Financial services
AI for Financial Services
Financial services firms use Kylon to run reviews and research where every decision needs evidence. Agents refresh KYC files against Box documents and HubSpot records, recompute loan covenant ratios from borrower statements, propose credit limits from Stripe payment history and triage new insurance claims against the policy schedule. Investment firms keep deals, diligence and the quarterly letter on one record. A named owner approves each decision.
Financial services use cases
Before: Re: dec pages, Moreno, Which one do I pick?, Quote compare v4, Waiting for signature, Moreno follow up. With Kylon: $710 a year saved for one household. Lined up 4 carrier quotes in one table. Workflow tools: Gmail, Outlook, Google Calendar, Google Sheets.
IrisInsurance Agencies: From First Quote to Renewal
Leads, quotes, proposals, binds and renewals for an independent agency
Before: Fwd: rate news??, signals_tracker_old, did anyone see this?, brief_mon_v2.docx, Re: Re: comps update. With Kylon: Desk briefed by 7am daily. Compiled 42 signals overnight.
How a Fintech Team Built a 24/7 Market Intelligence System
Agents watch market signals and compile a daily brief.
Before: Re: still need W-2, FW: bank stmt Oct, client_docs (old), 3rd reminder sent?, Re: status update?. With Kylon: 20 hours back to billable work. Chased 31 missing client docs. Workflow tools: Gmail, Outlook, Google Drive.
MaxAccounting Firms: Turning Client Follow-Up Back Into Billable Capacity
Agents chase client documents and track status across systems.
Before: FW: intro, Series A, DD list v4 (final), Q3_letter_draft3, IC moved again, need audited FS. With Kylon: 3 deals ready for IC Monday. Chased 17 open diligence items. Workflow tools: Outlook, Gmail, Google Drive, Google Calendar.
PipInvestment Firms and Family Offices: Every Deal, Request and Report on One Record
Deals, diligence and LP reports kept on one shared record.
Before: 3 late payments, credit_rpt_2025.pdf, limit: $50k??, memo uses old #s, limit_table_v3. With Kylon: Bad-debt risk cut on 3 accounts. Set 26 limits by the table. Workflow tools: Box, Google Drive, Salesforce, Slack, Stripe.
JoyAnnual Credit Review: Limits and Holds Set by Policy, Account by Account
Agents propose credit limits and holds from payment history.
Before: dinner, $240, OK?, pre-approval?, tender closes 11/20, Re: gift basket, is this over limit?. With Kylon: 2 over-limit gifts caught. Summed totals per counterparty. Workflow tools: Gmail, Google Calendar, Google Drive, Notion, Outlook, Slack.
Gifts and Entertainment Review: A Quarterly Register That Adds Up by Counterparty
Declarations matched to pre-approvals and totalled by counterparty.
Before: OK to sell today?, blackout extended, restricted list v1, is ACME restricted?, preclear_log.xlsx. With Kylon: Trades cleared within policy. Checked 22 requests vs tiers. Workflow tools: Google Calendar, Google Drive, Notion, Outlook, Slack.
AdaInsider Trading Preclearance: Clearances That Respect Tier, Blackout and Restricted List
Trade requests checked against tier, blackout dates and restricted list.
Before: FNOL #31: roof gone, photos_storm.zip, reserve band old, adjuster: unassigned, 31 claims, help. With Kylon: 22 claims moving the same day. Triaged 31 storm notices. Workflow tools: Airtable, Box, Gmail, Google Drive, Salesforce, Slack.
FNOL Claims Triage: New Claims Routed, Reserved and Assigned on Policy
New claims read against policies, with reserve and adjuster proposed.
Before: POA_2024.pdf (old), screening alert open, next review: blank, docs request #3, Harrow ready?. With Kylon: 10 accounts reviewed on time. Caught a 14-month-old address. Workflow tools: Box, Gmail, Google Drive, HubSpot, Outlook, Slack.
KYC Periodic Refresh: Each Due Account Reviewed Against the Whole Policy
Due accounts reviewed against documents, screening and the policy.
Before: Q3_financials.xlsx, certificate late, LN-2214: compliant, Re: Kemp ratios, who watches 2214?. With Kylon: 4 breaches flagged early. Recomputed ratios on 28 loans. Workflow tools: Box, Gmail, Google Drive, Outlook, Salesforce, Slack.
Loan Covenant Monitoring: Ratios Recomputed Whenever a Borrower's Numbers Change
Covenant ratios recomputed from borrower statements each quarter.
Before: MA-122 dup of 121?, bank_stmt_3.pdf, test charge failed, Merchant: any news?, reserve rate?. With Kylon: New merchant live this week. Re-ran MA-2026-122 checks. Workflow tools: Box, Gmail, Google Drive, HubSpot, Slack, Stripe.
Merchant Underwriting: Each Application Decided on Its Own Packet and Test Charges
Applications decided from the ticket, packet and test charges.
Questions teams ask
- How can an investment firm or family office use AI agents?
- Intros, diligence lists, portfolio updates and the quarterly letter to principals often run through a few people and their inboxes. In Kylon, that work runs in rooms on one shared record: agents read the email and documents, track requests and draft updates, and a partner approves everything that leaves the firm.
- Can AI agents monitor loan covenants?
- Yes. Each quarter agents compute ratios from borrower statements in Box, read compliance certificates in Outlook and test each loan against its covenants, recomputing whenever a borrower's numbers change. The result decides who watches the loan, and the portfolio manager approves each disposition and tier before Salesforce changes.
- Who approves the decisions the agents prepare?
- The person with authority for that decision. The KYC lead approves account status changes, the credit lead approves limits and holds, the underwriting manager approves merchant decisions before the merchant hears back, and a compliance officer issues every trade preclearance. Agents propose each decision with the evidence attached.
- Where should a financial services firm start?
- Start with KYC periodic refresh. Due accounts arrive on a schedule, the policy is written down and the evidence is spread across Box, screening results in Outlook and HubSpot. Agents propose each account's status against the whole policy and the KYC lead approves every change, which gives a clear audit trail from the first cycle.