How it works

From a signed client to a report they trust

Aegis is built around the way an agency actually runs: acquire, staff, operate, prove, keep. Each step below is a working part of the product, not a roadmap item.

  1. 01

    Set up the agency and its clients

    One workspace per agency, one record per client brand. Team members get roles — or a custom role you build — and an account manager can be confined to the brands they actually work on.

    • Roles with granular permissions, enforced by the API and again by the database
    • Custom roles that can never exceed what the person granting them holds
    • A full audit trail, with a tamper-evident chain over it
  2. 02

    Bring in products and creators

    Import creators from a CSV with validation, a preview and duplicate detection before anything is written. Connect a shop where the platform offers an API and the client grants access; where it does not, the manual path is a first-class workflow rather than an apology.

    • CSV import: upload, validate, preview, detect duplicates, confirm, then a report of what failed and why
    • Integrations use published APIs under permissions the customer grants
    • A capability a platform does not expose is documented, not worked around
  3. 03

    Score and shortlist

    Score creators against the client you are staffing: sales potential, audience fit, product fit, engagement, content quality, reliability and growth — weights you set per agency, client or campaign.

    • Every score shows its components, its reasons and how fresh the data is
    • A missing signal is named as missing; it is never quietly treated as a zero
    • Predictions come as conservative, expected and optimistic, with confidence
  4. 04

    Reach out, and follow up

    Sequences with templates, personalisation and AI drafting. They stop the moment somebody replies, respect quiet hours and a daily cap, and record every message that was sent and every one that was deliberately held.

    • Only authorised channels — nothing here bypasses a platform control
    • Replies are classified, assigned and prioritised in one inbox
    • Where a channel needs a person to send, the queue is part of the product
  5. 05

    Samples, content, sales

    Track a sample from request to delivery to the post it produced, with the cost of the product and the shipping attached — which is what turns "we sent fifty samples" into a number with a return on it.

    • Sample status, carrier, tracking, cost and the content it was meant to produce
    • Content with its metrics, approvals and usage rights
    • Overdue content escalates instead of being forgotten
  6. 06

    Analytics, attribution and profit

    Campaign, creator, content, product, order. Aegis joins what the data supports and labels the rest as unknown rather than filling the gap. Costs come off the top so each client shows a contribution profit rather than a revenue figure that flatters everybody.

    • Direct, estimated and unknown attribution are separate columns, never merged
    • Forecasts always carry an interval — a point estimate alone is not shown
    • Your own cost assumptions, applied consistently across clients
  7. 07

    Report, and let the client look

    Weekly, monthly, campaign and executive reports built from the metrics that exist, in a structure a client can follow: what happened, why, what worked, what did not, and what happens next. A client portal shows them their own brand and nothing else.

    • Reports export to PDF, CSV and Excel, and can carry your brand
    • The portal redacts internal notes, margins and other clients by construction
    • A shared report link is scoped, expiring and revocable

What Aegis will not do

It will not scrape a platform that has not authorised it, solve a CAPTCHA, work around a rate limit, or act on an account nobody gave it permission to act on. Where a platform does not expose a capability, the product says so and gives you a compliant way to do the work by hand.

It will not invent a metric, guarantee revenue, or present a prediction as a fact. If the data cannot support an answer, the answer is that the data cannot support it — which is less impressive in a demo and considerably more useful in a client meeting.

Try it on one client

A free trial with your own data is worth more than any demo. Import a list, run a campaign, and see whether the numbers hold up.