New features
A daily account check, a weekly optimization review, and a quarterly audit are not the same job.
Trying to force all three into one giant “audit” usually produces one giant pile of findings.
HireOtto now separates them by cadence. There is a focused daily operations audit, a deeper weekly optimization audit, and a broader account-health audit. I have also launched the first version of the GA4 MCP server, so you can investigate what happens after the click from the same AI workflow.
At a glance, you can now:
query GA4 properties, reports, and realtime data from your AI assistant
catch Google Ads delivery and pacing issues every day
review search terms, keywords, ads, and visibility each week
run a broader structural account review monthly or quarterly
test the first cross-platform tracking health check across your website, GTM, GA4, and Google Ads
This is the clearest step yet from “Google Ads tools in chat” toward a performance marketing stack that your AI assistant can operate with you.
At a glance
Query GA4 from your AI assistant
HireOtto now has a Google Analytics 4 MCP server.
You can use it to:
list the GA4 accounts and properties available to your Google login
inspect property details, data streams, key events, Google Ads links, and custom definitions
run flexible GA4 reports with the dimensions, metrics, filters, comparisons, and date ranges you need
check realtime activity
find compatible dimensions and metrics before running a report
export larger results to CSV
The server is read-only and still in beta. The core reporting and discovery workflows are working, but I expect the toolset to evolve quickly based on how people use it.
Example prompt:
List the GA4 properties I can access.
For property [PROPERTY_ID], compare the last 30 days with the previous 30 days.
Break down sessions, active users, key events, and total revenue by default channel group.
Show the top rows here and export the complete report to CSV.Add the GA4 server to your AI assistant using:
https://ga4.hireotto.com/mcpAuthenticate the connection, then ask it to list your accounts and properties as a quick check.
One audit rhythm, three different jobs
The original daily ops audit tried to cover pacing and search-term waste together. That was useful, but it mixed fast operational checks with slower optimization decisions.
The new audit system gives each cadence a clearer job.
Daily: catch delivery and pacing exceptions
The daily audit now focuses on what needs attention today:
campaigns that stopped or failed to serve as expected
unusual spend increases or drops against a recent baseline
over- or underspending against daily budgets
pacing across campaigns that share a budget
It deliberately leaves search terms, keywords, ads, locations, and devices to the weekly review, where there is more mature data to work with.
Example prompt:
Run the daily operations audit for Google Ads account [CUSTOMER_ID] using yesterday’s data.
Show all high- and medium-severity findings, explain what changed against the baseline, and give me a review plan. Do not make any changes.Weekly: find optimization opportunities
The weekly audit goes deeper once enough performance data has accumulated.
It can surface review candidates such as:
search terms that may belong in a negative list
converting queries that may deserve their own keyword
underperforming keywords
Manual CPC bid-review candidates
disapproved ads
ads performing below their ad group’s CTR
visibility lost to budget or rank
These are review candidates, not automatic changes. You decide what is genuinely irrelevant, strategically important, or worth testing.
Example prompt:
Run the weekly optimization audit for Google Ads account [CUSTOMER_ID].
Compare the last complete week with the previous week. Group the findings into:
1. act now
2. investigate
3. monitor
Show the evidence behind each recommendation and do not apply changes yet.Monthly or quarterly: review account health
The account-health audit is for broader structural and configuration checks.
It reviews areas such as:
auto-tagging and campaign settings
network, location, and language configuration
keyword count, match-type performance, and Quality Score components
negative-keyword coverage
geographic and device performance
Search impression share lost to budget or rank
optional RSA asset-count and pinning checks
It returns prioritized findings, supporting evidence, CSV exports, and an optional checklist that you can move into a spreadsheet or client deliverable.
Example prompt:
Run the account-health audit for Google Ads account [CUSTOMER_ID] for the last 90 days and compare it with the previous 90 days.
Include the checklist and RSA structure review. Separate tool-generated findings from your own interpretation, and flag any section that could not be completed.The audit tools are read-only and available on the Agency plan. You can review their findings first, then use HireOtto’s Google Ads tools to make only the changes you approve.
Also shipped: keyword-level bid control
For Search campaigns using Manual CPC, you can now set explicit bids when creating ad groups or positive keywords, update the default ad-group bid, and change bids on selected existing keywords.
HireOtto checks the campaign’s bidding setup before applying the change. It will not quietly switch an automated-bidding campaign to Manual CPC just because a prompt included a keyword bid.
Example prompt:
In ad group [ADGROUP_ID], set a max CPC bid of [AMOUNT] on keyword criterion IDs [KEYWORD_IDS].
Confirm the campaign uses Manual CPC first. If it does not, stop and explain the current bidding setup without changing it.What’s next
From separate servers to cross-platform workflows
I have also started testing the first HireOtto skill that works across multiple parts of the stack: a tracking health check.
It reviews a website, the published GTM container, the GA4 property, and the Google Ads account together. It then reconciles the conversion setup across all four layers, flags missing or inconsistent tracking, and produces a prioritized implementation plan.
Once that plan is approved, it can also prepare a review-first GTM import JSON and a QA checklist. It cannot run live Tag Assistant tests or complete signup and payment flows, so those steps remain explicitly marked for human validation.
This is the direction I want HireOtto to take.
The goal is not to build five disconnected chatbots for five ad platforms. It is to let AI workflows inspect the whole performance stack, compare what is happening across platforms, and give you a reviewable plan.
Meta, Microsoft Ads, Reddit Ads, and LinkedIn Ads are all on the roadmap, alongside more cross-platform skills for reporting, audits, tracking, and campaign operations.
If you want to beta test the tracking health check, reply with tracking and I will send you the setup.
Question
One question for you: what would save you more time next – a Meta, Microsoft Ads, Reddit Ads, or LinkedIn Ads server, or a cross-platform workflow that combines several of them?
Reply with the platform or workflow. I am using those answers to decide what moves up the roadmap.
— Suyash
Founder, HireOtto