Platform · overviewIn development · not for sale yet

One workspace,four jobs.

One place to manage your marketing, score it, have agents do the work and measure what came of it. It is being built now, in the order the ad platforms grant access. None of it can be bought yet. What exists today is the free Troy Score and Troy’s team.

Get a free Troy ScoreSee what Troy’s team does
Product preview · in development

The four jobs, one at a time.

  1. Every channel in one place.

    Plans, budgets, creative and calendars for search, social, email and content, side by side instead of in six tabs.

    • One plan across channels
    • Budgets and pacing at a glance
    • Approvals in one queue

  2. Know where you stand, and why.

    Troy Score reads your connected accounts and grades what it finds. Every point it gives or takes away comes with the evidence.

    • Free to run
    • Scores with reasons, not just numbers
    • A ranked list of what to fix first

  3. The work gets done, with your say-so.

    Agents draft, build and propose. Anything that reaches an audience or spends money stops and waits for a named person.

    • Five specialised agents
    • Spend limits you set
    • A full trail of who approved what

  4. One number everyone agrees on.

    Results tied back to the decisions that caused them, so next week's plan is built on what worked rather than what was loudest.

    • Cross-channel reporting
    • Every change linked to its effect
    • Plain-language weekly summary
Build order

Read, then draft, then act.

The product is built in the order the ad platforms grant access. That is also the order in which an agent can be trusted, so the two never pull against each other.

An order, not a schedule. Each stage ships when the access it needs is granted.

Stage 1 of 3 · First

Read

What ships
  • Connections to Meta, Google Ads, Shopify and Klaviyo.
  • Troy Score v1.
  • A weekly report in which every figure traces to an API response.
The access it needs
Read-only. The lowest tier everywhere.
Why it comes first
Reading is the lowest access every platform offers, and it cannot change anything in your accounts. It is also where trust starts: before an agent proposes a change, you can check that it reads your numbers correctly.
The rule that binds here
No number appears in a report unless it came from an API response.
Stage 2 of 3 · Then

Draft

What ships
  • The approval queue.
  • Copy and creative variants under brand rules.
  • Approved work pushed as paused ads or drafts.
The access it needs
Writes to paused and draft objects only.
Why it comes second
The next thing a platform allows is writing objects that are not live. A draft can be wrong at no cost: it sits switched off until a named person approves it, edits it or throws it away. Each of those decisions tells Troy whether that agent is good enough to ship.
The rule that binds here
Nothing reaches an audience without a named person approving it.Meta, on its ads AI connectors: ads are created paused
Stage 3 of 3 · Last

Act

What ships
  • Budget and bid suggestions, approved with one click.
  • Hard caps on every suggestion.
  • A change log, and a revert for every change.
The access it needs
Live changes, each one approved.
Why it comes last
Changing something live is the last thing an agent should be trusted with, so it ships last. Every change an agent proposes here is approved by you, sits inside a cap you set, and can be reverted.
The rule that binds here
No spending authority without a cap. Changes are batched weekly, because frequent edits reset an ad’s learning phase on Meta.Meta, on the learning phase

What Troy can touch in your accounts

  • Read your accountsNew at Read
  • Write paused ads and draftsArrives at Draft
  • Change a live budget or bid, each change approved, under a hard capArrives at Act
  • Change spend with nobody watchingNot in the first version

Each stage keeps what the one before it could do and adds one thing: read first, then draft, then act under a cap.

At Read, Troy can: read your accounts.

The approval queue

The approval queue is the core of the product.

Agents draft, build and propose. Nothing reaches an audience or spends money without a named person approving it, and the queue is where that happens.

See the queue at work

  1. 01

    Every write lands paused or as a draft.

    An agent never creates anything live. What it builds sits in your own account, switched off, until a named person approves it.

  2. 02

    Each tool has its own policy.

    Every tool an agent can use is set to one of three. Whichever it is, the first version makes no change to spend unattended.

    • Auto-approve
    • Always ask
    • Ask above a threshold
  3. 03

    Spend caps are hard.

    A cap is a limit the agent cannot pass. It is enforced at the platform level, not only in Troy’s code, and there is no spending authority without one.

  4. 04

    The change log is permanent.

    Every change is recorded with the approval behind it, and every change can be reverted.

  5. 05

    An approval can wait for days.

    A proposal keeps its state while it waits. If the decision needs the weekend, it takes the weekend.

Of the ad platforms’ own connectors, only Meta’s creates ads paused. On Google and TikTok the queue is entirely Troy’s job, which is why it ships before any live change does.

Connections · planned

Four first. The rest when access allows.

Troy connects to the tools you already use and does not replace them. None of these is connected today. This is the order they are planned in.

Showing 6 of 6

  • Meta

    Ads

    First · Read stage

  • Google Ads

    Ads

    First · Read stage

  • Shopify

    Store

    First · Read stage

  • Klaviyo

    Email

    First · Read stage

  • TikTok

    Ads

    Planned · No date

  • LinkedIn

    Ads

    Planned · No date

  • These are planned.

    None is connected today. While the platform is being built, a Troy strategist works from a user seat on your own accounts.

  • Access is each platform’s to grant.

    App reviews are discretionary and have no deadline. That is why this list has an order and no dates.

  • Nothing is built on scraping.

    Every connection uses the platform’s own access, and the tokens belong to you.

Scope

Not in the first version.

Saying what is left out is part of saying what it is.

  • Self-serve signup and billing.The first workspaces open by invitation. Yourself is a waitlist.
  • Unattended spend changes of any size.Every live change waits for a named person.
  • Attribution modelling, or any claim that the Score predicts results.The Score grades the state of your marketing. No claim of cause without an experiment.
  • Troy’s own AI models.
  • A CRM, an email-sending service or a social calendar.Troy connects to your tools. It does not replace them.
  • Handling client media money.You pay the ad platforms directly. Troy never holds media money.

What you can have today

Available nowTroy ScoreFreeAn independent read of your marketing, prepared by a strategist while the platform is built. Every point shows its evidence, and the fixes come ranked.How the Score works Available nowTroy’s teamFrom $4,000 a monthTroy’s strategists run your marketing against goals you set. Core is $4,000 a month and Scale is $8,500 a month.What the team does
Next

Start with what exists today.

The platform is being built. The Troy Score is free now, prepared by a strategist, and the list of fixes is yours whatever you decide.

Get a free Troy ScoreHow the agents work