MINDFAX OPS  /  The Operations Catalog

Every operation we can run for you.

A guided catalog of the AI agents we build, host and monitor for agencies and consultancies. Browse by category, read a real before→after example for each, and see the complexity level. Every agent runs on our infrastructure with live telemetry — and a human approves anything irreversible.

Level 1 · SimpleOne source to one output — a scheduled sync, a dashboard, an alert.
Level 2 · StandardSeveral tools with logic and branching — the everyday workhorse.
Level 3 · AdvancedOrchestration, generated drafts, exceptions and judgment across systems.
Gold marks the human gate: sending, deleting, paying or publishing always waits for a person's approval.
How to read this catalog

Three levels of complexity. One rule that never changes.

Complexity is about how much wiring, logic and judgment an automation needs — not about how valuable it is. A simple daily alert can save a client relationship; an advanced orchestration can quietly run half your ops. Use the levels to size the effort, then pick the plan by how many you want running and how fast you want new ones built.

LEVEL 1 · SIMPLE

Do one thing, reliably

A single trigger and a few steps across one or two tools. Set it once, it runs forever with monitoring on top.

e.g. a scheduled dashboard refresh, or a Slack alert when a metric drops.
LEVEL 2 · STANDARD

Move work across tools

Several integrations with branching logic — parse, decide, route, write. This is where most agency time actually goes.

e.g. a lead enriched and routed into the CRM, or an invoice drafted for approval.
LEVEL 3 · ADVANCED

Run a whole process

Multiple agents, generated drafts, exception handling and a human decision point. It operates a process, not a step.

e.g. contract-to-kickoff onboarding, or the monitoring layer over every other agent.
Levels × plans

What each plan runs

Plans differ by how many agents run and how fast we build new ones — not by a cap on complexity. Even Start can run an advanced agent if that's your highest-impact one. This is where each plan usually lands.

 
Start
$900 / mo
Operate
$1,600 / mo
Scale
$2,500 / mo
Agents running
1 production agent
Up to 3 production agents
An agent portfolio across your ops
New automations
First agent built & included
+ one new automation each quarter
Continuous build queue
Complexity you usually run
Your single highest-impact agent — often L1–L2, L3 if that's the one
A working set across L1–L3, including your first orchestration
A full L1–L3 portfolio with multi-agent orchestration
Hosting, monitoring & monthly report
Included
Included + priority direct channel
Included + monthly ops-review call
Guarantee
First agent in production within 30 days or the first month is refunded
Same 30-day guarantee
Same 30-day guarantee
01 · Client Reporting 02 · Client Onboarding 03 · Data Movement & Sync 04 · Lead & Pipeline Ops 05 · Content Ops 06 · Billing & Finance Ops 07 · Status, Comms & QA
CATEGORY 01

Client Reporting

Agents pull performance data from your analytics and ad platforms and assemble branded, on-time client reports, dashboards and alerts — so nobody rebuilds the same deck by hand.

SERVICE 01L3 · ADVANCED

End-of-month client report assembler

Pulls each client's metrics on schedule, drops them into a branded template with draft commentary, and queues it for sign-off.

Before

An account manager burns most of a week logging into GA4, Search Console, Google Ads and Meta, copying numbers into a deck and writing the "what happened".

→
After

On the 2nd business day the agent builds the branded report with draft commentary and drops it in a review queue.

GA4Search ConsoleGoogle AdsMeta AdsSlides / LookerSlack
Impact Reclaims about a week of manual reporting each month — and every client report ships on the same day.
Human approves: reviews the numbers & commentary and clicks send to the client.
SERVICE 02L1 · SIMPLE

Live client KPI dashboard refresher

Keeps an always-on dashboard current by syncing platform metrics into one place on a schedule.

Before

A marketer manually exports CSVs each week into a Looker Studio board and hopes nothing broke.

→
After

The agent refreshes every connector on schedule and flags any source that failed to load.

GA4Google AdsMeta AdsGoogle SheetsLooker Studio
Impact Dashboards stay current on their own; no more weekly CSV exports or a board nobody trusts.
Read-only — nothing irreversible. A human only investigates flagged gaps.
SERVICE 03L1 · SIMPLE

KPI drop & anomaly alert

Watches key metrics and posts an alert the moment something moves past a threshold.

Before

You find out at month-end that spend spiked or conversions cratered — after the damage.

→
After

The agent checks daily and posts to Slack when a metric moves >15% vs. baseline, tagging the owner.

GA4Google AdsMeta AdsSlack / Email
Impact Catch spend spikes and conversion drops the day they happen, not at month-end.
Read-only alert — no irreversible action.
CATEGORY 02

Client Onboarding

Agents turn a signed contract into a fully set-up project — accounts, tasks, folders, credentials and a welcome sequence — so "signed" is followed by momentum, not days of silence.

SERVICE 04L3 · ADVANCED

Contract-to-kickoff onboarding runner

When a contract is signed, it creates the project, tasks and folders, assigns owners and fires the welcome + credentials sequence.

Before

"Signed" is followed by days of nothing while someone manually spins up an Asana project, a Drive folder and a kickoff email.

→
After

Off the DocuSign "completed" event, the agent builds the project from a template, creates the folder and sends the welcome with an access checklist.

DocuSign / PandaDocAsana / ClickUpGoogle DriveHubSpotGmail
Impact Signed-to-kickoff drops from days to hours, and every new client is set up the same way.
Human approves: reviews the auto-created project & scope before the welcome goes out.
SERVICE 05L1 · SIMPLE

Credentials & access chaser

Tracks which client logins and ad-account grants are still missing and nudges until complete.

Before

An account manager keeps forgetting to chase the client for Meta Business Manager access; the project stalls.

→
After

The agent sends day-3 and day-7 reminders and pings the AM in Slack if access is still pending.

GmailSlackAirtable
Impact No project stalls waiting on access, and nothing gets forgotten.
Reminders only — no irreversible action.
SERVICE 06L2 · STANDARD

Intake form → client workspace setup

Turns a completed intake form into structured client records across the CRM and PM tool.

Before

Someone copy-pastes a Typeform intake into HubSpot and Notion by hand, with the usual typos.

→
After

The agent standardizes the fields and creates the client record, project and brief page in one pass.

TypeformHubSpotNotion / Airtable
Impact A new client is fully set up in one pass — no copy-paste, no typos.
Human approves: optional review of the created records before they go live.
CATEGORY 03

Data Movement & Sync

Agents keep the same record consistent across your CRM, project tool, sheets and billing — so nobody re-types data and the tools stop disagreeing with each other.

SERVICE 07L2 · STANDARD

CRM ↔ project-tool two-way sync

Keeps deal and client status consistent between the CRM and the PM tool so both always agree.

Before

A "won" deal in HubSpot doesn't reach Asana until someone re-enters it; statuses drift apart.

→
After

The agent mirrors stage and status changes both directions and logs any conflict for a human.

HubSpotAsana / ClickUp
Impact The CRM and the project tool always agree; no re-entry, no drift.
Human approves: resolves flagged sync conflicts only.
SERVICE 08L2 · STANDARD

Meeting-notes → CRM logger

Turns call transcripts into a clean CRM activity with summary, action items and next steps.

Before

A strategist means to type up the call and log it, then forgets; the context is lost.

→
After

The agent summarizes decisions and action items and writes them to the contact's HubSpot timeline with tasks.

Fireflies / OtterHubSpotAsana
Impact Every call is logged with action items; context stops dying after the meeting.
Human approves: reviews the summary before it becomes client-visible tasks.
SERVICE 09L1 · SIMPLE

Spreadsheet ↔ database keeper

Syncs a working Google Sheet with the system of record so exports and imports stop.

Before

Someone re-uploads a campaign tracker CSV into Airtable every week, and versions drift.

→
After

The agent watches for row changes and keeps both sides matched automatically.

Google SheetsAirtable
Impact One source of truth — exports and imports simply disappear.
Kept in sync automatically — no irreversible action.
CATEGORY 04

Lead & Pipeline Ops

Agents capture, enrich, route, respond to and nurture inbound leads — so nothing sits cold and your team spends time on conversations, not data entry.

SERVICE 10L2 · STANDARD

Inbound lead enrichment & router

Cleans a new lead, enriches it with company data, filters junk and creates the CRM record with an owner.

Before

A Typeform lead lands in an inbox and someone eventually looks up the company by hand.

→
After

The agent standardizes fields, drops personal-email domains, appends industry/size, and creates the HubSpot contact for the right rep.

Typeform / CalendlyEnrichmentHubSpotSlack
Impact Every lead is enriched and assigned in seconds, with junk filtered out before it hits the pipeline.
Internal records only — a human handles first outreach.
SERVICE 11L2 · STANDARD

Speed-to-lead first responder

Sends an instant, personalized acknowledgment and booking link the moment a qualified lead arrives.

Before

A lead waits until next day for a reply — and conversion quietly dies in the gap.

→
After

The agent replies within minutes with a relevant case study and Calendly link, then stops on reply or booking.

Web formsGmailCalendlyCRM
Impact Leads answered in minutes instead of next-day — protecting the conversions you already paid for.
Human approves: the message templates up front; sends can require approval if you prefer.
SERVICE 12L2 · STANDARD

Multi-touch nurture sequencer

Runs a timed nurture sequence for leads that didn't book, stopping the instant they engage.

Before

The owner means to follow up between client work, and the lead goes cold.

→
After

The agent runs a 5-touch / 14-day sequence and hands off to a human on any reply.

CRM / EmailSlack
Impact No lead goes cold in the gap between client work.
Human approves: sequence content once; a person takes over on reply.
SERVICE 13L3 · ADVANCED

Renewal & upsell watchdog

Watches contract end-dates and triggers a staged renewal/upsell workflow with draft materials.

Before

Renewals are handled in a last-minute scramble; upsells never get proposed.

→
After

At 60 days out the agent drafts a strategy review, prepares scope options at 30, and pings the AM to decide what to send.

CRMGmailSlackDocs
Impact Renewals are worked 60 days out and upsells actually get proposed — protecting recurring revenue.
Human approves: every client-facing renewal or upsell message (revenue-affecting).
CATEGORY 05

Content Ops

Agents draft, route for approval, schedule and repurpose content across briefs, calendars and channels — so the pipeline keeps moving and nothing publishes without a human's yes.

SERVICE 14L2 · STANDARD

Content brief → first-draft generator

Turns an approved brief into a structured first draft for a human editor.

Before

A writer stares at a blank doc and rebuilds the same structure from scratch each time.

→
After

The agent reads the Notion brief and keyword targets and produces an outline + first draft in the house format.

Notion / DocsSearch ConsoleSEMrush
Impact Writers edit instead of starting from a blank page, so content throughput goes up.
Human approves: a person edits and approves — nothing auto-publishes.
SERVICE 15L2 · STANDARD

Content approval router

Moves a draft through defined review tiers with SLAs and escalates if approvals stall.

Before

Approvals get lost in email threads; assets sit for days waiting on a nudge.

→
After

The agent routes reviewer → client with reminders and escalates to a manager after 24h of silence.

Notion / AirtableSlackForms
Impact Approvals stop stalling in email; SLAs are enforced automatically.
Human approves: each reviewer's approval is the gate — the agent only routes and nudges.
SERVICE 16L2 · STANDARD

Publish & repurpose distributor

Once content is approved, schedules it and spins channel-specific variants.

Before

A social manager manually cuts each blog into LinkedIn and X posts and schedules them one by one.

→
After

The agent drafts the variants from the approved piece and queues them in the scheduler for review.

Buffer / HootsuiteNotionLinkedIn
Impact One approved piece becomes every channel variant, queued for a human's yes.
Human approves: the scheduled queue before it publishes externally.
CATEGORY 06

Billing & Finance Ops

Agents handle invoicing, payment chasing and receipt capture — so cash is collected on time and money-related steps always wait for a person's approval.

SERVICE 17L2 · STANDARD

Recurring invoice generator

Creates and drafts recurring client invoices from the source-of-truth on schedule.

Before

Someone manually raises the same monthly retainer invoices in QuickBooks and occasionally forgets one.

→
After

The agent reads the retainer table and drafts each invoice for a bookkeeper to approve and send.

Airtable / SheetsQuickBooks / XeroStripe
Impact Every retainer is invoiced on time — no missed revenue.
Human approves: every invoice before it's sent (money-related).
SERVICE 18L2 · STANDARD

Overdue-invoice chaser

Detects unpaid invoices and runs a polite, escalating reminder sequence.

Before

Someone eyeballs the AR aging report and awkwardly chases payments when they remember.

→
After

The agent watches invoice status, sends staged reminders, and flags accounts past a threshold for a human call.

QuickBooks / XeroStripeGmailSlack
Impact Cash comes in faster and overdue balances get chased without the awkward reminders.
Human approves: tone up front; final escalation steps stay human.
SERVICE 19L2 · STANDARD

Receipt & vendor-invoice capture

Extracts vendor, amount and date from incoming bills and creates the accounting record for approval.

Before

Vendor bills are keyed from email into the ledger by hand, with the occasional wrong number.

→
After

The agent reads the PDF, extracts the fields and creates a draft bill routed for approval.

Gmail / OutlookOCRQuickBooks / XeroDrive
Impact Vendor bills are booked straight from the PDF instead of keyed by hand.
Human approves: each bill before it's booked or paid (money-related).
CATEGORY 07

Status, Comms & QA / Monitoring

Agents keep clients and teams updated, watch KPIs and health, and flag anomalies before a human notices — including the monitoring layer that watches every other agent.

SERVICE 20L2 · STANDARD

Weekly internal status digest

Compiles cross-tool metrics into a formatted status report for leadership on schedule.

Before

A PM spends Friday afternoon pulling project completion, closed deals and open tickets into a summary.

→
After

The agent queries each system Friday at 4pm and posts a clean digest to the leadership channel.

Asana / ClickUpHubSpotHelp deskSlack
Impact The Friday status report writes itself.
Internal, read-only — no irreversible action.
SERVICE 21L3 · ADVANCED

Ops health & pipeline monitor

Continuously watches the automations and data pipelines themselves and alerts on failures or anomalies.

Before

You discover a broken connector when a client report ships empty — after the client has already seen it.

→
After

The agent checks that each sync ran and each source returned data, and pages the owner on failure. The QA layer over everything else.

Pipeline monitoringSlackPagerDuty / Email
Impact Broken connectors are caught before a client ever sees an empty report.
Detects and alerts — a human triages. This is the telemetry that makes failures visible early.
How to choose

Start with your highest-impact agent. Grow the portfolio.

Most agencies begin with the one automation that hurts most — usually reporting or onboarding — prove it in production, then add the next. You never buy a level; you subscribe to an operated capability that grows with you.

Start
$900/mo
For your first automation done right.
  • 1 production agent — any level, your highest-impact one
  • Hosting + monitoring with real telemetry
  • Monthly report: processed, flagged, escalated
  • Fixes and adjustments included
Start the conversation
Where most agencies land
Operate
$1,600/mo
For running core ops on agents.
  • Up to 3 production agents across levels
  • A new automation from your hours-lost map each quarter
  • Priority support, direct channel
  • Everything in Start
Start the conversation
Scale
$2,500/mo
For operating on agents end-to-end.
  • An agent portfolio across your operation
  • Continuous build queue
  • Monthly ops-review call
  • Everything in Operate
Start the conversation

Don't see yours? It probably fits.

This catalog is a starting menu, not a limit. Tell us the most repetitive task in your operation and we'll map it to a level, an example and a plan — honestly, including when it isn't worth automating.

Talk to us on Telegram

Prefer email? Write to us  ·  Back to overview