By December 31, 2027, percent-of-spend will be a minority pricing model for white-label Google Ads fulfillment. Not unpopular. Minority. Most fulfillment deals you sign will be flat per account, because the hours agencies used to resell will no longer be worth reselling.

I managed Google Ads the slow way for the better part of a decade. Hand-built match types. Negative-keyword mining at 1am. Watching Quality Score tank because a client swapped a landing page without telling me. That work took hours, so white-label fulfillment was priced around hours. Wholesale still runs 10 to 15% of spend with a $300 to $500 floor, while retail runs 10 to 20%, with $1,500 to $5,000 a month typical for mid-market. The logic was honest: someone had to do the mechanical work.

Google spent 2026 deleting that logic. Starting September 1, campaigns using auto-created assets or campaign-level broad match were auto-upgraded to AI Max, pulling matching, asset generation, and URL routing closer to one platform-managed workflow. If Google does the matching and much of the building, what exactly is the reseller marking up?

What changed in 2026: Three billable tasks moved into the platform

First, matching changed. AI Max entered open beta on May 6, 2025, claiming 14% more conversions. Google revised that figure to 7% when it went generally available in April 2026. After an August 5 email notice, Google began auto-upgrading eligible Search campaigns to AI Max on September 1. Dynamic Search Ads got a longer runway, until February 2027. For newer placements, the direction is clearer still: Conversational Discovery ads and Highlighted Answers require AI Max for Search or PMax with text customization. Microsoft made a similar move in July 2026, retiring Predictive Matching into its AI Max equivalent.

Then interpretation moved inside the platform. On August 10, Google added agentic capabilities around Ask Advisor: AI Overviews summaries, insight cards, natural-language questions about impression share, and prompt-generated dashboards. That overlaps with the weekly white-label report a junior used to rewrite from screenshots. Google’s 2026 roadmap points in the same direction, describing agentic commerce, sponsored listings, and Direct Offers inside AI Mode, with Gemini 3 and Veo 3 generating creative.

I used to tell clients account structure was the edge. I was wrong. Structure is becoming easier to buy; useful conversion signals and convincing pages are harder to fake. That distinction runs through all five predictions.

Prediction 1: Percent-of-spend pricing becomes the minority by December 2027

The math stops working when the work stops taking the same amount of time. A percentage fee made sense when every extra $10k in spend could mean more queries to mine, bids to adjust, and ad groups to split. Donutz Digital names the conflict: percentage fees rise with spend, rewarding budget increases even when the data says to hold steady. Many contracts also pair a minimum-fee-or-percentage clause with a $1,000 to $5,000 setup fee.

Wholesale puts another squeeze on the reseller. A $300 to $500 floor at 10 to 15% leaves room only if retail pricing supports the markup. The white-label guide warns that gross margin below 35% stops working once you add your own account management. That is before a client asks why its fee rose when its spend rose but the account work did not.

AI Max attacks the old justification. Matching, asset assembly, and URL expansion increasingly happen inside Google. The reseller still needs to judge the output, but it has less reason to charge for every extra dollar flowing through the account. If a $20k-a-month account once took ten hours of mechanical attention, a provider that can reduce those hours can offer a flat price, such as $1,499 with no cut of spend, without pretending the work scaled with budget.

Here is the check: count white-label deals signed in Q4 2027. If more than half still price wholesale as a cut of spend, I missed. I expect most to read as flat per account, with any spend tiers tied to genuine differences in work rather than an automatic raise when the budget goes up.

Prediction 2: Daily action logs become normal; weekly check-ins become a warning

Say you spend $20k a month across three Search campaigns and one PMax build. Under the old model, your white-label partner might log in Tuesday and Friday, pull search terms, add negatives, nudge bids, and send a PDF. I lived that routine and billed it with a straight face. The work was real then.

Now Google can reroute URLs inside AI Max, while Ask Advisor can produce fresh summaries without waiting for someone to assemble a report. The account still needs oversight, but a Tuesday-and-Friday visit is a poor substitute for seeing what changed between visits. If your partner reports weekly activity but cannot show a daily action log, that is the tell.

The agencies that keep clients through 2027 will show dated records: what changed, why it changed, and what happened to CPA. The ones still selling a weekly call and a branded report are selling the part Google automated with prompt-generated dashboards and summaries. A dashboard is not management. It is a tidier way to look at an account that may or may not have been managed.

Pull 20 white-label service agreements at the end of 2027. If most still promise weekly optimization with no daily execution record, I am wrong about the speed of this shift. Until then, ask for the log before you admire the report.

Prediction 3: “AI Google Ads agency” stops earning a premium

Search interest already shows the blur. Queries like ai google ads agency and ai agents for google ads pull impressions but almost no clicks because the label does not tell a buyer what gets done. Private-label searches also turn up vendor listicles repeating familiar names, including The HOTH, Digital Success, SEOReseller, DashClicks, Vendasta, WSI, Boostability, and Clicks Geek. The lists mix productized platforms and embedded partners, often without a checkable position on pricing or automation.

Once AI Max handles more matching and Ask Advisor handles first-pass analysis, every competent partner has access to much of the same foundation. That does not make their work identical. It makes uses AI a weak reason to pay extra. If your pitch in 2027 is still “we use AI,” you are telling the client you have electricity. The badge stops meaning anything; the execution behind it has to earn the fee.

By 2028, I expect agents to be routine in fulfillment offers. Buyers will ask a better question: what does this partner do that Google does not? The answer should involve business context, decisions the platform cannot make for you, and proof tied to closed-won revenue. That is why the traditional agency model is running out of room, not why every agency disappears.

The test is the premium. If standalone “AI Google Ads agencies” can still charge more for the label alone at the end of 2027, I am wrong. Ask a prospective partner what it did yesterday, not which adjective it put before agency.

Prediction 4: Conversion data and landing pages beat bid management

Smart Bidding and AI Max can decide who sees an ad. They still learn from the signals an advertiser sends and the experience a page provides. The 2026 practitioner advice is blunt: clean tracking, Enhanced Conversions and CRM imports using stored GCLIDs, with Closed-Won or Qualified Lead as primary conversions, plus clear offers and pages that answer the customer’s question. The line worth keeping from that piece is: “If it is not on the page, the AI cannot say it.”

I spent years splitting ad groups to fix CPA. That lever has less room to move now. The lever that remains is what the model learns from and where the click lands. Feed it raw form fills when you need qualified pipeline, and it learns the wrong lesson. Send a promising click to a page that does not explain the offer, and the bid cannot rescue it.

That changes what a white-label partner sells. Bid management becomes table stakes because Google bids in every auction. Partners distinguish themselves by getting useful offline conversion data back into Google Ads, defining the primary conversion correctly, and building pages that match the ad’s promise. This is where many retainers fail: they price the report while leaving tracking and pages to the client. I have written more about why the typical retainer is not built for performance.

Pull 20 fulfillment offers in December 2027. If most still headline bid adjustments and negative keywords, with no mention of CRM imports or page builds, I called this shift early. For now, inspect the conversion definition and the landing page before you ask who is touching the bids.

Prediction 5: One named strategist replaces the anonymous bench

When execution runs around the clock, judgment about limits becomes more valuable. Someone has to set the CPA ceiling, decide which budgets cannot be crossed, name the services the business wants to sell, and define a qualified lead. Those are not chores you hand to a bidding system and hope for the best.

My prediction is that a white-label pod serving 20 accounts moves away from three juniors doing routine checks toward one senior setting boundaries and answering for performance. What the deck calls synergy, I call two people doing the same job. One owner who can say no to bad spend beats four checkers who approve it politely. By 2028, agencies will pay for a named strategist, not a bench of routine account checkers.

That is the part of my old job I would keep. The strategist helps decide the offer, approves the landing-page promise, and answers when closed-won revenue misses. Routine bids, budget moves, search-term work, and asset tests should not wait for the next meeting. A managed service pairing agents with human oversight is already built around a flat fee, continuous work within guardrails, and logged actions with reasoning. The human is not there to perform every action. The human is there to own the boundaries and the outcome.

Check the headcount in December 2027. If the average white-label pod still needs two or more full-time operators per 10 client accounts, the automation story stalled and I am wrong. Ask who owns your number by name. “The team” is not a name.

Where I could miss, and who should ignore this

I will name the miss plainly. If Google pauses the AI Max push, or if PMax reporting stays opaque enough that clients refuse to trust it, more manual work remains valuable. Manual structure also keeps a role in three corners: regulated lead gen where claims need legal review, sub-$2k spend accounts where tracking volume may not support learning, and ecommerce catalogs with margins thin enough that one bad URL expansion hurts.

If that is your book, keep the hourly pod and ignore my pricing prediction. For local services, SaaS trials, or Shopify stores spending $10k and up, I expect the mechanical middle to keep shrinking. The test is what your partner ships, not what its sales deck calls it. Search-term sheets and bid-change logs point toward labor. CRM imports, page builds, and a daily action log tied to CPA point toward outcomes.

What I would do with a fulfillment contract now

Do not renew a percent-of-spend wholesale deal on autopilot. Productized platforms offer predictability; embedded partners offer fit. Both typically assume you set retail while the partner fulfills at a lower rate. Protecting that margin gets harder when the fulfillment fee climbs with spend but the work does not.

Ask your current partner four questions this quarter, in writing:

  1. What changed in my accounts each day last week, why, and with what result?
  2. How do Closed-Won or Qualified Lead values reach Google Ads, and how fast?
  3. Who builds the landing page for each ad group, and what does it cost me?
  4. What is my flat per-account price if spend doubles and the work does not?

A partner who answers in logs and numbers stays. A partner who answers in slides goes.

If the answers are vague, test an alternative on one live account. groas runs Search, bidding, budgets, and pages around the clock for a flat $1,499 per service with no cut of spend. It logs actions with reasoning and pairs execution with a named strategist. That is the model these predictions point to: flat pricing, daily execution, and signals and pages treated as the product.

Run your current fulfillment against it for 30 days on CPA and qualified pipeline. Keep the budget and the definition of a good lead the same. Keep the winner. My money is on the system that stopped billing hours. Set a calendar reminder for December 2027 and grade me then. I prefer being checked to being believed.