Merton AI

Agentic AI for sales & trading

Meet Merton,
your desk’s
newest analyst.

Merton is a private AI agent trained on market data, alternative data, and your desk’s own history — built to think like your best traders, work alongside your team, and never leave the building.

AUDITED BY US BULGE-BRACKET SECURITY TEAMS
merton — desk session · EON CAPITAL
LIVE
80+ years
of senior sales & trading judgment encoded into Merton’s training
100%
private — your data trains and informs only your agents
5
model providers supported, chosen per task without compromising security

How merton works

An analyst that already knows your book

Merton sits next to your salespeople and traders, not in front of them. It’s trained on your desk’s data, your firm’s playbook, and the market itself — then put to work on the tasks that used to eat up your day.

Private by design

Secure, private agents trained on market data, public and alternative data, plus your desk’s own data — used only for you, never pooled across clients.

Senior judgment, encoded

Trained directly by DataDock’s executives, drawing on more than 80 years of combined senior sales and trading experience.

Behavioral simulation

Runs DataDock’s simulation engines to analyze customer profitability and surface optimal trading decisions across different behavioral lenses.

Multiplies your best people

Augments your most effective salespeople and traders — supercharging productivity and revenue while reducing cost per seat.

Best model for the job

Runs on Claude, Gemini, AWS Bedrock, Titan, and open-source models — chosen per task, without compromising security or data integrity.

Lighter operational load

Automates the routine work currently handled by junior, operational, and support staff — freeing your team to focus on clients.

Merton in action

From question to hedged trade in five turns

A real exchange between a trader and Merton — analysis, structuring, distribution, execution, and hedging, in one conversation.

client eon capital
underlying xle
3-mo p&l +$1.05mm
structure APR 65C → MAY 62C ROLL
1
Analyze the relationship
Tina · Trader
What has the P&L been with Eon Capital over the last 3 months?
Merton · AI Agent
Based on DataDock simulations, P&L for Eon over the last 3 months ranges from +$800k to +$1.2mm depending on behavioral lens. On an aggressive 20-day lens: P&L was +$1.05mm, total revenue incl. commissions was $1.23mm.
2
Find the driver
Tina · Trader
What are the largest drivers of that revenue?
Merton · AI Agent
Biggest trade: client bought 50k XLE April 65 calls @ $1.50 (stock ref $60.5) on March 25 — driving nearly all the positive P&L (~$1.2mm). All other trades with this counterparty netted between –$100k and +$100k.
3
Structure the idea
Tina · Trader
Should we suggest a roll to this customer?
Merton · AI Agent
Yes — show a May roll at $0.70 offer. Client sells 50k XLE April 65 calls and buys 50k May 62 calls. Market mid is $0.67, theo mid $0.65. Makes sense given $0.10 current value and only 6 days to expiry.
4
Distribute the risk
Tina · Trader
The customer is bidding $0.68. Who should we show this trade to?
Merton · AI Agent
Try Neptune Capital, Orion Capital, and Artemis Capital — all active in XLE and sellers of calls this quarter. Target ≥50% internalization. If you need to lay off more, Bank Street is our go-to at ~$0.03 over theo.
5
Execute & hedge
Tina · Trader
We got $0.69 and crossed the entire 50k. How should we hedge?
Merton · AI Agent
Hedge delta via 3% participation VWAP. Eon typically has low initial delta edge and tends to buy vol — holding 20 days has been P&L-positive historically. No need to unwind vega right now.
Outcome
Trade complete — booked, crossed & hedged
$0.69
Fill — one cent through the client’s bid
50k
Contracts crossed in full
VWAP 3%
Delta hedge in motion
Held
Vega kept — historically P&L-positive vs. Eon

Security & trust

Audited where it matters most

Merton’s cyber and AI security has been reviewed and audited by major financial institutions, including US bulge-bracket banks. Your data trains your agents — and stays yours.

Your data, your agents Market, public, alternative, and your own desk data — used only for your firm.
Bulge-bracket audited Reviewed by the security teams of major financial institutions before deployment.
Model-agnostic, security-first Claude, Gemini, AWS Bedrock, Titan, and open-source models, chosen without compromising data integrity.

Common questions

What desks ask us first

Does our data train models for anyone else?

No. Your market, public, alternative, and proprietary desk data trains agents for your firm only. Nothing is pooled across clients, and nothing leaves your environment.

Which models does Merton run on?

Merton selects between Claude, Gemini, AWS Bedrock, Titan, and open-source models per task — balancing capability, latency, and cost without compromising security or data integrity.

Who trained Merton’s trading judgment?

DataDock’s executives — with more than 80 years of combined senior sales and trading experience — trained Merton directly, alongside DataDock’s behavioral simulation engines for customer profitability analysis.

Has the security been independently vetted?

Yes. Merton’s cyber and AI security has been approved by and audited by many large financial institutions, including US bulge-bracket banks.

What roles does Merton actually take on?

Merton augments your most effective salespeople and traders, and automates tasks currently performed by junior, operational, and support staff — from relationship P&L analysis to structuring, distribution lists, and hedging recommendations.

get started

Put Merton on your desk

See how Merton handles your book, your customers, and your playbook — in a private session with your own data.