Use this banner section for site-wide announcements, news updates, big changes and more.
.svg){kind=logo}
When taking over existing PPC c, it surprising how often businesses don't fully understand what their PPC agency has actually been doing with their budget. That's not a communication problem. It's an accountability problem. And it's expensive.
Most founders and marketing leaders are in the same position: you're receiving monthly reports, the agency is confident on calls, and the numbers look plausible. But you have a nagging sense that something isn't right. Lead quality is uneven. The pipeline hasn't moved the way you expected. The report is full of CTR and impression data but light on commercial outcomes.
The issue is that auditing a PPC account properly used to require either deep platform expertise or paying another agency to review the first one. Neither option is appealing. AI has changed that. You can now conduct a structured, rigorous audit of your own Google Ads or LinkedIn account without being a PPC specialist, using a combination of the right data exports and the right prompts. This guide walks you through exactly how to do it.
Before getting into the audit process, it's worth understanding why the standard agency reporting pack is structurally insufficient as an accountability tool.
Most agencies report on the metrics they can most easily look good against. Click-through rate, impressions, cost per click, and conversion volume are all visible, attributable to agency activity, and relatively easy to improve in ways that don't reflect genuine commercial progress. CTR can be inflated by clickbait headlines that fail to convert. Conversion volume can be boosted by counting low-quality form fills that never progress to qualified pipeline. Impression share sounds impressive but tells you nothing about whether the right people are seeing your ads.
The metrics that actually reflect whether PPC is working for a B2B business are: cost per sales-qualified lead, pipeline contribution from paid channels, and closed revenue influenced by PPC. These require CRM integration and a longer reporting window, which many agencies don't set up because it creates clearer accountability for underperformance.
Companies waste an average of 21% of their Google Ads budget on irrelevant clicks, outdated keywords, and poor targeting configurations. In disorganised accounts, wasted spend can reach 30 to 50% of total monthly budget. The agency reporting pack will rarely surface this clearly. An audit will.
You don't need to be a PPC expert to run an effective AI-assisted audit. You do need two things: account access and the right data exports.
Account access is non-negotiable. If your agency has set up campaigns under its own credentials or has not granted you admin access to your own Google Ads account, that is itself a red flag that warrants an immediate conversation. Your account, your data, your history, your ownership. Any agency that doesn't provide full admin access from day one is creating lock-in, intentionally or not. Your ad account history (the performance data, keyword performance, audience insights, conversion tracking setup, and algorithm learnings) is a valuable business asset. You should own it unconditionally.
Once you have access, the data exports you need for a complete AI audit are:
From Google Ads: campaign performance report (last 90 days), search terms report (last 90 days), keyword performance report, conversion actions report, audience performance report, and the account change history.
From your CRM: lead source data showing which deals entered pipeline from paid search, and close rate by lead source if available.
If you're also running LinkedIn Ads, export your campaign performance, audience demographics, and lead gen form submission data.
Most of these exports live under the "Reports" tab in Google Ads. The account change history (under "Tools and Settings" then "Change history") is particularly valuable and almost never included in agency reporting. It shows exactly what the agency has done in your account over the past 90 days, or hasn't done.
There are two ways to run an AI-powered PPC audit, depending on your technical setup.
Option 1: Direct connection via MCP (faster, more accurate)
If you're using Claude, you can connect your Google Ads account directly via MCP (Model Context Protocol). This lets Claude pull live account data and run structured analysis without you having to export anything manually. A comprehensive Google Ads audit via MCP takes roughly 45 to 60 minutes and covers conversion tracking, budget allocation, keyword performance, Quality Scores, search term analysis, and a prioritised action list. This is the most efficient approach and the one that catches the issues manual exports can miss.
Option 2: Manual data exports pasted into AI (still very effective)
If you're not set up for MCP, export the reports listed above as CSVs, paste them into Claude or ChatGPT, and work through the audit sections below with the prompts provided. This takes two to three hours but is entirely achievable without technical setup. Use Claude for the most structured analysis, particularly for larger data sets where context window length matters.
Either way, the audit structure is the same. Work through each section in order. Don't skip conversion tracking.
This is always the first check, because everything downstream depends on it. If conversion tracking is broken, inaccurate, or measuring the wrong events, every bidding decision Google's algorithm makes is based on bad data. Smart Bidding learns from the signals you give it. Bad signals produce bad outcomes.
What to look for:
Compare conversion counts in Google Ads against your CRM for the same period. If the discrepancy exceeds 10%, something is wrong. Either conversions are double-counted (a very common issue, particularly where both a Google tag and GA4 are firing on the same event), conversions are being recorded that don't represent real leads (page views, button clicks, or scroll events set up as conversions), or conversions are being missed entirely.
Check which conversion actions are set to "Include in conversions" in the campaign bid settings. If phone calls, chat interactions, or newsletter sign-ups are included alongside demo requests and actual lead forms, the bidding algorithm is optimising for a mixed signal that includes low-quality actions alongside high-value ones.
The AI prompt to use:
Paste your conversion actions report and CRM lead data, then ask: "Review these conversion actions and compare them to my CRM lead volumes for the same period. Identify any discrepancies above 10%, flag any conversion actions that appear to be low-quality signals (page views, scroll events, newsletter sign-ups), and tell me whether the current conversion setup would cause Google's Smart Bidding to optimise towards the wrong outcomes."
This is where wasted spend hides most visibly. Your agency bids on keywords, but those keywords trigger actual searches that may be completely irrelevant to your business. The search terms report shows you what people were actually searching when your ads appeared.
What to look for:
Open your search terms report and sort by spend. Look for: searches that are clearly irrelevant to your product (competitor brand names you're accidentally targeting, job-seekers searching for roles in your category, students researching the topic for an essay), searches with high spend and zero conversions over a 90-day window, and searches that suggest the wrong audience type (free, DIY, small business terms if you're targeting enterprise, for example).
Then check your negative keyword lists. For a B2B account that has been running for more than six months, a healthy negative keyword list should contain hundreds of entries. A list with fewer than 50 entries suggests the agency hasn't been doing the basic hygiene work that prevents budget waste on irrelevant clicks.
Also check the account change history for how frequently negative keywords have been added. If the agency hasn't touched the negative keyword lists in 60 or more days, budget has been leaking to irrelevant searches the whole time.
The AI prompt to use:
Paste your search terms report and ask: "Analyse this search terms data. Identify the top 20 irrelevant or low-intent searches by spend that should be added as negative keywords. Flag any patterns suggesting systematic budget waste (job-seeker queries, competitor names, free/DIY terms). Estimate total wasted spend if possible based on cost data."
A disorganised account is a profit leak. Poor structure makes budget allocation inefficient, blurs performance data, and suppresses Quality Scores.
What to look for:
Check whether brand, competitor, and non-brand campaigns are properly separated. Mixing branded keywords (searches for your company name) with non-brand keywords in the same campaign makes it impossible to assess true demand generation performance. Brand campaigns typically have much higher conversion rates, which artificially inflates the blended numbers and makes non-brand performance look better than it is.
Look at budget allocation across campaigns. Is the majority of spend on branded terms rather than demand generation keywords? Over-reliance on branded keywords can produce a good-looking ROAS while failing to attract any new customers. Branded traffic cost share can range from 16% to 56% across accounts; make sure the split aligns with your actual growth objectives.
Check whether Performance Max campaigns are running alongside Search campaigns with no brand exclusions. Without brand exclusions on PMax, it will cannibalise branded search traffic and make it appear to be generating new demand when it isn't.
The AI prompt to use:
Paste your campaign performance report and ask: "Review this campaign structure. Identify whether brand and non-brand campaigns are properly separated. Calculate what percentage of total spend is going to branded versus non-branded terms. Flag any budget allocation decisions that look like they're prioritising easy wins (brand, remarketing) over demand generation. Identify any campaigns that appear to be cannibalising each other."
Quality Score is Google's rating of your ad relevance, and it directly affects how much you pay per click. A Quality Score of 7 or above is healthy. Scores of 4 or below mean you're paying a penalty on every click in that ad group, sometimes 20 to 40% more per click than a well-optimised competitor.
What to look for:
Pull your keyword performance report and filter for keywords with Quality Scores of 4 or below. Calculate the total spend on these keywords over the past 90 days. This is the cost of poor ad relevance, and it represents directly recoverable budget.
Check ad copy for message match with landing pages. The most common cause of poor Quality Score in B2B accounts is ad copy that makes a specific promise (for example, "CRM for construction teams") but sends traffic to a generic product homepage that doesn't reflect the same message.
The AI prompt to use:
Paste your keyword performance data and ask: "Identify all keywords with Quality Score 6 or below. Estimate the CPC premium being paid on these keywords versus a Quality Score of 7+ benchmark. Calculate the approximate monthly over-spend from poor Quality Scores. Group low Quality Score keywords by likely cause (poor ad relevance, poor landing page experience, or low expected CTR) based on the available data."
This is the audit section agencies are least comfortable with, and the most revealing. The change history log shows every modification made to your account, with dates and details. It is the most objective measure of how actively your account is being managed.
What to look for:
Download the change history for the past 90 days. Look at the frequency and nature of changes. A well-managed B2B PPC account should show: weekly search term reviews with negative keyword additions, regular ad copy testing with at least one active test per campaign, bid strategy adjustments in response to performance data, and audience refinements as data accumulates.
What raises concerns: no changes for stretches of three or more weeks, changes that are exclusively automated (Smart Bidding adjustments that the agency didn't initiate), a flurry of activity immediately before monthly reporting periods followed by inactivity, and no evidence of landing page or ad copy testing.
Some agencies effectively put accounts on autopilot after the initial setup, collecting the retainer while Smart Bidding runs on its own. The account may not visibly deteriorate in the short term, but it won't improve either, and it will fall behind as competitors actively optimise.
The AI prompt to use:
Paste your change history export and ask: "Analyse this PPC account change history. Calculate the average frequency of changes per week. Identify any gaps of 14 or more days with no changes. Categorise changes by type (negative keywords, bid adjustments, ad copy, audience, structural). Based on this activity pattern, assess whether the account appears to be actively managed or primarily running on automation. Highlight any patterns that suggest reactive rather than proactive management."
For B2B companies, this is the section that connects PPC spend to commercial outcomes. It requires CRM data alongside your Google Ads data, but it's the only check that tells you whether PPC is actually contributing to business results.
What to look for:
From your CRM, pull the source data for leads that became sales-qualified in the past 90 days. Cross-reference with your Google Ads conversion data for the same period. Calculate your actual cost per SQL from paid search, not cost per form fill.
If your agency is reporting on cost per lead (CPL) but you've never had a conversation with them about cost per SQL or pipeline contribution, that is the conversation to have next. An agency optimising for CPL without visibility into lead quality will systematically optimise for volume over commercial relevance.
For B2B SaaS companies running sales-led models in particular, the gap between a marketing-qualified lead and a sales-qualified opportunity is where most PPC ROI gets lost or hidden. The audit should reveal whether your agency has any visibility into what happens after the form fill.
The AI prompt to use:
Paste both your Google Ads conversion data and your CRM pipeline source data, then ask: "Compare these paid search conversion figures to the CRM pipeline data for the same period. Calculate cost per SQL from paid search. Identify any significant discrepancy between the number of leads Google Ads reports as conversions and the number that appear as qualified opportunities in the CRM. If possible, calculate what proportion of total PPC spend is producing pipeline versus producing unqualified form fills."
Running the audit produces a prioritised list of issues. Before you act on it, it's worth separating findings into three categories.
Immediate fixes are issues that are currently costing money and can be resolved in a day: broken conversion tracking, obvious irrelevant search terms, campaigns with no negative keyword lists, Performance Max cannibalising brand. These should be fixed regardless of what you decide to do about the agency relationship.
Structural issues are problems with account architecture that require a more considered plan: campaign structure that conflates brand and non-brand performance, bidding strategies feeding on the wrong conversion signals, Quality Score problems that require ad copy and landing page work. These take weeks to address properly.
Strategic misalignment is when the agency is technically competent but optimising for the wrong outcomes. If every campaign is set up to maximise lead volume and the agency has never discussed cost per SQL or pipeline contribution, the issue isn't the account settings; it's the brief. This is the conversation to have with your agency before considering a change.
Present the findings to your agency as a collaborative document, not an accusation. Ask them to explain what they see in the data and what their plan is to address each issue. A good agency will engage with the detail honestly and produce a clear action plan. An agency that becomes defensive, dismisses the findings, or can't explain what the data shows is confirming the concern.
If you're considering making a change, the guide to choosing a B2B technology marketing agency covers how to evaluate specialists, structure the selection process, and avoid the mistakes that result in the same problems recurring with a different provider.
To summarise, these are the most useful prompts to run across a full AI-assisted account audit. Use them in order, with the relevant data exports pasted before each one.
Conversion tracking check: "Review these conversion actions and CRM lead volumes. Flag discrepancies over 10%, identify low-quality conversion signals, and assess whether the setup would cause Smart Bidding to optimise towards the wrong outcomes."
Wasted spend check: "Analyse this search terms data. Identify the top irrelevant searches by spend, flag patterns of systematic budget waste, and estimate total wasted spend over the period."
Structure and budget check: "Review this campaign structure. Identify brand versus non-brand budget split, flag any campaigns that appear to be cannibalising each other, and assess whether spend allocation aligns with demand generation objectives."
Quality Score check: "Identify all keywords with Quality Score 6 or below. Estimate the CPC premium being paid and calculate the approximate monthly over-spend from poor ad relevance."
Activity audit: "Analyse this change history. Calculate average weekly change frequency, identify inactivity gaps, categorise changes by type, and assess whether the account shows active or passive management."
Pipeline attribution: "Compare these conversion figures to the CRM pipeline data. Calculate cost per SQL, identify the discrepancy between reported conversions and qualified pipeline, and assess what proportion of spend is producing commercial outcomes."
Run all six, compile the outputs into a single document, and you have a complete picture of account health that would previously have required either deep PPC expertise or an external audit engagement. The whole process, with data exports and AI analysis, takes a working day. Most accounts will surface at least one significant issue. Many will surface several.
The goal isn't to catch your agency out. It's to have a better-informed conversation about what is working, what isn't, and what needs to change. That conversation is easier, faster, and more productive when you arrive at it with data rather than instinct.
Lever Digital is proud to be a 2026 UK Paid Media Awards finalist, recognised for outstanding performance-led paid media campaigns across B2B and SaaS.
.png)
