\n At the Hubbis Malaysia Wealth Management Forum 2026, Damien Piper, Executive Director – Growth at Unique AI, examined how agentic AI is moving from generic productivity use cases into more specific, workflow-driven applications across wealth management, asset management and investment banking.
\n
\nPiper\u2019s presentation focused on the practical realities of deploying AI in regulated financial institutions, including hallucination controls, document intelligence, client data integration, security, on-premise deployment, compliance workflows and front-office augmentation.
\n
\nThe central message was that AI in wealth management only creates value when it is precise, secure, connected to relevant data and designed around the work that relationship managers, compliance teams and operations staff actually need to perform.\n <\/p>\n
Key Takeaways<\/p>\n
\n\tGeneric AI Is Not Enough For Wealth Management: Piper said standard AI tools can support broad productivity, but they do not solve the harder problems that require deep understanding of financial documents, client data, portfolios and house views.
\n\tPrecision Drives Adoption: If AI outputs are not accurate and context-specific, users will not adopt the tools, limiting the business value that financial institutions expect.
\n\tHallucination Was An Early Barrier: Unique AI had to develop controls such as hallucination-checking and prompt-extension engines to make AI usable in regulated financial services environments.
\n\tClient Data Is Essential To Real Use Cases: Piper said AI becomes materially more useful when it can work with CRM data, portfolio systems, market research and internal policy documents.
\n\tAI Should Augment Relationship Managers: Piper stressed that the aim is not to automate away the RM, but to support advisers through agentic workflows that perform specific tasks alongside them.
\n\tSecurity And Deployment Models Matter: Wealth and client data are difficult to place in generic cloud environments, making guardrails, secure zones and on-premise deployment important.
\n\tAgentic AI Can Support Front, Middle And Back Office: Use cases include investment insights, KYC, onboarding, meeting summaries, fund analysis, due diligence, RFP drafting, reconciliation, compliance and source-of-wealth narratives.
\n\tDomain-Specific Document Understanding Is Critical: Financial factsheets, term sheets, policies, directives and research reports require AI systems that can interpret wealth and asset management content accurately.
\n\tCommunity Feedback Helps Shape Product Development: Piper said Unique AI works with clients through strategy board discussions, allowing institutions to influence the future development of the platform.
\n\tAI Platforms Need To Complement Existing Tools: Large institutions may still use enterprise productivity tools, but specialist platforms can operate alongside them on wealth-specific workflows.
<\/p>\n
\u00a0<\/p>\n
Piper began by describing Unique AI\u2019s origins in Switzerland around five and a half years ago, as generative AI began to change expectations around how technology could support financial services.<\/p>\n
The firm, he said, was built around the recognition that deploying AI in wealth managers, asset managers and investment banks would be fundamentally different from deploying generic AI tools. Financial institutions operate with client confidentiality, regulated workflows, complex documentation and high accuracy requirements.<\/p>\n
Early challenges were substantial. Hallucination was widespread, meaning AI systems could produce confident but inaccurate outputs. Unique AI therefore had to build controls designed to check, extend and constrain prompts so that responses could become more reliable.<\/p>\n
The next challenge was document understanding. Banks needed AI to work with factsheets, internal policies, corporate documentation, research and client materials. But these documents are often highly structured, domain-specific and difficult for generic models to interpret properly.<\/p>\n
Piper said this is particularly clear in wealth and asset management, where small details in a factsheet, rating, policy document or investment report can materially change the output.<\/p>\n
\u201cYou cannot just drop a generic AI tool into a wealth management environment and expect it to understand the business,\u201d Piper said. \u201cThe documents, the data and the workflows are too specific.\u201d<\/p>\n
Why Precision Matters<\/p>\n
Piper said precision became the central requirement.<\/p>\n
For AI to become useful in wealth management, it needs to connect unstructured information, such as PDFs, factsheets, emails and policy documents, with structured data from CRM, portfolio and market research systems. It must then recall that information accurately and apply it within the correct business context.<\/p>\n
If the system is not precise, users lose confidence. If users do not trust the output, adoption falls. If adoption falls, the business value disappears.<\/p>\n
This shaped Unique AI\u2019s architecture. Piper said the firm had to redesign its platform from the ground up to support higher-precision retrieval, data integration and agentic workflows.<\/p>\n
\u201cThe lesson was simple,\u201d he said. \u201cWithout precision, there is no adoption. Without adoption, there is no business value.\u201d<\/p>\n
For wealth managers, this distinction matters because the most valuable AI applications are not necessarily chat-based. They are task-based. They help a user complete a process, prepare advice, review a portfolio, analyse a document, draft a report or capture regulatory information.<\/p>\n
The Agentic AI Platform<\/p>\n
Piper described Unique AI as an agentic AI platform for financial services, using retrieval-augmented generation and agentic workflows to improve customer experience, agility and operational efficiency.<\/p>\n
The platform is designed to connect to embedded data, including documents, wikis and directives from SharePoint, client insights, Salesforce and portfolio data, public data sources, and real-time integrations such as FactSet and Quartr.<\/p>\n
It includes an agentic framework, model orchestration, connectors, data and access management, enterprise security and compliance, and modular architecture. It is also designed to be LLM and infrastructure agnostic, allowing institutions to work with models such as GPT, Mistral, LLaMA, Claude or their own hosted models.<\/p>\n
This flexibility matters because financial institutions differ significantly in their technology environments, risk appetite and cloud policies. In some cases, especially where client and wealth data are involved, firms may require on-premise deployment or secure internal environments.<\/p>\n
Piper said this is one reason financial services AI is more complex than consumer AI.<\/p>\n
\u201cChatting with a public AI model is easy,\u201d he said. \u201cPutting AI safely inside a bank, connected to client data and business processes, is a much harder problem.\u201d<\/p>\n
Real-World Deployment In Wealth Management<\/p>\n
Piper said Unique AI now works with more than 40 financial services clients, with around 30,000 users across front, middle and back office functions. The firm\u2019s client base includes major banks and financial institutions, and its platform is being used across Europe and Asia.<\/p>\n
Unique AI opened in Asia around a year ago, with Singapore serving as its regional hub. Piper said a number of implementations are now live in Singapore and Hong Kong, including multilingual use cases involving Chinese, Malay and Bahasa.<\/p>\n
The point, he said, is that the technology is no longer theoretical. It is being used in real financial institutions to address specific operational and advisory problems.<\/p>\n
In the front office, the need is not simply for chat or email summarisation. Relationship managers need agents that can retrieve CRM data, compare investment information, access market research, generate investment proposals and explain recommendations using the institution\u2019s own house view.<\/p>\n
This is where domain-specific design becomes important. An investment proposal generated from generic market views may not match the bank\u2019s approved perspective, product shelf or client suitability standards.<\/p>\n
\u201cThe front office does not need another chatbot,\u201d Piper said. \u201cIt needs agents that can do the work around the RM, using the bank\u2019s own data, language and controls.\u201d<\/p>\n
Front Office Use Cases<\/p>\n
Piper outlined several front-office applications for agentic AI.<\/p>\n
Client research agents can combine internal and public data to profile clients, identify opportunities and support personalised communication. Meeting insight tools can capture, transcribe, summarise and analyse client meetings, then help record key outputs in CRM systems.<\/p>\n
Investment insight agents can create personalised investment pitches, next-best-action prompts and portfolio commentary. Portfolio review tools can explain trends, allocations and changes over time, while fund insight capabilities can benchmark and compare criteria and performance across products.<\/p>\n
Customer chatbots can support basic investment and servicing questions, while investment advice tools can help advisers connect client objectives, portfolio data and market research into more useful recommendations.<\/p>\n
These tools are not positioned as replacements for advisers. They are designed to reduce preparation time, improve consistency and help RMs focus on client engagement.<\/p>\n
Middle Office And Compliance Applications<\/p>\n
Piper said demand is also rising from compliance and regulatory functions.<\/p>\n
KYC, onboarding and source-of-wealth processes are particularly well-suited to AI because they involve large volumes of documentation, repeated checks and structured outputs. Unique AI\u2019s platform includes tools that can pre-fill onboarding questionnaires, flag missing documents, identify discrepancies and draft structured KYC overviews.<\/p>\n
Source-of-wealth narratives are another use case. These can be time-consuming for advisers and compliance teams, especially where a client\u2019s wealth journey involves multiple companies, transactions, inheritances, liquidity events or jurisdictions. AI can help assemble clear, compliant narratives while retaining human oversight.<\/p>\n
The platform can also support account reviews, suitability checks, regulatory procedures and document-driven compliance work. Directive agents can retrieve information from internal policies, procedures and regulations, while procedure update tools can analyse and adapt internal documents to reflect local regulatory requirements.<\/p>\n
Piper said these processes are often painful for institutions, but they are exactly where AI can remove manual friction.<\/p>\n
\u201cOnboarding, KYC and source-of-wealth work are document-heavy and process-heavy,\u201d he said. \u201cThat makes them natural areas for agentic AI to support staff, provided the controls are right.\u201d<\/p>\n
Back Office And Operational Efficiency<\/p>\n
Piper also highlighted increasing demand from back-office functions.<\/p>\n
Financial institutions process large numbers of term sheets, factsheets, fund reports, transaction documents, broker reports and unstructured PDFs. Much of this information needs to be ingested, extracted, reconciled, compared and entered into other systems.<\/p>\n
AI can support reconciliation, due diligence, RFP and DDQ drafting, macro broker research synthesis, earnings call analysis, ESG analysis and supplier compliance checks. Scheduled workflows can also run recurring tasks, such as weekly report generation or daily data refreshes.<\/p>\n
The value comes from turning unstructured information into usable outputs, while embedding the process into daily workflows.<\/p>\n
Piper said this is where agentic AI differs from conventional automation. Rather than automating one narrow task, agents can handle multi-step workflows that involve retrieving information, reasoning over documents, producing a draft or output, and reporting what has been done.<\/p>\n
Community-Led Product Development<\/p>\n
Piper said Unique AI has learned that collaboration between clients can accelerate development.<\/p>\n
The firm runs strategy board meetings where clients discuss the direction of wealth management and vote on future platform priorities through a process the company calls \u201cUnicopoly\u201d. This allows institutions in different markets to share use cases and influence the product roadmap.<\/p>\n
He said this global feedback loop matters because a solution developed for one market may become useful in another. A feature designed for a project in Korea, for example, may later become relevant for institutions in Switzerland or Southeast Asia.<\/p>\n
This network effect is important in an area where technology is changing quickly. Financial institutions want to know not only whether a vendor can solve today\u2019s problem, but whether it can keep adapting as AI capability, regulation and business expectations evolve.<\/p>\n
\u201cWorking with clients in one market helps us build better tools for clients elsewhere,\u201d Piper said. \u201cThe use cases travel faster than people expect.\u201d<\/p>\n
Complementing Enterprise AI Tools<\/p>\n
Piper said large financial institutions will continue to use tools such as Copilot or enterprise GPT platforms for broad productivity tasks, including work in Word, Outlook and general document drafting.<\/p>\n
However, he argued that these systems do not solve the harder wealth management problems by themselves. The most valuable use cases require secure access to client data, portfolio systems, research platforms, house views, compliance processes and the institution\u2019s own workflows.<\/p>\n
Unique AI\u2019s role, he said, is to complement these enterprise tools by operating in a secure environment inside the bank on the problems that require domain expertise.<\/p>\n
That distinction is important for wealth managers evaluating AI strategies. Generic productivity tools may improve individual efficiency, but specialist platforms are needed where workflows involve regulated advice, client information, investment suitability, KYC, source-of-wealth checks or portfolio-specific recommendations.<\/p>\n
Building AI Around The Work<\/p>\n
Piper\u2019s conclusion was that the value of AI in wealth management depends on whether it is designed around real work.<\/p>\n
The industry does not need AI for its own sake. It needs technology that can reduce manual effort, improve accuracy, help advisers prepare better, support compliance teams and make operations more efficient.<\/p>\n
For Malaysian wealth managers, the implications are practical. AI adoption should not begin with a generic chatbot. It should begin with the workflows that create the most friction: onboarding, KYC, client servicing, proposal generation, portfolio review, compliance documentation, research synthesis and operational reconciliation.<\/p>\n
Piper\u2019s message was that AI agents can support these processes continuously, but only if the architecture is secure, the data connections are precise and the outputs remain explainable and controlled.<\/p>\n
\u201cAI agents can work around the clock, but in financial services they still need to be accountable,\u201d he said. \u201cThey have to report, show their sources and operate within the bank\u2019s rules.\u201d<\/p>\n
From Experimentation To Embedded AI<\/p>\n
Piper\u2019s presentation framed agentic AI as a shift from experimentation to embedded capability.<\/p>\n
The first phase of AI adoption focused on prompts, chat, summarisation and productivity. The next phase is about connecting models to trusted data, integrating them into workflow, and allowing agents to perform specific tasks with appropriate oversight.<\/p>\n
For wealth managers, this is where AI becomes more operationally meaningful. It can help RMs prepare for client meetings, support investment conversations, draft proposals, complete KYC work, process documents and handle repetitive internal tasks.<\/p>\n
But the test is not whether the technology is impressive. The test is whether it is adopted, whether it improves the work, and whether it can operate safely in a regulated environment.<\/p>\n
Piper\u2019s message was that agentic AI should be judged by those standards. It is not a replacement for the human adviser, but a way to give advisers, compliance teams and operations staff more capacity, more context and better execution support.<\/p>\n","protected":false},"excerpt":{"rendered":"At the Hubbis Malaysia Wealth Management Forum 2026, Damien Piper, Executive Director – Growth at Unique AI, examined…\n","protected":false},"author":2,"featured_media":67333,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[6],"tags":[405,7537,3382,3386,3395,3391,3381,3385,3394,3390,3383,3387,3396,3392,3380,3384,3393,3389,3388,3202],"class_list":["post-67332","post","type-post","status-publish","format-standard","has-post-thumbnail","category-agentic-ai","tag-ai-agents","tag-artificial-intelligence-agents","tag-asia-private-banking","tag-asia-private-banking-news","tag-asia-private-banking-online-training","tag-asia-private-banking-training","tag-asia-wealth-management","tag-asia-wealth-management-news","tag-asia-wealth-management-online-training","tag-asia-wealth-management-training","tag-asian-private-banking","tag-asian-private-banking-news","tag-asian-private-banking-online-training","tag-asian-private-banking-training","tag-asian-wealth-management","tag-asian-wealth-management-news","tag-asian-wealth-management-online-training","tag-asian-wealth-management-training","tag-e-learning","tag-training"],"_links":{"self":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/67332","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/comments?post=67332"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/67332\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media\/67333"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media?parent=67332"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/categories?post=67332"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/tags?post=67332"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}