driven services to prospective clients benefits from the same production values that make any message land — for firms building out client education content, the same fundamentals covered in Video Marketing Agency: Why Your Brand Needs One in 2026 apply directly, since a short, well-produced video explaining "how our AI planning tool works and where a human reviews it" does more to build the kind of trust this milestone illustrates than a paragraph of text ever will.

Firms should also resist the temptation to measure early success by adoption numbers alone. A high sign-up rate for a new client-facing AI feature means very little if the actual usage pattern shows clients trying it once and reverting to the old channel — that is the financial-services equivalent of a robotaxi ride that impresses on the way in but leaves the rider unwilling to book a second one. The more meaningful metric, borrowed directly from how the robotaxi program's own credibility is measured, is repeat usage over a meaningful volume: are the same clients coming back to the tool for the second, fifth, and twentieth interaction, and is the error or correction rate staying low as that volume grows. A firm that tracks this from the first week of launch, rather than waiting for a quarterly review, catches reliability problems while they are still cheap to fix and before they have had the chance to color a client's broader opinion of the firm's technology.

Security and identity verification deserve the same deliberate treatment. As advisory firms add more digital touchpoints — client portals, mobile check-ins, remote onboarding — the authentication layer becomes part of the trust equation the same way visible safety systems are part of what made the robotaxi program credible. The considerations laid out in Biometric Authentication in Mobile Apps: Face ID, Fingerprint, and Beyond are directly applicable for a UAE advisory firm building or upgrading a client mobile app, since a client who trusts an autonomous vehicle with their physical safety will reasonably expect their financial advisor's app to protect their account with at least that level of engineered care.

Finally, firms should keep an eye on the regulatory direction other markets are taking with AI generally, since it shapes client expectations even outside the UAE's own framework. Clients increasingly compare notes across jurisdictions, and the kind of structured thinking described in Australia's AI Regulation Roadmap: Inside the New National Standards and Office of AI is a useful reference point for how governance-minded AI deployment is being formalized elsewhere — a pattern that reinforces the same lesson the robotaxi milestone teaches: visible structure and accountability around an AI system are what let trust scale past the early-adopter stage.

Where This Kind of Work Typically Falls in Scope and Cost

Advisory firms considering their first serious AI-assisted client tool often ask where the investment lands. The honest answer depends on scope, but it maps reasonably well onto three tiers of engagement.

Tier	Typical scope for a financial advisory firm	Starting price
Essential	A single well-defined automation — e.g., an intake or scheduling agent, or a client-facing FAQ assistant with clear human handoff	$1,000
Growth	Multiple connected agents — meeting-prep summaries, portfolio-update alerts, document collection — plus the client-facing explainer content around them	$2,000
Enterprise	A full client-facing AI layer across the practice, including custom agent workflows, secure client portal integration, and ongoing reliability monitoring	$4,000+

Most firms taking their first real step into this kind of automation start at the Essential or Growth tier, proving out one reliable, bounded use case before expanding — the same staged path the robotaxi program itself took from limited routes to millions of kilometres.

Key Takeaways

• Dubai's robotaxi program passing 4 million kilometres with 97% satisfaction is a regional trust signal, not just a transport milestone — UAE clients have direct experience trusting well-engineered autonomous systems.
• Financial advisory clients extend trust to AI tools the same way: through consistent reliability at volume, not a single impressive demo.
• Any client-facing AI feature should be explicit about its boundaries — what it decides versus what a human advisor reviews — to avoid the ambiguity that erodes trust fastest.
• Start automation with narrow, well-bounded, repetitive tasks (scheduling, document intake, meeting-prep summaries) before layering in anything resembling advisory judgment.
• Pair automation with visible, honest communication about how it works, and treat authentication and security as part of the same trust-building system.
• Most firms should scope their first AI agent work at the Essential or Growth tier and expand once reliability is proven, mirroring how the robotaxi program itself scaled gradually.

Dubai's robotaxi milestone is a useful reminder that UAE clients are ready to trust well-built autonomous systems — the opportunity for financial advisors is building that same discipline into their own client-facing tools rather than assuming trust will follow automatically. If you want help figuring out where to start, book a meeting with our team.

Frequently Asked Questions

What exactly did Dubai's robotaxi program achieve to reach this milestone?

According to UAE mobility reporting from August 2026, the fleet surpassed 4 million kilometres driven on public roads with a 97% rider satisfaction rate. That figure reflects accumulated real-world operation, not a limited pilot, and it is the specific data point this article is grounded in.

Why should a financial advisor care about a transport milestone?

Because the underlying story is about trust in autonomous systems, not about cars specifically. The same trust-building mechanics — consistency, transparency, and proven reliability at volume — apply directly to any AI tool a financial advisory firm puts in front of clients.

Does this milestone mean UAE clients are ready for fully autonomous financial advice?

No, and the article does not claim that. Physical-safety trust and financial-decision trust are different categories with different stakes and different regulatory expectations. What the milestone shows is that the trust-building process works when done deliberately, not that the two domains are interchangeable.

What is an AI agent in the context of a financial advisory practice?

An AI agent is a system that can carry out a defined task with some degree of autonomy — for example, gathering a client's recent portfolio activity and preparing a summary before a meeting — rather than simply answering a single question. It differs from a basic chatbot by acting across multiple steps toward a goal.

What is the difference between AI automation and full AI-driven advice?

Automation handles bounded, repetitive tasks like scheduling, document collection, or status updates. AI-driven advice would imply the system is making investment judgments, which carries far higher regulatory and trust requirements and is not the recommended starting point for most firms.

Why does the article recommend starting with narrow, bounded tasks?

Because trust is built incrementally through demonstrated reliability, the same way the robotaxi program built trust route by route before scaling to millions of kilometres. A narrow task is easier to get right consistently, and getting it right consistently is what earns the right to automate more.

How does the 97% satisfaction figure relate to consumer trust more broadly?

It shows that a large population sustained trust across repeated use, not just a first try. That distinction — repeat trust versus first-impression trust — is the more useful lesson for any business introducing AI to clients, financial advisory or otherwise.

What kind of AI tools are UAE wealth management clients most likely to already be exposed to?

Government-backed and highly visible programs like the Dubai robotaxi fleet, along with AI features embedded in major banking and government digital services, mean many UAE clients already have some reference point for what a well-engineered AI experience feels like.

Is this trend specific to Dubai, or does it apply across the UAE?

The specific milestone is a Dubai program, but the broader pattern of visible, government-supported AI and automation adoption extends across the UAE, making the trust dynamics described here relevant to advisory firms serving clients throughout the country.

What is the first step a financial advisory firm should take toward this kind of automation?

Audit existing client touchpoints — intake, scheduling, document requests, status updates — for inconsistency or lack of clarity about what is automated. Fixing that inconsistency is typically the cheapest and highest-impact first move before adding new AI features.

How much does a basic AI agent implementation typically cost?

Based on Scult's service tiers, a single well-defined automation such as an intake or scheduling agent typically starts at the Essential tier, around $1,000, depending on scope and integration requirements.

What does the Growth tier typically include for a financial advisory firm?

The Growth tier, starting around $2,000, typically covers multiple connected agents — such as meeting-prep summaries, portfolio-update alerts, and document collection — along with the client-facing content explaining how those tools work.

When does a firm need the Enterprise tier?

The Enterprise tier, starting at $4,000+, fits firms building a full client-facing AI layer across the practice, including custom workflows, secure portal integration, and ongoing reliability monitoring rather than a single point tool.

How long does it typically take to build and launch a first AI agent for client-facing use?

Timelines vary by scope and integration complexity, but a single bounded automation at the Essential tier is generally the fastest to design, build, and validate, since it involves fewer systems and less judgment logic than a multi-agent Growth or Enterprise build.

What are the biggest risks in adding AI tools to a financial advisory client experience?

The main risks are ambiguity about what the system decides versus what a human reviews, and inconsistency across touchpoints that undermines confidence built elsewhere. Both are avoidable through clear scoping and transparent client communication rather than technical limitations.

Does adding AI agents to client tools create new compliance considerations?

Any automation that touches client data or communicates information relevant to financial decisions should be reviewed against the firm's existing compliance framework, particularly around what the system is permitted to state as fact versus general information. This is a scoping conversation to have early in any build.

How does biometric authentication relate to this trend?

As advisory firms add more digital client touchpoints, the authentication layer becomes part of the overall trust picture, similar to how visible safety systems supported trust in the robotaxi program. Strong authentication, such as the approaches described in Scult's biometric authentication guide, protects the credibility of every other AI feature built on top of it.

What role does video content play in building trust around new AI features?

A short, clear video explaining how an AI tool works and where a human advisor is involved tends to build more client confidence than written disclosures alone, because it makes the system's boundaries visible rather than assumed.

Should a financial advisory firm publicize its use of AI tools to clients?

Generally yes, framed honestly. Being transparent about what is automated and what remains human-reviewed tends to build more trust than staying quiet about it, based on the same pattern that made the robotaxi program's public reporting effective.

What happens if an AI tool underperforms after launch?

A single poor experience can reset client trust in every AI touchpoint the firm offers, not just the one that failed. This is why the article recommends prioritizing consistency and bounded scope over an ambitious but fragile first feature.

How is this different from a robo-advisor platform?

A robo-advisor typically automates portfolio construction and rebalancing based on preset rules. The automation this article recommends starting with is narrower still — operational and communication tasks — as a lower-risk way to build the same reliability track record before considering anything closer to a robo-advisor's scope.

What does "hybrid trust" mean in this context?

It refers to combining automated efficiency with clear human oversight points, so clients get the speed of automation without losing the judgment layer they expect from a financial advisor. It is the same balance the robotaxi program strikes with human oversight systems monitoring the autonomous vehicles.

Are UAE regulators addressing AI use in financial services specifically?

The UAE has been active in shaping AI governance broadly, and firms should track relevant guidance as it applies to financial services specifically. This article does not detail UAE-specific financial AI regulation, since that was not part of the sourced trend, but it is a reasonable area for firms to monitor alongside frameworks emerging in other markets.

Why does the article reference Australia's AI regulation roadmap for a UAE-focused topic?

Clients and firms increasingly look across jurisdictions for signals about how AI governance is maturing. Australia's approach is a useful structural reference point for the kind of formal governance thinking that reinforces the same trust principles the robotaxi milestone illustrates, even though the regulatory frameworks themselves are separate.

What is the most common mistake firms make when adopting AI client tools?

Launching one polished, high-visibility feature while leaving the rest of the digital client experience inconsistent. Trust is built cumulatively across every touchpoint, not from a single showcase moment.

Can a small, independent financial advisory firm realistically use this kind of automation?

Yes. The Essential tier is specifically designed for a single well-scoped automation, which is an accessible starting point for smaller practices rather than requiring the full multi-agent build of the Enterprise tier.

How does this trend affect client onboarding specifically?

Onboarding is often one of the most repetitive, document-heavy parts of the client relationship, making it a strong candidate for early automation — provided the handoff to a human advisor for anything judgment-related remains clear throughout.

What is the difference between AI automation and AI agents as terms?

Automation broadly refers to systems executing predefined steps, while AI agents specifically imply some degree of autonomous decision-making within a bounded scope, such as deciding which of several relevant documents to flag for a client meeting. The distinction matters when scoping a project, since agent-based work generally involves more design around decision boundaries.

How should a firm measure whether an AI tool is actually earning client trust?

Track consistency over time rather than one-off feedback — repeat usage, low error or correction rates, and whether clients raise fewer questions about how the tool works as they use it more. This mirrors how the robotaxi program's satisfaction score reflects sustained use rather than first impressions.

Does this trend apply equally to independent advisors and larger wealth management firms?

The trust-building principles apply equally, though the scale of implementation differs — an independent advisor might start with a single Essential-tier tool, while a larger firm may move toward a Growth or Enterprise-tier build across multiple client-facing systems.

What should be in a client-facing explainer about an advisory firm's AI tools?

At minimum, a plain-language description of what the tool does, what it does not do, and at what point a human advisor is involved. This directly addresses the ambiguity that most often undermines client trust in AI features.

How does mobile app security connect to this trend if a firm doesn't have a robotaxi-style product?

Any digital client touchpoint carries the same underlying trust requirement — reliability and visible safeguards — regardless of industry. A firm's mobile app authentication is one of the most direct places clients experience that trust firsthand.

What is the risk of moving too fast into advisory-judgment automation?

Attempting to automate open-ended advisory judgment before proving reliability on simpler tasks risks a visible failure that undermines confidence in every other AI feature the firm has built, similar to how a single serious incident could have set back the robotaxi program's public trust regardless of its accumulated kilometres.

Should a firm involve compliance or legal review before launching a client-facing AI agent?

Yes, particularly for anything that touches client data, financial information, or communications that could be construed as advice. Early involvement avoids costly rework later in the build.

What ongoing costs should a firm expect beyond the initial build?

Beyond the initial tiered investment, firms should budget for ongoing monitoring and refinement to maintain reliability as client usage grows, similar to how the robotaxi program required continuous operation and oversight to sustain its satisfaction rate across millions of kilometres.

How does this milestone affect client expectations for response speed?

Clients increasingly compare digital experiences across industries, so an advisory firm's automated tools are implicitly benchmarked against well-engineered systems clients already trust elsewhere, raising the baseline expectation for responsiveness and clarity.

What is the role of transparency in building AI trust with financial clients?

Transparency about what a system does, does not do, and where a human is involved is the single most consistent factor separating AI tools that build trust from those that erode it, based on the pattern this article draws from the robotaxi program's approach.

Is there a risk of over-explaining AI tools to clients?

Some risk exists if explanations become overly technical, but the greater risk in financial services is under-explaining, since ambiguity about automated decision-making tends to concern clients more than a slightly longer explanation would.

How do you scope a first AI agent project with a development partner?

Start by identifying one specific, repetitive, well-defined task with a clear success measure, then define exactly where human review fits before any build begins. This scoping conversation is typically part of an initial engagement regardless of tier.

What happens during an AI Agents & Automation engagement at the Essential tier?

Work at this tier typically focuses on scoping and building a single automation — such as a scheduling or intake agent — including defining its boundaries and testing it against realistic client scenarios before launch.

Can existing client portal software be integrated with new AI agents, or does it require a rebuild?

In most cases integration is possible without a full rebuild, though the scope depends on the existing platform's flexibility. This is typically assessed during initial scoping rather than assumed either way.

What is a reasonable first success metric for a new client-facing AI tool?

Consistency of correct output over a defined volume of real interactions is a more meaningful early metric than raw usage numbers, mirroring how the robotaxi program's credibility rests on sustained performance across millions of kilometres rather than a single successful ride.

How does this trend relate to broader AI adoption trends in financial services generally?

It reinforces a pattern seen across financial services globally: client comfort with AI grows fastest when firms are transparent about scope and consistent in execution, rather than when firms lead with ambitious but unproven capability claims.

Should smaller advisory firms wait for larger competitors to prove out AI tools first?

Waiting has a cost, since clients are already forming expectations from experiences like the robotaxi program regardless of when a given firm moves. Starting with a low-risk, well-scoped Essential-tier build lets smaller firms build their own track record without waiting for the market to settle.

What is the biggest misconception firms have about AI agent adoption?

That adoption requires an all-at-once transformation. The more reliable path, illustrated by how the robotaxi program itself scaled gradually from limited routes to millions of kilometres, is incremental expansion from one proven use case to the next.

How does this connect to a firm's broader digital marketing strategy?

Client-facing AI tools and the content explaining them are part of the same trust-building system as a firm's broader marketing, since prospective clients evaluating an advisory firm increasingly factor in how modern and transparent its digital experience feels.

What should a firm avoid saying when marketing a new AI-assisted service to clients?

Avoid language implying full autonomy or that a human is no longer needed in the loop unless that is precisely true, since overstating capability creates the exact expectation gap that erodes trust once a client experiences the tool's actual limits.

Is 97% satisfaction a realistic benchmark for a financial advisory AI tool to aim for?

It is a useful directional benchmark for what sustained, well-engineered automation can achieve, though financial services carries different stakes and measurement methods than a transport satisfaction survey, so firms should define their own realistic success criteria rather than importing the figure directly.

How should a firm handle client data privacy when deploying AI agents in the UAE?

Any agent that touches client financial data should be scoped with data handling and storage in mind from the start, including where data is processed and who can access it. This is a standard part of scoping a build responsibly and should be addressed before launch rather than retrofitted afterward.

What is the next step for a firm that wants to explore this further?

The most practical next step is a scoping conversation to identify one bounded, high-value task worth automating first, which is exactly the kind of discussion a firm can start by booking a meeting with a team experienced in this type of build.