AI in asset management: a buyer's guide to agentic AI for investment managers

Where does AI in asset management pay off first, and why?

Investment management runs on documents as much as on prices: annual reports and filings in the issuer's own language, broker research and call transcripts, indentures and loan agreements, investment management agreements and side letters, ISDA schedules, custodian and prime broker files, administrator NAV packs, capital call notices, board books, consultant questionnaires, factsheets and client reports. Much of the cost sits in skilled people reading those documents, comparing them with a record or a rule, and keying what they find. That is the work AI does well, because the answer is on a page and can be checked.

Margins leave little room for that cost. McKinsey reports that global assets under management reached a record $147 trillion by June 2025, yet pretax operating margins were 33% in 2024, up only about one point on 2023 even as revenue pools grew by double digits, and industry cost reached $167 billion. That is an estimated 11 basis points of cost on every dollar managed each year (2024 cost over mid-2025 assets). PwC's 2025 global report finds profit per unit of assets down about 19% since 2018, with 89% of managers reporting pressure on profitability. Fee rates rarely rise, so time recovered in research, operations and reporting is one of the few levers left.

Agentic AI pays off first where three things hold: the volume is high, the inputs are documents or messages, and a qualified person already reviews the result. In those places agents read, check and draft, and the analyst, controller, compliance officer or client manager decides faster with the evidence in front of them.

Which processes should asset managers start with?

Start with work that is document-heavy, measured in basis points, breaks, days or deadlines, and already reviewed, so the human check stays in place while the reading and keying go away. Leave the investment decision, the trade, the NAV sign-off, the vote and every regulatory filing with the people who own them today.

• Research: questions answered across filings, broker research, transcripts and internal notes with the page cited, and local-language annual reports and financial statements turned into English beside the original, so more names get read on the same budget.
• Credit and fixed income: issuer financials, indentures and covenants read and compared, held issuers watched for filings, ratings and court news, and credit reviews drafted for committee.
• Portfolio management: pre-trade questions such as which funds may buy a security answered with the clause quoted, investment policy statements checked, and managers and sub-advisers reviewed against their mandates.
• Trading and best execution: the quarterly best-execution pack built with every exception explained, transaction costs read across equities, bonds and FX, and broker reviews and trade errors documented.
• Operations and reconciliation: books matched to custodian and prime broker files with every break explained, failing trades and settlement instructions chased, and corporate action notices turned into one checked record before the election deadline.
• Fund accounting and NAV: each administrator NAV checked against the manager's own books, fees and expenses recalculated from fund terms, and the fund administrator's service levels tracked.
• Private markets: capital calls and distribution notices read, wire details verified and booked, portfolio company packs turned into KPIs, data rooms read for red flags, and side letter obligations tracked.
• Risk, compliance and surveillance: liquidity and derivatives programs run with board reports drafted, guideline alerts triaged to real breaches, communications alerts closed with reasons, and every ad and factsheet checked against the SEC Marketing Rule.
• ESG and stewardship: issuer sustainability reports read in their own language, exclusion screens tested with evidence, and proxy votes drafted against the house policy with the rationale written down.
• Distribution, client reporting and fund governance: consultant databases and RFP answers kept current from approved data, institutional reports and factsheets drafted and checked, and 15(c) board materials and prospectus updates assembled with a cited change log.
• Wealth and investor services: advisor meeting packs and follow-ups drafted, new-account and subscription documents checked for gaps before they are submitted, and investor KYC files kept current.

What should we demand from a vendor of AI in asset management on citations, sign-off and audit?

Demand that every extracted figure, every match and every drafted sentence links to the exact passage it came from: the line of the 10-K or the local-language annual report, the clause of the indenture or the investment management agreement, the row of the custodian file, the paragraph of the side letter. An answer without a source should be treated as unverified, especially where an auditor, a fund board, a client or an SEC examiner will later ask how a conclusion was reached.

Demand that accuracy is measured on your own documents before go-live and monitored after, field by field and decision by decision, against what your experienced staff concluded. Ask for error rates and for the cases the system sends to a person, not a single headline number.

Demand maker-checker sign-off wherever responsibility sits with a person: the analyst who publishes a note, the portfolio manager who trades, the controller who signs off a NAV, the compliance officer who closes an alert or approves an ad, the valuation committee that sets a mark, the treasurer who files a report. The person who signs should see the evidence, the rule applied and the draft, and the system should record who prepared it, who approved it and when.

Demand an audit trail you can export. Every agent action and every human decision should be logged in a record that cannot be altered, and you should be able to hand that record to an auditor, a board or an examiner in a standard format without asking the vendor first.

What about security, entitlements, data residency and the systems we already run?

Investment data includes material nonpublic information, client and investor personal data, holdings and trading intentions, and commercial terms with clients, brokers and service providers. Ask where the data lives, who can see it and what happens to it afterwards. Botminds runs as multi-tenant SaaS, in a dedicated cloud environment, or fully on-premises, with the same controls in each, so data can stay in the region your obligations require. Your inputs and outputs are not used to train models.

Entitlements matter more here than in most industries. Research behind an information barrier, deal-team data rooms, private fund side letters and client portfolios should each be visible only to the roles entitled to them, and the AI should never answer a question with a document the person asking cannot open. Look for encryption in transit and at rest (Botminds uses TLS and AES-256), tenant isolation, single sign-on with MFA, role-based access to each dataset and workflow, and an audit log that records every human and agent action. Botminds holds SOC 2 Type II and ISO 27001 and supports GDPR, with audit reports available on request.

Ask how the vendor works with the systems you already run: the order management system, portfolio accounting, the fund administrator's files and portal, the custodian and prime broker feeds, the CRM, research management and the shared mailboxes where notices and requests still arrive. A good answer reads from those systems and writes back through approved routes, so staff are not asked to keep a second copy of the truth. Expect your own security, vendor due diligence and model risk review of the AI vendor, and plan for it from the start.

How do we run a pilot of agentic AI in asset management?

Pick one process, one team and a fixed window, such as 6 to 10 weeks. Agree before you start what you will measure, and take a baseline from the current process so the comparison is real.

• Use your own recent documents, including the hard ones: scanned capital call notices, local-language annual reports, long indentures with amendments, month-end break files and side letters with most-favored-nation terms.
• Measure cycle time end to end, such as hours to read an earnings season, breaks still open at NAV cut-off, or business days after quarter end to send client reports.
• Measure touch time for the analyst, controller or compliance officer, and count the corrections they make.
• Count what was found or protected: NAV differences caught before publication, fee errors caught before billing, guideline breaches confirmed inside the cure period, wire changes stopped before payment.
• Check that every output a person signs can be traced to its source page, that entitlements held, and that the audit log shows who did what.
• Decide in advance what result means expand, adjust or stop.

Should we build, buy point tools, or use a platform?

Building in-house gives control but means owning document reading, citations, review screens, entitlements, audit and model changes for every process, usually with a technology team already stretched by data and trading systems. Point tools solve one step well, but a manager with research, operations, fund accounting, compliance and reporting work ends up with many vendors, many due diligence reviews and evidence spread across many places.

Budgets are not growing to cover the gap. Substantive Research found that research budgets at the largest asset managers fell 0.2% in the first half of 2025 while assets grew, and FactSet's 2025 survey found only 12% of asset managers and owners believe their reporting fully meets client needs, with 60% citing a lack of automation and 70% data availability and quality.

A platform with a library of ready solutions sits between the two: each solution fits a specific investment process, and all share one security model, one audit log and one way of citing and signing. Botminds is the Agentic Solution Engine: it builds governed, agent-driven solutions from a specification and delivers them live. Its investment management library covers 15 areas, from research, credit and portfolio management to trading, operations, fund accounting, private markets, risk, compliance, ESG and stewardship, distribution, client reporting, fund governance, wealth management and investor services.

Whatever you choose, check that you can export your documents, extracted data and decisions, so changing course never means losing your records.

What changes for analysts, portfolio managers, operations and compliance teams?

The job moves from searching and typing to judging. Analysts start from a cited read of the filing or the call and spend their time on the thesis; controllers review the NAV differences that matter instead of ticking every line; compliance officers work the alerts that are real; client teams check a drafted report instead of building it. Corrections should become rules for the next case, so the same fix is not made twice.

Decisions stay with people: the portfolio manager who trades, the investment committee that approves a deal, the valuation committee that sets a mark, the chief compliance officer, the fund board and the treasurer who signs a filing. Plan training on reading the evidence trail, and update procedures and compliance manuals to say where AI drafts and where people decide.

What does it cost to wait?

Every month without change is another month of paying skilled people to search and re-key, and of breaks, fee errors and late reports found after the fact. On a $10 billion fund, one basis point is $1 million a year, so small leaks add up. Since the US moved to T+1 in May 2024, DTCC reports industry affirmation by 9 p.m. on trade date rose to 94.80% in April 2026 from 88.39% before T+1, while the DTC fail rate rose to 3.01% from 2.83%. Business email compromise, the fraud behind many changed wire instructions, cost victims $2.77 billion in 2024 in complaints to the FBI.

Rules and standards with fixed dates are also in force or arriving, and each one asks for evidence that is hard to assemble by hand:

• Marketing Rule: in force since November 2022, it requires advisers to substantiate what their ads say, and the SEC's Division of Examinations issued another risk alert on testimonials, endorsements and third-party ratings on December 16, 2025.
• Form PF: on August 31, 2026 the SEC and CFTC moved the compliance date for the 2024 Form PF amendments to July 1, 2027, while they consider an April 2026 proposal to scale them back.
• Names Rule: funds whose names suggest a focus must invest at least 80% of assets accordingly under the 2023 amendments, with compliance dates moved in March 2025 to June 11, 2026 for groups with $1 billion or more and December 11, 2026 for smaller groups; in 2026 the SEC extended the related Form N-PORT reporting dates to November 17, 2027 and May 18, 2028.
• Settlement: the EU, the UK and Switzerland move to T+1 on October 11, 2027, compressing affirmation, instruction and funding work for every manager that trades there.
• Private fund reporting: ILPA's updated Reporting Template applies from the first quarter of 2026 for funds still in their investment period and for funds launched on or after January 1, 2026, and limited partners increasingly expect it.

Solutions to look at first

• Investment Research AssistantAnswers investment teams' questions across filings, broker research and notes, all cited
• Financial Statement TranslationTurns local-language filings and financial statements into English for global analysts, cited
• AI Credit AnalysisAnswers credit analysts' questions across filings, indentures, calls and notes, all cited
• Pre-Trade Compliance ChecksTells portfolio managers which funds may buy a security and how much, quoting the clause
• Best Execution ReviewBuilds the quarterly best-execution committee pack for heads of trading, exceptions explained
• Position and Cash ReconciliationReconciles books to custodian and prime broker files and explains every break for middle office
• NAV OversightChecks each administrator NAV against the manager's own books and sizes errors for controllers
• Capital Call ProcessingReads GP capital calls and distribution notices, checks wires and books them for LP operations
• Liquidity Risk Management ProgramRuns a fund's Rule 22e-4 program: liquidity buckets, the 15% limit, minimums, board reports
• SEC Marketing Rule ComplianceChecks every ad, factsheet and pitch book against the SEC Marketing Rule and FINRA 2210
• Proxy Voting ResearchApplies the house voting policy to each proxy, drafts vote rationales and the Form N-PX
• Consultant Database ManagementKeeps every consultant database profile current from approved fund data, for RFP teams
• Institutional Client ReportingAgents draft, check and release monthly and quarterly reports for every institutional mandate
• 15(c) Board ReportingAgents assemble quarterly board books and answer the annual 15(c) request list, cited
• Client Meeting PreparationAgents draft the prep pack, notes and follow-ups for each advisor meeting; advisors approve

Every investment management solution →

A checklist for any vendor

1. Does every extracted figure, match and drafted sentence link to the exact page it came from?A good answer is a live demo on your document: click a covenant, a NAV difference or a fee calculation and land on the line of the filing, agreement or custodian file it came from.
2. How will you measure accuracy on our documents before go-live, and keep measuring after?Look for field-level and decision-level results on your own sample, compared with answers your staff confirmed, and ongoing monitoring that catches drift.
3. Where does a person sign, and what does the system record when they do?Research notes, NAV sign-off, guideline breaches, ad approvals, valuation marks and filings should each need a named maker and checker, with the evidence shown and both signatures logged.
4. What happens to the cases the system is unsure about?They should go to a person with the reason stated and the evidence attached, never be passed through silently or dropped.
5. How do entitlements work, and can the AI answer from a document the user cannot open?The answer should be no: access is set per dataset and role, information barriers hold, and deal, client and side letter material stays with the people entitled to it.
6. How does it work with our order management system, portfolio accounting, fund administrator, custodian and CRM?It should read from those systems and the shared mailboxes, and write back through approved routes, with no second copy of the truth.
7. Can we export the audit trail, our documents, extracted data and decisions at any time?Yes, in standard formats, so an auditor, board or examiner can see who did what, and you keep your records whatever you decide later.
8. Where can it be deployed, and where does our data live?Options should include multi-tenant SaaS, a dedicated cloud environment and on-premises, with the same controls in each and data kept in the region you need.
9. Are our inputs or outputs used to train your models or anyone else's?The answer should be a plain no, written into the contract.
10. Which security certifications and audit reports can you share for our vendor due diligence?Expect a current SOC 2 Type II report and ISO 27001 certificate, plus encryption, SSO with MFA, role-based access and an unalterable audit log.

Questions