Client reporting automation you work with in plain words: agents tie every figure to the book of record, explain each break with its source, and a different person releases
Quarterly packs out by business day 3, not day 10, each figure tied to the book of record.
See one case, screen by screen ↓Institutional Client Reporting is client reporting automation for asset managers: seven AI agents draft, check and release the monthly and quarterly reports every institutional client receives on its segregated mandate. They build each pack from the book of record, tie every figure to its source, check it against the client's own investment management agreement, compare it with last quarter and trace every break to its cause. People work with the agents in plain words — ask why a figure doesn't tie and the answer cites the journal and the clause. The analyst decides each break, and a different person, the team lead, releases.
AI in asset management: a buyer's guide to agentic AI for investment managers →
Updated
Client reporting in asset management means every institutional client gets a report on its own segregated mandate each month and a full quarterly investment report each quarter: returns gross and net, attribution, holdings, transactions, commentary and a compliance certificate, each on the terms its investment management agreement sets. Without client reporting automation the work is serial. Performance signs off, then analysts pull figures into templates — in client reporting software, Excel or Word — one mandate at a time, check them by eye against printouts and cross-foot totals by hand. A number that breaks goes back to performance by email. Client terms live in a spreadsheet copied from agreements and side letters, some in German or Japanese, and changes that arrive by email get missed.
The risk sits in what nobody had time to check. Attribution that no longer reconciles to the gross excess return because a template changed, a pack valued at the New York close when the agreement names the 16:00 London fixing, Japanese commentary that says selection helped when attribution says it hurt, a certificate that leaves out a passive breach. A wrong figure means a reissued pack, a call from the client's consultant and a working paper the SEC can ask for under Rule 204-2. For EU accounts, MiFID II requires a periodic statement at least every quarter (Delegated Regulation 2017/565, Article 60) and a notice by the end of the business day when a portfolio falls 10% from the start of the reporting period (Article 62). A missed term is a breach of the agreement.
Estimated split, in business days, for a quarter-end of 612 mandates like the demo firm's: by hand (last pack about business day 10) and with the solution (the business day 3 target). Our estimate, not a measured result.
Seven agents draft, tie out, check and deliver every pack; the analyst decides, performance confirms, the team lead releases.
Quillfield Capital's institutional client reporting team owes 612 quarterly packs to 214 institutional clients in 9 currencies. It is Monday, October 5, 2026, business day 3 of the Q3 cycle. The analyst works with the agents as a conversation: they report what they found, cite it, and propose; he asks back, decides, and hands the pack to someone else to release. Here it is, screen by screen, in the working solution.
Leo's Today screen is a conversation. The Drafting agent says all 612 quarterly packs for 214 clients are drafted from the book of record and 1,184,316 figures were tied. The Tie-out agent says 17 breaks need him and 12 share one cause. The Requirements agent flags Versorgungswerk Rheinmark's FX fixing, and the Anomaly agent has escalated Ashbourne Rail's hedge ratio to the team lead. 396 of 612 packs are already with clients.
Tie-out agent: “17 breaks need you. 12 share one cause: the September 12 attribution template change. One decision can fix all twelve.”
The tie-out grid shows every mandate against six report sections — performance, holdings, attribution, commentary, compliance certificate and client terms — with a tick or a break in every cell. Leo selects Calder Basin Pension, Global Equity Core, $2.92B, due Wednesday, October 14 (business day 10). Attribution and client terms both show a break. Beside the grid: 1,945 of 1,946 figures tied, and total attribution of 0.42%, from the attribution engine, marked as the one break.
Total attribution is 0.42% against a gross excess return of 0.47% — a 0.05% residual where the tolerance on totals is 0.00%. The Requirements agent cites Schedule C §3.2 of Calder Basin's agreement: attribution gross of fees, reconciled to the gross excess return. The Anomaly agent notes the residual was 0.00% in each of the last eight quarters. The agents offer three options; restating gross of fees is recommended.
Option B — show 0.05% as a “fees and other” line — is marked “Fails Schedule C §3.2”.
Clicking 0.42% opens the lineage the Tie-out agent traced: the report figure, attribution run ATT-Q3-4471-r1 on a net basis since September 12, change PERF-2291 that set the return basis from gross to net, the performance system's 3.86% gross against 3.81% net, and the Q3 fee accrual true-up of $1,459,203 in journal JE-2026-0930-88412. Proof: $1,459,203 ÷ $2,918,406,233 = 0.0500%.
“The five basis points are the Q3 fee accrual true-up.” Checked by the Anomaly agent against eight prior quarters.
PERF-2291 switched 12 accounts' attribution to the net-of-fee series. The Tie-out agent proves each residual against that account's own fee accrual — from 0.03% at Wattle Coast Superannuation Fund to 0.06% at Sable Ridge, Meridian, Granite Peak and Oakhaven — so one restatement gross of fees closes every one, as the clients' agreements require.
Leo approves option A for all 12 accounts with the reason code “Methodology per client agreement”. At 10:52 Tomás Ferreira, the performance analyst, confirms re-run ATT-Q3-4471-r2: allocation 0.11%, selection 0.31%, currency 0.05%, total 0.47%, residual 0.00%. The Narrative check agent rewrites the one sentence on page 4 that depended on the figure, and Elena Vasquez, the portfolio manager, accepts it.
Narrative check agent: “Rewrote one sentence on p. 4 … Elena Vasquez's approved commentary is otherwise unchanged; I sent her the redline as its owner.” The redline: 27 → 31 basis points.
Leo sends the pack to the team lead; he cannot release a pack he prepared. Sara Nwosu, who prepared and edited nothing in it, sees what was checked: 1,946 of 1,946 figures tied to the book of record, 14 of 14 client-term checks passed under Schedule C, anomalies explained, the restatement approved and confirmed, the commentary accepted. She signs and releases to the client portal and to Northgate Consulting by SFTP, nine days before its October 14 deadline.
“You did not prepare or edit this pack (maker-checker).”
With the other eleven restated packs added, 152 are ready. The Release agent ranks them and says why for each: 86 are low risk — every figure tied first time, no edits, same template as Q2 — and the other 66 have restated figures, edited commentary, a large mandate or a translated version, so they need a closer read. Sara selects the 24 lowest risk and releases them with a reason and her signature. Packs she edited are excluded automatically.
Release agent: “I deliver only what a team lead releases.”
Versorgungswerk Rheinmark's agreement is in German. The Requirements agent reads it into nine checks, with each clause beside the English: 7 of 9 pass. Anlage 3, Ziffer 4.1 requires the 16:00 London closing rate; the draft used the New York close, a €184,221 (0.02%) difference in market value. On “Monatsultimo” it does not guess — it asks which date to use, because October 31, 2026 is a Saturday.
“‘Monatsultimo’ can mean the last calendar day or the last bank business day. It matters in October: October 31, 2026 is a Saturday. Which should the monthly report use?”
The Calendar agent builds every due date from 214 clients' agreements, side letters and emails, each citing its clause. The cycle view shows packs due and released by business day and by office. Q4 2025, Q1 and Q2 2026's last public-markets packs went out on business days 10, 11 and 10; Q3's target is business day 3. Wrong-figure reissues: 7 in Q2 2026, 0 in Q3 so far.
Private-markets packs follow the fund administrator's NAVs and are due by business day 12 under their agreements.
The same Q3 cycle, seen by five of the people who carry it — what quarter-end looked like, and what it looks like now.
Seven agents draft, tie out, check and deliver every pack; the analyst decides, performance confirms, the team lead releases.
Reads agreements, side letters, client emails and CRM preferences into each mandate's reporting schedule — contents, deadline, language, recipients, channel — builds the cycle calendar and watches for changes. On the demo it read 731 agreements and side letters and proposed 4 changes.
Applies each client's reporting terms, read in the original language, as checks the pack must pass: gross or net, attribution basis, FX fixing, sections, disclosures, custody legend and certificate wording. 7,812 client-term checks ran on 612 packs.
Builds each pack from the book of record — holdings, transactions, returns, attribution, risk and ESG figures — and drafts portfolio commentary variants from attribution and the portfolio managers' notes, every sentence tied to its figures.
Checks every number against its source: recomputes, cross-foots every table, compares accounting with the custodian, follows the lineage of each break to its cause and finds every mandate with the same cause. 1,184,316 figures tied; 17 breaks traced to 5 causes.
Compares each pack with last quarter, the strategy composite and the rules in Settings — return outliers, tracking error, holdings count, attribution residuals and the MiFID II 10% depreciation check for EU accounts — and escalates what it cannot explain.
Checks commentary against the numbers — direction, size and period — in every language, and checks required disclosures and the compliance certificate against the investment compliance system. 23,940 sentences in 4 languages and 489 certificates checked.
Ranks ready packs by risk, assembles the PDF, data file, portal view and language versions, checks every approval and signature, delivers after release to the portal, SFTP or secure email, and files receipts and evidence.
Maker, checker, release. The agents propose; a named person decides.
Ask in plain words, or tell the agents what to do. They answer with sources and act only with your approval.
Why doesn't Calder Basin's attribution tie to its excess return?
Total attribution is 0.42% against a gross excess return of 0.47%. The attribution run used the net-of-fee series after change PERF-2291 on September 12. The 5 basis points are the Q3 fee accrual true-up of $1,459,203, which is 0.0500% of $2,918,406,233. The agreement requires attribution gross of fees.
Which other packs have the same break?
Eleven more accounts share the cause: PERF-2291 switched their attribution basis to net. Each residual equals its own fee accrual, from 0.03% (Wattle Coast Superannuation) to 0.06% (Sable Ridge, Meridian, Granite Peak, Oakhaven). One restatement is proposed for all twelve; it needs your approval and Tomás Ferreira's confirmation.
What does Versorgungswerk Rheinmark's agreement say about FX rates?
The agreement is in German. Anlage 3, Ziffer 4.1: foreign currencies are converted into euros at the 16:00 London closing rate. The draft used the New York close, so market value is off by €184,221 (0.02%).
Get the lowest-risk packs ready for Sara to release
Ranked all packs ready for release by risk: restated figures, edited commentary, size and language raise it. The 24 lowest are all tied first time on Q2's template. I cannot release anything: Sara Nwosu releases, and never a pack she prepared or edited.
13 screens from the working solution, on its sample data. Pick one to see it large.
The tolerances, rules and limits every pack is checked against are settings. Changes go to the Head of Client Reporting for approval and are logged.
| Setting | Default | Choose from |
|---|---|---|
| Tolerance on totals and attribution effects | 0.00% | Displayed returns ±0.01% rounding · market value ±$1 per account or local equivalent · positions exact |
| FX rate source | Client's named fixing | Read from each client's agreement |
| Anomaly rules | Tracking error change over 50% quarter on quarter | Return outside 2 standard deviations of composite dispersion · holdings count change over 25% · attribution residual above 0.00% |
| MiFID II Art. 62 depreciation (EU accounts) | 10% since start of period | Same-day notice |
| Person decides a break within | 4 business hours | Reminder at 75% of the window · escalate to the Head of Client Reporting at 100% |
| House release target | Business day 3 | Warn client service 2 business days before the agreed deadline |
| Confidence below which a finding goes to a person | 0.92 | No agent finding below it changes a pack on its own |
| Retention of reports, working papers and receipts | 7 years | With legal hold for named clients or accounts |
Every figure is labelled: a target the solution is built to, an estimate, a typical published result, or a proven one.
“demo” = seen in the working solution, on its sample data · “target” = the design goal, measured in the live solution · “estimated” = our estimate · “typical” = published figures (an industry survey of asset managers and asset owners, September 2025 — 12% say their reporting meets all client needs, 60% cite a lack of automation, 70% data availability and quality; an industry poll of asset managers — 63% say client reporting is less than half automated; SEC Advisers Act Rule 204-2, books and records; Commission Delegated Regulation (EU) 2017/565, Articles 60 and 62). Quillfield Capital and every person, client, consultant, issuer and account named on this page are fictional — characters and sample data in the working solution.
Institutional Client Reporting is client reporting automation for asset managers: seven AI agents draft, check and release the monthly and quarterly reports every institutional client receives on its segregated mandate. They build each pack from the book of record, tie every figure to its source, check it against the client's own investment management agreement, compare it with last quarter and trace every break to its cause. People work with the agents in plain words — ask why a figure doesn't tie and the answer cites the journal and the clause. The analyst decides each break, and a different person, the team lead, releases.
In this client reporting software for asset management, the work starts when performance signs off: the Drafting agent builds every pack overnight from portfolio accounting, the performance and attribution systems, risk and ESG data. The Tie-out agent recomputes and cross-foots every figure and compares accounting with the custodian; the Requirements, Anomaly and Narrative check agents test client terms, history and commentary. The analyst's day starts with the agents' briefing and only what needs a person: on the demo, 17 breaks out of 1,184,316 figures, each explained with its source and options to choose from.
Yes. The Requirements agent reads each agreement, side letter and client email in its original language into checks — gross or net, attribution basis, FX fixing, sections, disclosures and certificate wording — and shows the clause beside the English. On the demo it ran 7,812 client-term checks on 612 packs. When wording is ambiguous it asks, and records the answer.
The Tie-out agent follows the break through its lineage to the cause and cites it, then finds every mandate with the same cause. On the demo, a 0.05% attribution residual traced to a $1,459,203 fee accrual true-up after a template change, and the same cause hit 12 accounts. The analyst approved one restatement and the performance analyst confirmed one re-run.
Yes. On the demo the agents read agreements in 9 languages and check commentary in English, German, French and Japanese. A German agreement requiring the 16:00 London FX fixing failed a pack valued at the New York close, and Japanese commentary saying stock selection helped was flagged because attribution shows −0.06%.
Every release files its evidence — released reports and data files, tie-out evidence for every figure, decisions and reason codes, approvals and delivery receipts — on a hash-chained audit trail kept 7 years by default. Under SEC Rule 204-2, the books and records rule, advisers keep client communications about performance and the working papers behind them. Your compliance team decides what your firm must keep. Auditors export the whole pack in one step.
Yes. Automated client reporting here never means unattended: the agents never send or release anything. The analyst decides every break with a reason code, restated figures need the performance analyst, certificate exceptions need the compliance reviewer, and the team lead releases — never a pack they prepared or edited. The release rules — no preparer releases their own pack, no agent delivers a pack a person has not released — cannot be switched off.
The Agentic Solution Engine builds and deploys it from your requirements and documents — your client agreements, report templates, a sample of past quarterly packs and the systems they come from — and it goes live once every quality gate has passed. We will walk you through it on your own mandates first.
Factsheets out on business day 3, not day 10, with every figure and disclaimer checked.
1-min demoPortfolio Commentary GenerationEvery sentence tied to its attribution figure, and PMs review in minutes, not hours.
1-min demoGIPS Compliance ReportingComposite errors found before the verifier finds them, with every move proven twice.
1-min demoWe'll run Institutional Client Reporting on a sample of your own past quarterly packs, agreements and source data.