Skip to content
How It Works

Built for Asset Managers.
Fully Managed. Always Running.

Cloud-hosted on dedicated infrastructure, managed by our team. 26 AI agents working as a single unified team automate your IR, compliance, operations, research, and content 24/7. Nothing to install, nothing to maintain.

14,700+ Knowledge Chunks Indexed26 Purpose-Built AI AgentsAlways Running 24/7Works With Your Existing Systems
AssetBoxxClient ExperienceDeliveryBDRSecurityOps IntelContentKnowledge BaseCommand CenterExecutive Asst.

One Conversation.
Twenty-Six Specialists Behind It.

Ask a question in Slack. The right agent picks it up, pulls context from your systems, and delivers one answer.

Your Slack MessageCommand CenterKnowledge BaseInvestor RelationsCompliance MonitorResponse AssembledYour Answer in Slack

You Ask

Type a question in Slack. The system reads intent and routes to the right specialist: IR, compliance, research, or operations.

Agents Work

Each agent pulls what it needs: Salesforce records, Addepar portfolio data, compliance libraries, fund documents. Scoped to the right fund and LP automatically.

You Get One Answer

One question, one sourced response with citations, right in Slack.

Every Document. Every Filing. Every LP Communication.
Searchable in Seconds.

Your firm's institutional knowledge becomes searchable through plain English questions in Slack.

Consistent Answers, Every Time

Junior team members get sourced answers from official compliance libraries, fund documents, and approved talking points, not whoever happens to be available.

What Your Team Can Search

Compliance rules and regulatory guidance. Investment research and economic data. Fund documents and PPMs. LP communications history. DDQ response libraries. Meeting transcripts and notes. All indexed across 13 RAG collections with 14,700+ knowledge chunks.

Built-In Privacy Controls

Fund-level data isolation enforced at the query level. LP-specific information restricted by role. Compensation, personnel, and side letter data blocked from general queries.

SalesforceBloombergAddeparGmailJuniper SquareProcessing PipelineIndexChunkEmbedVector Database"What is the process for..."Sourced answer with citationsFrom official SOP documents

AI Costs That Stay Predictable

Smart model routing keeps costs low. Simple tasks use cheap models. Important tasks use powerful ones.

Low CostPennies per task

Email triage, document sorting, template filling, routine compliance checks

Fast, lightweight AI models

StandardA few cents per task

DDQ draft responses, meeting briefings, daily summaries, LP communications

Mid-tier reasoning models

PremiumStill cents, not dollars

Quarterly letters, investor-facing reports, complex compliance reviews, market commentary

Top-tier models, only when it matters

FreeIncluded

Search, indexing, document classification, spam detection

Local models on dedicated infrastructure

Why It Stays Cheap

Before anything reaches a premium model, context is compressed first: fund performance data, LP history, and compliance requirements get distilled into a focused brief. Costs stay 70 to 95% lower than raw input.

If Something Breaks, It Fixes Itself.

The system detects problems, resolves them, and tells you what happened.

Service Health

Last check: 47s ago
Email Processing

99.9%

Compliance Monitoring

99.9%

Knowledge Base

99.9%

Slack Agents

99.9%

Draft Service

Restarting...

System Monitor

99.9%

LP Communications

99.9%

Prospect Enrichment

99.9%

Document Sync

99.9%

Meeting Processing

99.9%

Alert System

99.9%

Regulatory Watch

99.9%

Security Scanning

99.9%

Operations Brain

99.9%

Content Pipeline

99.9%

Quarterly Letters

99.9%

DDQ Response Engine

99.9%

Performance Reporting

99.9%

Capital Call Processing

99.9%

ESG Data Collection

99.9%

Market Commentary

99.9%

Research Aggregation

99.9%

Portfolio Analytics

99.9%

Brand Guardian Review

99.9%

LP Onboarding

99.9%

Continuous Monitoring

A watchdog checks every service every 60 seconds. Failures trigger automatic restarts, most resolving in under a minute.

Smart Alerts, Not Noise

Failed auto-restarts produce a single Slack alert. Follow-ups post as threaded replies. Only genuine issues reach the critical channel.

Full Recovery on Reboot

If infrastructure restarts, every service comes back online automatically. No intervention required.

Automation Applied to the Right Problems.

The tasks that consume the most asset management time, handled automatically.

Before (Manual)

1

Gather fund performance data, draft narrative sections, incorporate portfolio commentary, review for compliance, coordinate approvals across investment team and legal, format for each LP class.

2-3 weeks

Total: 2-3 weeks per fund, per quarter

After (AssetBoxx)

Agents pull performance data from Addepar, draft narrative sections from templates and market commentary, run compliance checks against firm guidelines, and present final drafts for human review and approval. Class-specific versions generated automatically.

Total: 2-3 days with human review and approval

Time saved: 10-15 business days per fund, per quarter

Most of This Happens Without You Asking

Not a chatbot. An operations layer running on schedules and triggers.

24/7 Operations

Always running. Always watching.

12 AM
3 AM
6 AM
9 AM
12 PM
3 PM
6 PM
9 PM
TriageMonitoringReportingSecuritySync

Continuous

Every 2.5 minEmail poller checks inboxes for LP inquiriesExecutive Assistant
Every 60 secCompliance monitor checks flagged itemsCompliance Monitor
Every 60 secWatchdog checks all servicesCommand Center
Every 2 minSecurity scanning runsSecurity
Every 5 minDocument sync checks for new filingsDocument Manager
Every 60 secSlack monitor checks unanswered questionsKnowledge Base
Every 10 minPortfolio data refresh from connected systemsPortfolio Analytics

Daily

7:00 AMOvernight LP inquiry reportInvestor Relations
7:30 AMMarket data briefingResearch Analyst
8:00 AMMorning triage report with LP prioritiesExecutive Assistant
9:00 AMCompliance alert summaryCompliance Monitor
4:00 PMEnd-of-day operations summaryFund Operations
5:00 PMRegulatory watch digestRegulatory Watch

Weekly

Monday 8 AMWeekly LP communications summaryInvestor Relations
WeeklyKnowledge base sync across all collectionsKnowledge Base
Sunday 4 AMFull security auditSecurity
WeeklyRegulatory change digestRegulatory Watch
WeeklyContent calendar review and topic suggestionsContent Expert

Ongoing

ContinuousLP prospect enrichment pipelineProspect Enrichment
On DemandDDQ response draftingDDQ Response
On DemandQuarterly letter productionQuarterly Letter Writer
On DemandCompliance review on all investor-facing contentBrand Guardian

Plugs Into What You Already Use

Connects to your existing systems. Nothing changes for your team.

CRMLP relationships, pipeline, and communicationsMarket DataReal-time market data and researchPortfolio ManagementPerformance data and portfolio analyticsCommunicationWhere your team already worksDocumentsSubscription docs, side letters, and agreementsFund AdministrationCapital calls, distributions, and LP portalAssetBoxxFully Managed

Integrations are modular. Start with your CRM and communication tools or connect all of them.

Gets Better the More You Use It

Every correction teaches the system how your firm communicates with investors.

Your Inbox Gets Smarter

LP inquiries, consultant requests, and regulatory notices are automatically categorized. Within a few interactions, the system learns your firm's communication priorities.

Learns from every interaction

Drafts Match Your Firm's Voice

Edits you make before sending are stored per LP and per fund. The system uses them as examples for future drafts. Fewer corrections over time.

Per-LP and per-fund style calibration

Knowledge Gaps Surface Automatically

Questions the system cannot answer from your compliance library or fund documents show you exactly which SOPs, policies, or approved talking points need updating.

Surfaces what your knowledge base is missing

0

Purpose-Built Agents

A single unified team across IR, compliance, operations, research, and content

0+

Knowledge Chunks Indexed

Across 13 RAG collections

0-15

Days Saved on Quarterly Letters

Per fund, per quarter

0s

Health Check Interval

Every service, every minute

0

RAG Collections

Compliance, research, operations, and more

How We Know It Works

2-3 days

quarterly letter production time, down from 2-3 weeks per fund.

Source: AssetBoxx production deployment

75%

of asset managers cite operational inefficiency as their top scaling constraint.

Source: BCG Global Asset Management Report, 2024

14,700+

knowledge chunks indexed across compliance rules, investment research, and economic data.

Source: AssetBoxx RAG system, 13 collections

Every asset manager we worked with had the same bottleneck: institutional knowledge trapped in spreadsheets, compliance workflows held together by email, and IR teams drowning in manual processes.

David Ward, CEO, AssetBoxx

Common Questions

Initial deployment follows four phases: onboarding (week 1-2) to map your fund structures, LP relationships, and compliance requirements; agent configuration (week 3-4) to connect your systems and load your knowledge base; supervised break-in (week 5-8) where you approve everything and agents learn your standards; and production (week 9+) where calibrated agents operate with earned autonomy. The system improves continuously as it absorbs your firm's communication patterns, compliance preferences, and institutional knowledge.

No. You interact entirely through Slack with Approve, Edit, and Reject buttons. Behind the scenes, all services are monitored 24/7 with a watchdog checking every service every 60 seconds. Failures trigger auto-restart, and AI diagnostics resolve issues autonomously. Our team handles all configuration, integrations, updates, and security patching so your firm never needs to touch infrastructure.

Your data lives on dedicated infrastructure with no shared resources between firms. Each deployment is completely isolated at the database, file storage, and AI model level. Fund-level data segregation ensures no cross-fund data leakage. Contractual data processing agreements with all AI providers explicitly prohibit training on your data. The entire system has zero public internet exposure.

Yes. Most firms begin with IR and compliance agents, which address quarterly letters, DDQ responses, and compliance monitoring immediately. From there, teams add research, operations, or content agents based on their priorities. The modular architecture means each agent can be deployed individually without disrupting existing workflows.

Integrations are modular. The agents are system-agnostic; only the connector changes. We support Salesforce, Bloomberg, Addepar, Slack, Gmail, DocuSign, and Juniper Square out of the box, with additional integrations configured during deployment.

Monthly platform fee plus AI usage. Tiered routing keeps AI costs low: cheap models for document classification and triage, premium models only for investor-facing quarterly letters and compliance reviews where quality matters.

A central Command Center routes queries to the right agent based on topic. When a request spans multiple agents, such as a quarterly letter that needs performance data, market commentary, and compliance review, the orchestrator coordinates their responses and assembles a unified output. Each agent maintains its own Slack channel, knowledge context, and approval workflow.

Every investor-facing output goes through human approval before it reaches anyone. If you edit a draft, the correction is stored and improves future output for that fund and LP. The system learns from corrections, not just prompts.

Compliance is embedded in every agent's workflow, not bolted on as a separate review step. The Compliance Monitor agent checks all investor-facing content against SEC marketing rule requirements, FINRA guidelines, and your firm-specific compliance manual. The Brand Guardian ensures tone and positioning consistency. Human compliance officers make all final decisions.

See It Running Live

No slide deck. We show you the live system: Slack channels, agent responses, dashboards. On a call. With your questions.

Every number on this page comes from the production system you will see on the call.