Cognition.Finance Insights

Artificial Intelligence in Investment Management – Use Cases

Artificial Intelligence in Investment Management – Use Cases

Artificial Intelligence in Investment ManagementArtificial Intelligence in Investment Management: AI and ML Use Cases

Artificial Intelligence in Investment management sector is at a nascent and experimental stage but is evolving fast with use cases spanning the entire asset management value chain.  By 2025 artificial intelligence in investment management firms will be pervasive and will play an integral role in process efficiency and decision efficacy.

Unless an asset manager is living under a rock, the drumbeats of cognitive technologies are everywhere, and headlines about chatbots, machine learning, deep learning, neural networks, computer vision dominate the investment management trade journals.

In this thought paper, Cognition.Finance, a leading publication covering cognitive technologies in the broader financial services space, examines how artificial intelligence will impact asset management, what is the current state of AI and cognitive techniques in the investment management space, and what are the use cases across the asset management value chain for leveraging AI and related concepts.

The asset management sector is immense with trillions of dollars of the asset under management (AUM) and set to increase tremendously with the rise in global wealth, particularly in the emerging markets.

According to PWC, the assets under professional management of investment management firms will double by the year 2025.

By 2025, AuM will have almost doubled – rising by 6.2% a year, from US$84.9 trillion in 2016 to US$145.4 trillion in 2025, with the fastest growth seen in the developing markets of Latin America and the Asia Pacific.

While active management will continue to grow and play an important role, reaching $87.6 trillion by 2025 (60% of global AuM), PwC predicts growth in passive management to reach $36.6 trillion by 2025 (25% of global AuM).

Alternative asset classes – in particular, real assets, private equity, and private debt – will more than double in size, reaching $21.1 trillion by 2025, accounting for 15% of global AuM.

There is a ‘great divide’ between asset and wealth managers who have acted to ensure they are fit for growth and those who have not.

The industry’s involvement in niche areas such as trade finance, peer-to-peer lending, and infrastructure will dramatically increase.

A Brief look into the state of Investment Management sector and the Asset Managers:


  • The growth in global wealth, particularly discretionary wealth, driving the increase in AUM across the world.
  • Product innovations are covering every imaginable niche of the investment world.
  • Digital technologies are improving the front-end client and advisor experience


  • The tectonic shift of a significant AUM into low-cost passive investment strategies and vehicles
  • Regulatory burdens increasing in all major global jurisdictions
  • Pressure on fees and hence compression of margins
  • A competitive landscape that is intense with upstart startups, technology giants, and traditional players competing for a slice of the investor pie.

As the industry participants are well aware, the asset management industry is vast, varied, and versatile. The asset managers encompass traditional mutual funds, exchange-traded funds (ETFs), closed-end funds, Unit Investment Trusts (UITs), separately managed accounts (SMAs), hedge funds, private equity, and venture capital.  Furthermore, there are asset custodians, fund administrators, prime brokers, and other service providers and you can see the sheer breadth of the industry. Artificial intelligence will impact each of these subsectors of asset management differently regarding scope and magnitude.

The advent and growth of artificial intelligence and related cognitive technologies is both a blessing and a bane, a fantastic opportunity, as well as a severe threat.  Understanding and acting on how artificial intelligence impacts investment managers and asset servicers is at the top of the executive agenda.

Now let’s dig into how artificial intelligence in investment management and asset management works and what areas are ripe for disruption.

Why is Artificial Intelligence important for investment management and asset servicing firms?

AI and cognitive technologies are useful in the end-to-end investment management value chain:

  • AI can help asset managers improve the experience of clients, advisors, and intermediaries with intelligent segmentation, personalization, and onboarding.
  • Machine learning, deep learning, and related techniques can help in aiding the investment research and portfolio management functions including research, manager due diligence, aggregation and reporting of performance, and manager attribution.
  • Machine learning and Robotic Process Automation (RPA) can bring efficiencies to many middle offices and back-office functions such as intelligent rebalancing, collateral management, trade reconciliation, and automated compliance.

Essential Use Cases for deploying AI in the Asset Management Universe (including Alternative Investments such as Hedge Funds and Private Equity firms.)

  1. AML/KYC

  2. Client Segmentation

  3. Account Opening and Onboarding

  4. Investment Research

  5. Manager Due Diligence

  6. Performance Measurement and Attribution

  7. Collateral Management

  8. Capital Call Processing

  9. Adherence to IMA (Investment Management Agreement) and IPS (Investment Policy Statement) term

Anti-Money Laundering (AML) and Know Your Customer (KYC):

While brokerages and banks face most of the burden relating to AML/KYC, asset managers that have direct to consumer operations through marketplaces or their channels need to comply with the tenets of AML/KYC.

The traditional AML/KYC processes of asset managers are somewhat antiquated, rigid, rule-bound, and siloed making it difficult to attain an accurate picture. Sometimes, inability to bring together all the information leads to false positives wasting time and annoying the customers.

Artificial intelligence can help make the AML/KYC process smoother and more effective and here’s how.

AI-enabled link analysis, clustering, and pattern recognition by analyzing large data encompassing complex webs of relationships and culling out evidentiary information relating to customers’ financial and non-financial activity is a crucial enabler in AML/KYC capability.  These links and patterns allow human beings to step in and conduct additional due diligence where necessary.

Instead of reinventing the wheel, asset management firms can use third parties to help them in AI-enabled KYC and AML. For example, recently FICO (Fair Isaac and Co) has acquired Tonbeller developed anti-financial crime solutions which use unsupervised Bayesian learning algorithms and other techniques to discern deeper aspects of customer behavior which in turn drives further investigations into the suspicious activity.

AI and ML techniques go beyond data crunching in the background and enable the engagement with clients to foster AML/KYC without the pain. For example, machine learning based compilation and aggregation of data can pre-fill many of the form fields, and for the rest, a natural language enabled conversational agent can help walk the customers through the process.

With supervised learning, the machine learning algorithms can “understand” the regulations and help not only in adherence but filing the SARS (Suspicious Activity Reports) to the regulatory authorities.

In addition to the Tonbeller mentioned above, firms like Feedzai and Nice Actimize provide AI-enabled AML/KYC solutions.

Client Segmentation and Personalization:

Artificial Intelligence in Investment Management

Artificial intelligence and machine learning can be an effective means to segment customers and personalize the journeys and interactions to drive additional AUM.

Instead of traditional demographic and psychographic profiles, AI and ML can help companies a vast set of data which includes demographics, values and lifestyles, interactions and transactions to micro-segment the customers and personalize their journey.

Franklin-Templeton, a large U.S. based asset management firm claims to have added new AUM worth $600 million by effective client segmentation and personalization. Fractal Analytics, the AI-startup, partnered with Franklin-Templeton to implement the technology.

“Merging AI, external data and a customer’s own internal data sources to create detailed segments is the only way for marketers to get an edge when it comes to driving superior customer experience,” said Shashi Upadhyay, CEO at Lattice Engines.

Another startup, Lattice Engines, an AI-enabled customer segmentation engine has launched a data cloud for its customers. Leveraging a next-generation segmentation engine, Lattice’s can create unlimited segments using a combination of the Lattice Data Cloud, which provides more than 16,000 attributes on 200 million companies globally, and their internal data, including CRM and ERP data, transaction data, product usage data, support and service data, etc. The segments are then made available in company’s execution systems (e.g., marketing automation, CRM, or ad targeting platforms) to enable teams to run hyper-targeted campaigns across the entire buyer’s journey – from awareness to engagement to conversion to expansion.

“Merging AI, external data and a customer’s own internal data sources to create detailed segments is the only way for marketers to get an edge when it comes to driving superior customer experience,” said Shashi Upadhyay, CEO at Lattice Engines.

Account Opening and Onboarding:

Machine learning techniques like chatbots, NLP and Text Analytics can help in making account opening and onboarding smooth.

An asset manager can use machine learning to aggregate client data from numerous sources. And this information could be pre-filled into the account opening forms leaving the customer to validate the data and enter any missing fields.  For the incomplete fields, a natural language conversational agent can help guide the customers.

With multi-custodial account aggregation, machine learning engines can locate the assets that the customer wants to transfer and initiate automatic ACAT and non-ACAT transfers to marshal the assets for funding the account.

Companies like Rage Frameworks, Quovo, and others use machine learning to extract and normalize data from unstructured data sources and offer account aggregation services for held-away assets including alternatives.

Investment Research: AI-enabled Fixed Income and Equity Research:

Artificial intelligence is slowly but surely stepping into the terrain of investment research, for long a sacred cow among the buy-side firms. Given the information explosion, mainly external, non-financial data, it has become difficult for human equity research analysts to sift through the noise for appropriate signals.

Instead, a combination of big data, machine learning, deep learning, and predictive analytics are changing the way investment research works.

Fixed Income Research:

For fixed-income managers, monitoring the credit quality of issuers, particularly ongoing and real-time intelligence about various events and the impact on issuers is a critical success factor.  Whether it is micro or macro events, each has a varying level of relevance and implications. For example, company level factors such as the debt-equity ratio or executive shuffle will undoubtedly impact the firm’s outlook.  And industry level factors such as OPEC production limits will influence producers, refiners and downstream players to a differing degree. And geopolitical events, such as government trades and tariffs will impact a range of players in the industry value chain.

Compiling, collating, synthesizing, and analyzing all these information borders on the impossible for a human analyst.  However, machine learning algorithms can aggregate, extract, normalize, and summarize all this information, including unstructured data.   In addition to offering the nuggets of useful information through a dashboard, AI can analyze and measure the impact and predict the outcomes.

Within the Fixed Income sector, Muni bond fund managers need to follow issuers at the county, city, town level and much of the data is unstructured (or available through EMMA.) Also, local news articles showcase what is happening at a local level, and this information combined with census data, housing, and real estate data, and labor/income data can add tremendous depth and granularity of analyzing the credit quality of the muni issuers.

Equity Research:

artificial intelligence in investment management

Most of the use cases in Fixed Income apply to the Equity side of the house, but perhaps the data and sources are exponentially more significant, and the implications order of magnitude higher.

Today, equity research analysts can use a lot of non-traditional data to glean insights. A lot of this is noise and correlation alone is not causation.  So, asset management firms and heads of equity research must determine what the factors that drive stock performance and include those parameters into the decision-making matrix are.

For example, the following are some of the non-traditional sources, much of it unstructured that equity research analysts are leveraging, thanks to big data, machine learning, deep learning, and predictive analytics type concepts and technologies.

News Articles and Blog Posts: ML techniques such as NLP (Natural Language Processing) and NLU (Natural Language Understanding or also known as “Machine Comprehension”) are helpful in compiling, extracting, and analyzing the news articles for critical pieces of information as well as overall sentiment analysis.   These news articles may be about a company, or industry or broader economy and the ML algorithms can intelligently evaluate and assign the degree of impact at various levels (economy, industry/sector, company, stock).

Tweets and Message Boards: The activity and sentiment from stock message boards and stock tweets may be a useful indicator to measure investor psyche around the security.

Obituaries: Morbid as it may sound, obituaries of key executives of companies could provide that information edge in some cases.

Job Activity and Movements: Analyzing Linkedin data for job changes and recruitment activity could provide indicators on which employer is hot or not.

Satellite Imagery: If a parking lot is full, does it bode good tidings? Or if a factory lot is full of trucks, is that a signal of positive momentum? Today, analyzing the satellite imagery – though the technology is at a nascent stage – is a potential game changer.

IOT Data Streams:  Internet of Things (IoT) data streams and location beacons can provide tremendous information for analysts.

Automated Investing: In addition to supporting the human investment research analysts with inputs and nuggets, there are machine learning and deep learning algorithms that analyze the markets, economies, industries, and individual securities and automate the construction of the portfolio and the buy-sell process.

Random Forest Capital uses machine learning algorithms to automate the investment process, and Franklin Templeton acquired the company. “Random Capital approaches investment management from the perspective of data science, in which machine learning within fully non-parametric statistical models are applied to the problem of expected gains in financial investments. Rather than having humans look at each event within the marketplace, machine learning employs statistical algorithms over thousands of variables and millions of observations that are capable of detecting persistent effects across all aspects of data.”

Manager Due Diligence:

Wealth managers, fund of funds managers, institutional investors, and third-party investments consultants evaluate money managers and conduct due diligence. As a part of this process, research teams rely on third-party databases for profile information and supplement and complement information with typing in a lot of unstructured data to compile a full picture of the money managers.

For example, a manager may submit his/her bio as a PowerPoint or a PDF document. Similarly, managers may transmit performance numbers an Excel attachment. Today, due diligence teams enter this data manually – which is expensive, slow, and error-prone.

Instead, today machine learning can help extract this unstructured data and combine it with regulatory filings and third-party manager profile databases to gain a holistic perspective of managers and conduct due diligence that is more comprehensive and effective.

Performance Measurement and Attribution:

For wealth managers, fund of funds, family offices, and third-party investment consulting firms aggregating performance numbers from various money managers is a painful chore. Today, beyond the traditional account aggregation which spans direct custodial feeds, machine learning algorithms can extract the performance numbers and normalize across the rest of the assets for measuring overall portfolio performance and attribution – at a manager level, portfolio level, account level and indeed sleeve level.

Firms like Quovo, AltX, Rage Frameworks, and ByAllAccounts and others such firms are helping asset managers and wealth managers extract unstructured performance and portfolio holdings information.

Collateral Management:

Collateral management is a complex and manual process as terms are embedded into many unstructured documents. AI and cognitive technologies can unlock this data and help firms fulfill the contractual obligations by multi-way matching.

Capital Call Processing:

In the Private Equity sector, Capital Calls are a critical process and often it is manual as to how the whole capital call operations flow.  Machine learning algorithms can extract the meaning and data and handle the claims for straight thru processing.

IMA and IPS Adherence:

Each IMA (Investment Management Agreement) with LPs (Limited Partners) and each IPS (Investment Policy Statements) with clients and institutional investors are littered with specific terms, thresholds, covenants, and restrictions.

Traditional pre-trade compliance cannot handle the unstructured data and details embedded in these documents. Again, machine learning and text analytics come to the rescue and remove a significant hurdle in complying with the terms of these contracts.

As you can see artificial intelligence in asset management is a tremendous value-add and has pervasive applications across the investment management value chain.


If you wish to obtain a PDF version of Artificial Intelligence in Investment Management, please contact us.

Our experts are available to conduct a custom briefing on how artificial intelligence impacts your firm and how to leverage the AI technologies to leapfrog the competition.

We can help in creating an AI-driven Transformation Roadmap for an asset management firm spanning all critical areas of the value chain.

We welcome AI software vendors and service providers to share client success stories or case studies about AI and cognitive technologies in the asset management space.

Artificial Intelligence in Investment Management: An AI guide to Asset Managers is current as of October 2018. While we strive to provide accurate and up to date information, given the dynamic nature of the AI field, some of the information may be outdated.


This Is A Custom Widget

This Sliding Bar can be switched on or off in theme options, and can take any widget you throw at it or even fill it with your custom HTML Code. Its perfect for grabbing the attention of your viewers. Choose between 1, 2, 3 or 4 columns, set the background color, widget divider color, activate transparency, a top border or fully disable it on desktop and mobile.

This Is A Custom Widget

This Sliding Bar can be switched on or off in theme options, and can take any widget you throw at it or even fill it with your custom HTML Code. Its perfect for grabbing the attention of your viewers. Choose between 1, 2, 3 or 4 columns, set the background color, widget divider color, activate transparency, a top border or fully disable it on desktop and mobile.