Enterprise DNA

Omni by Enterprise DNA

Enterprise DNA Resources

Step-by-step how-tos. Practical AI operating-system thinking for owners, operators, and teams doing real work.

220k+

Data professionals

Omni

AI agents and apps

Audit

Map the manual work

Guide Intermediate Omni Ops

Automate Media Buying Reconciliation

See how agencies can match insertion orders, platform spend, and invoices with AI to catch billing errors before month-end closes.

Sam McKay |
Automate Media Buying Reconciliation

Month-end reconciliation is where media margin disappears

Media buying reconciliation looks simple from a distance.

You have an approved media plan. You issue insertion orders. Platforms spend money. Vendors send invoices. Finance needs to know that billed spend matches approved spend and actual delivery.

Then month-end arrives.

A media buyer exports spend from Google Ads, Meta, LinkedIn, DV360, The Trade Desk, TikTok, or whichever mix a client is running. Someone opens the insertion order, usually as a PDF or spreadsheet. Someone else finds the vendor invoice in an inbox or accounts payable system. They compare dates, campaign names, platform fees, management fees, taxes, credits, makegoods, and pacing adjustments.

Most agencies do this through a patchwork of spreadsheets, folders, email threads, Slack messages, and memory.

The work gets harder as the agency grows. A $1M agency may have a manageable number of media accounts, but a $10M agency can be reconciling dozens of client campaigns across hundreds of line items each month. The operating problem is not that your team doesn’t know media. It is that skilled people are being asked to repeat a checking process that software should own.

For marketing and creative agencies, we often see annual leakage in the $60K to $180K range from operational gaps. Media reconciliation is rarely the only cause, but it is a regular source of avoidable write-offs, missed vendor credits, unrecovered overspend, and unbilled fees.

If a buyer finds a discrepancy after the client invoice has gone out, the agency usually absorbs it. If they do not find it at all, it may become a quiet margin loss that never gets discussed.

The aim is not to remove financial review. The aim is to automate the evidence gathering, matching, exception detection, and follow-up work so your team reviews the few items that need judgment.

You can see the wider operating model in Omni for marketing and creative agencies. For this guide, I want to stay focused on how to automate media buying reconciliation from insertion order through to final invoice approval.

What your team is manually reconciling today

Most agency leaders underestimate the number of comparisons happening inside a normal reconciliation cycle.

Take one paid media campaign with a monthly budget of $75,000. The client-approved plan may include six channels, separate monthly allocations, audience notes, flight dates, agency fees, and specific reporting requirements. The work then branches from that original approval.

Your team may need to compare:

  • Client-approved budget against the insertion order
  • Insertion order amounts against vendor invoices
  • Planned flight dates against actual platform delivery dates
  • Planned spend against actual spend by campaign, ad set, or line item
  • Vendor fees against contracted rates
  • Agency management fees against the commercial agreement
  • Credits and makegoods against the original discrepancy
  • Platform currency and tax treatment against finance records
  • Actual spend against the amount your agency intends to bill the client

That comparison is rarely clean. Campaign naming changes halfway through the month. A platform reports spend in a different timezone from the invoice. A vendor bills a service fee that was not included in the original sheet. A campaign may have been paused because the client changed creative. Finance sees an invoice total that does not match the account manager’s client billing schedule.

Each issue might take 10 minutes to resolve. Multiply that across a portfolio and it becomes a serious operating load.

Account managers often get pulled into the process because they hold the client context. They are already spending a large part of their week producing reports, updating decks, and answering status questions. In many agencies, AMs spend around 30% to 50% of their time on reporting and client communication. Adding invoice detective work turns their highest-value people into workflow glue.

The consequence is not only slower month-end close. It affects client trust.

When a client sees a billing correction two months later, they do not experience it as an internal admin problem. They experience it as uncertainty around money. Your agency may have done good media work, but the commercial relationship takes a hit.

The three-way match that should run automatically

At its core, media buying reconciliation needs a reliable three-way match:

  1. Approved commitment: the client-approved media plan and insertion order.
  2. Actual delivery: spend and delivery data from media platforms or vendors.
  3. Billed amount: vendor invoices, platform statements, and the agency’s client billing record.

An AI-enabled reconciliation workflow collects those documents and data points into one controlled process. It does not just put files in a folder. It reads the information, aligns the records, applies agreed rules, and flags exceptions for a person to approve.

That distinction matters.

A dashboard that shows media spend can be useful. It will not tell you that the invoice includes a 12% platform fee when the insertion order allowed 10%. Nor will it connect a $3,400 underdelivery to a campaign pause approved in an email thread.

A proper reconciliation workflow needs both structured data and agency context.

The structured data includes dates, amounts, line items, purchase order numbers, currencies, fees, impressions, clicks, and billing entities. The context includes naming conventions, negotiated vendor terms, client approvals, known campaign changes, and your own margin rules.

This is the sort of back-office operating work that Omni Ops is designed to handle. The agent carries the repetitive process. Your people make the calls where commercial judgment is required.

How an AI reconciliation agent works end to end

The best way to understand this is to walk through a monthly cycle.

1. It collects source documents without chasing people

The workflow begins by pulling source material from the places it already lives.

That may include:

  • Approved media plans from Google Drive, SharePoint, or your project system
  • Insertion orders from a document folder or email inbox
  • Actual spend from ad platform APIs and scheduled exports
  • Vendor invoices from accounts payable, email, or accounting software
  • Client billing data from your finance system
  • Campaign changes and approvals from email, Slack, or a project management tool

The agent assigns every record to the right client, campaign, period, and vendor. It does not need every source to be perfectly standard on day one, but it does need a practical mapping layer.

For example, “Acme Q3 Awareness”, “ACME_AWR_Q3_2026”, and “Acme Brand Campaign July-Sep” may all refer to the same campaign. A good workflow learns the approved naming relationships and asks for confirmation when confidence is low.

This is an important implementation point. You don’t get value by forcing a full systems replacement. You get value by connecting the workflow around the systems your agency already uses.

2. It extracts the facts from invoices and insertion orders

Invoices and insertion orders are often semi-structured documents. Some are PDFs, some are spreadsheets, some are portal downloads, and some are simply an email with a number in the body.

The agent extracts fields such as:

  • Vendor name and billing entity
  • Invoice number and invoice period
  • Insertion order number
  • Campaign and line-item descriptions
  • Net media cost
  • Platform, technology, or data fees
  • Tax and currency
  • Planned and actual flight dates
  • Credits, makegoods, and adjustments
  • Payment due date

It then compares those fields against the approved source of truth.

This is where automation removes a lot of manual copying. Your media coordinator should not need to rekey 40 invoice totals into a tracker, then search through PDFs to understand why three totals look different.

The agent creates a reconciliation record for each campaign and line item. It preserves links to the source documents so your finance lead or media director can inspect the evidence when necessary.

3. It matches actual platform spend to the approved plan

The next job is comparing delivery to plan.

The agent pulls actual spend at the right level of detail. For one vendor that may be by campaign. For another, it may need to be by insertion order or line item. It then calculates variances against agreed tolerances.

An agency might decide that a spend variance below 1% can be automatically marked as matched, while anything above $500 or 2% needs review. The right threshold depends on your average spend, client agreements, and the cost of reviewing exceptions.

The agent should not blindly mark every difference as a mistake. It classifies the discrepancy.

Common classifications include:

  • Underspend due to campaign pause
  • Overspend due to pacing or budget expansion
  • Invoice fee above contracted rate
  • Missing credit from a prior-period adjustment
  • Spend posted outside the campaign flight
  • Duplicate invoice risk
  • Currency conversion difference
  • Invoice received before platform delivery data is complete
  • Client-approved budget change not reflected in the insertion order

That classification saves time because the reviewer starts with an explanation and evidence, not a blank spreadsheet cell.

What a useful exception report looks like

A useful reconciliation output is short enough for a media lead to act on.

It should show the client, campaign, vendor, planned amount, actual spend, invoice amount, variance, likely cause, confidence level, and recommended next action.

For example:

ClientVendorPlannedActualInvoiceVarianceRecommended action
Retail Client AMeta$18,000$17,460$18,000$540Check unspent balance and client billing treatment
B2B Client BLinkedIn$9,500$9,830$10,190$360Review invoice fee against approved 3.5% rate
Tourism Client CProgrammatic Vendor$25,000$22,900$25,000$2,100Request makegood or credit confirmation

The key is that the report separates matched items from exceptions. If 85 of 100 line items reconcile cleanly, your team should spend their time on the 15 that do not.

That is a different month-end experience.

Instead of sending messages like, “Can someone check the LinkedIn amount?”, your team receives a draft exception summary with direct links to the insertion order, spend export, invoice, and relevant approval trail.

The same workflow can produce a client-ready reconciliation note. It should not send client communications automatically without review, but it can draft the explanation and give the account manager a clean starting point.

This is where the Reporting Agent can support the finance workflow. It pulls performance data across connected platforms and drafts the monthly report and AM email summary. When spend and billing are already reconciled, the reporting pack is less likely to contain numbers that need correcting after the client sees them.

Where agencies lose money in the process

The obvious loss is an overbilling error that the agency fails to dispute. But that is only one category.

Here are the margin leaks we tend to see.

Unclaimed credits and makegoods

A vendor may agree to a credit for underdelivery, a placement issue, or a technical error. If the credit is not tied to the original campaign and tracked through to the next invoice, it disappears into email history.

A reconciliation agent can keep an open-items register. If a promised $1,800 credit does not appear on the next invoice, it flags the issue before payment approval.

Fees that drift from agreed rates

Agency teams negotiate rates, then campaigns move quickly. A fee that looks modest on one invoice can become expensive over a year if it is applied inconsistently across multiple accounts.

The agent can compare billed fees against your contracted fee table. That is a rules-based check. It should happen every month, not only when a finance manager happens to notice a pattern.

Agency fees not billed correctly

The reverse problem also matters. Your agency may have earned a management fee, trafficking charge, or reporting fee that was never captured in client billing because the actual media activity changed during the month.

This is a common problem when account teams manage six to 10 accounts each. They keep client delivery moving, but commercial administration trails behind. Headcount becomes the only scaling lever because every additional account creates more manual checking.

The Account Health Agent helps on the client side by watching accounts daily, flagging risks and opportunities, and drafting next-step messages. Paired with reconciliation, it can highlight accounts where spend, delivery, scope, or margin are moving in the wrong direction before month-end.

Late discovery of overspend

By the time a spreadsheet check catches an overspend, the money is already spent. Daily or weekly variance monitoring gives the media lead a chance to intervene while the campaign is still live.

You do not need to turn every minor variance into an alert. That creates noise. You need rules that identify the differences that could affect client budget, agency margin, or vendor payment approval.

Build the process around controls, not just automation

There is a temptation to think AI reconciliation means allowing an agent to approve vendor invoices and send client adjustments without anyone looking.

That is not the right design for most agencies.

The right design is controlled automation. The agent does the repetitive work. A named owner approves exceptions above defined thresholds.

A practical control model could look like this:

  • Auto-match items within a set amount and percentage tolerance
  • Require media lead review for material spend variances
  • Require finance approval for invoice payment exceptions
  • Require account manager review for any client-facing billing change
  • Maintain an audit trail with source documents and decision history
  • Review the matching rules quarterly as vendors, clients, and pricing change

Your controls should match the size and complexity of the agency. A $2M agency does not need a heavyweight enterprise finance program. It does need a dependable process that does not rely on one person remembering where every approval lives.

If your documentation is inconsistent, start there. Standardize the minimum fields on every media plan and insertion order. Client name, campaign ID, vendor, flight dates, committed spend, fee structure, billing currency, and approval reference will get you a long way.

For more operating ideas beyond media reconciliation, the EDNA guides library is a useful place to find workflows that can be applied across account management, reporting, and delivery.

What to automate first

Do not try to automate every channel, vendor, and edge case in the first month.

Start with the area where manual effort and financial exposure are both high. For many agencies, that means your three biggest media-spend clients or the vendors producing the most invoice exceptions.

A sensible first phase might cover:

  1. One client group with repeat monthly media activity.
  2. Two or three high-volume platforms.
  3. Approved insertion orders and vendor invoices.
  4. Actual spend matching at campaign level.
  5. A weekly exception report and month-end approval workflow.

Run it alongside your current process for one or two closes. Compare what the agent catches against what your team finds manually. Adjust naming rules, variance thresholds, and exception categories.

Then expand to more vendors and more detailed line-item matching.

The objective is not to create another dashboard that needs managing. The objective is to reduce the number of hours your team spends assembling evidence and increase the speed at which they resolve real commercial issues.

If you want to identify the best first workflow in your own agency, Book a 60-min Omni Audit. We use the session to map the current process, identify the highest-value automation opportunity, and outline what needs to connect. You get three practical outputs in 60 minutes, with no deck.

The wider agency opportunity

Media reconciliation is an excellent place to begin because the work has clear inputs, repeatable rules, and a direct connection to cash and margin.

But it also exposes a broader issue. Agencies often have capable people doing work that should be carried by operating systems.

Your content team gets hit with rising asset volume and rising per-piece production cost. Your account managers spend too much time assembling updates. Your media team spends days each month matching files that should already be connected.

The Content Production Agent addresses another part of that model. It produces first-pass content from briefs in the right brand and format, so the team edits instead of starting from a blank page. The Reporting Agent supports monthly client communication. The Account Health Agent keeps account risks visible between meetings.

Each workflow gives your people more room for decisions that clients will actually pay for.

The agency that improves reconciliation does not suddenly need fewer media experts. It needs fewer experts spending Friday afternoon searching for an invoice, reopening last month’s workbook, and trying to remember why a number changed.

That matters when you are trying to grow without simply adding account managers, coordinators, and finance support at the same pace as revenue.

Start with the leak you can measure

If you are reviewing media invoices manually, you can measure the opportunity quickly.

Look at the last three months and ask:

  • How many hours did the team spend collecting data and checking invoices?
  • How many invoice discrepancies were found after payment or client billing?
  • What credits or makegoods remained open?
  • How much spend could not be matched cleanly to an approved insertion order?
  • How often did an account manager need to explain a number after the report went out?
  • Which clients create the most exception work relative to their fee income?

Those answers give you the real baseline. They are more useful than a generic automation promise.

From there, the right automation design is usually clear. Connect the sources, create a consistent campaign and vendor mapping, define materiality thresholds, automate the three-way match, and put people in charge of exceptions rather than spreadsheet assembly.

To see where reconciliation fits alongside the rest of your agency operations, review the AI audit for marketing and creative agencies. If you want a working plan for your specific accounts, systems, and margin pressures, Book my Omni Audit.