We implement Anthropic's financial-services suite in your shop — reference agents, skills, and data connectors, wired to your providers and tuned to your process. Every output staged for professional review.
Implementation engagement, not a plugin
Claude for Financial Services is a service we deliver, not a product we sell. We are the implementation, integration, and enablement partner.
Anthropic ships the models and an open FSI suite — reference agents, skills, and data connectors, delivered on Claude Cowork, Claude Code, or the Managed Agents API, plus Microsoft 365 tooling (Claude for Excel and PowerPoint). The suite is open source under Apache-2.0.
We make it real in your environment: installed, connected to your market-data and document systems, taught your models and templates, and rolled out to your analysts and ops team with the controls your compliance function expects.
Best wedge first: fund administrators and fund finance/ops teams, then independent equity-research shops and boutique advisory, small PE/VC funds and family offices, and RIAs. We typically start where there is already a market-data subscription in place.
What We Deliver
Nine deliverables that take you from assessment to an adopted, governed deployment.
A paid, fixed-fee audit of your workflows, data systems, and security/compliance posture, with a recommendation on which agents and which Claude plan fit. Not free discovery.
Claude Cowork, Claude Code, the Managed Agents API, and Claude for Excel/PowerPoint — stood up for your team.
The official reference agents and skills relevant to your desk, installed and configured for how you work.
We tune the agents to your models, templates, and house style with your real documents so the output is yours.
Connected to your market-data and document providers, running on your own subscriptions.
Custom Claude Skills and MCP servers for the workflows you need that do not exist yet (billable custom work).
Access controls, audit trails, human-sign-off gates, and a compliance posture your firm can stand behind.
Workshops, internal champions, and playbooks. Adoption is a real deliverable, not a throw-in.
A retainer for new skills, new connectors, model upgrades, monitoring, and support.
The official agents
Anthropic publishes these as open source (Apache-2.0). We deploy the ones that fit your desk, connect them to your providers, and tune them to your templates and review process.
Pitch materials and meeting preparation.
Market work and earnings coverage.
Build models and review marks.
Reconciliation, close, and statement checks.
Document parsing and rules checks for KYC files, staged for your compliance officer.
Skill bundles
Installed per team. The core bundle carries the data connectors, so it goes in first.
Valuation and modelling in Excel, in your templates.
Sell-side materials and process management.
Coverage, earnings, and idea generation.
Sourcing through portfolio monitoring.
The operational core: reconciliation, accruals, and NAV.
KYC document handling.
From Anthropic’s general finance plugin: close, reporting, and controls.
Anthropic also publishes a wealth-advisor reference implementation (pre-meeting prep, rebalance review, a Marketing Rule pre-check). It is not actively maintained upstream, so we treat it as a starting point and maintain your copy. Names are as published in Anthropic’s financial-services repository at the time of writing.
Connectors
Every market-data connector runs on your own vendor entitlement. We do not resell data.
Fundamentals, estimates, ratings, and private markets.
What management said and what the market heard.
Portfolio monitoring and data rooms.
Your own warehouse, through the finance plugin.
The surfaces your team works in.
Example Workflows
Representative tasks across research and operations — every output staged for human sign-off.
Build and update comparable, DCF, and LBO models in Excel against your assumptions.
Draft pitch decks and CIMs in PowerPoint and Excel from your data and templates.
Turn filings and transcripts into structured, source-attributed earnings notes.
GL reconciliation, month-end close, NAV tie-out, and LP statement audit support.
Assemble and structure KYC screening packages for analyst review.
Investment banking, equity research, private equity, wealth management, fund admin, and operations.
How It Runs
Runs on Claude’s own runtime and wires into the providers your desk already pays for.
Pick the surface that fits your team.
Require your own provider subscriptions.
Built for professional review.
Engagement & Investment
Implementation pricing is bespoke; these are typical starting points, not a price list.
from $1,500
A fixed-fee assessment and pilot to size the work and prove value.
$8,000–25,000+
Scoped by desks/agents, connector count, and custom skills.
$2,000–5,000/mo
New skills, new connectors, model upgrades, monitoring, and support.
Important
AI Agents Plus provides the implementation, integration, custom build, training, and support — billed by us. Claude itself (Cowork seats, Claude Enterprise, or API) and any third-party market-data and system subscriptions (FactSet, S&P, PitchBook, Morningstar, and the rest) are your own subscriptions, billed by Anthropic and those vendors. We are not reselling Claude licenses or data feeds; part of our value is advising on the right plan and helping with procurement, including AWS Marketplace billing.
How We Work
A staged engagement. A focused first deployment usually takes a few weeks.
01
Audit workflows, data, and compliance posture; recommend the agents, connectors, and Claude plan.
02
Stand up the runtime and install the official reference agents and skills for your desk.
03
Tune to your models and templates and wire connectors to your data providers.
04
Workshops and champions drive adoption; a retainer keeps it improving.
FAQ
Common questions from research shops, funds, and fund administrators
No. Claude for Financial Services runs on Claude’s own runtime (Cowork, Claude Code, or the Managed Agents API) — it is independent of Hermes. If you also want an always-on operations and communications assistant, a Hermes AI employee can be deployed alongside it as a complementary layer; Claude-for-FSI is the senior specialist.
Yes. Connectors to providers like FactSet, S&P Global, PitchBook, Morningstar, Daloopa, and others run on your own subscriptions. A good fit is a firm that already pays for at least one market-data provider — we help scope and connect the rest.
Outputs draft analyst work product for professional review with source attribution, on a SOC 2 / FedRAMP posture, and your data is not used for training. It does not make investment recommendations, execute trades, or post to ledgers — every output is staged for human sign-off.
Fund administrators and fund finance/ops teams are the strongest wedge (GL recon, month-end, NAV, LP statements — ops-flavored, low MNPI exposure), followed by independent equity-research shops and boutique advisory, small PE/VC funds and family offices, then RIAs.
A focused first deployment usually takes a few weeks. The exact timeline depends on desks, connectors, and any custom skills. We start with a fixed-fee readiness assessment that scopes the work precisely.
Start with a fixed-fee readiness assessment — we will recommend the agents, connectors, and plan that fit your firm.
Compare adjacent offerings and choose the right implementation path for your goals.
Read practical guides and implementation insights aligned with this service area.