Benjamin Christie runs Gourmet Ads, a digital advertising business that helps food brands reach household grocery buyers and home cooks online.
The company has been around for 18 years. Its advertising customers include supermarkets, food and beverage brands, and global advertising agencies. The engineering and product teams are small, which means every operational idea competes with product work, client work, reporting work, and the thousand small jobs that come with running an established advertising business.
That is why Benjamin pays attention to any workflow that helps the team see what needs attention without adding more reporting work.
The reporting problem Zapier already solved
Since 2018, Zapier has helped Gourmet Ads solve one of the least glamorous problems in advertising: daily reporting for Prebid partners.
Some partners provide real-time reporting through APIs. Others send daily reports by email. The formats are inconsistent: CSV files, Excel sheets, and tables in the body of a message. Of about 30 advertising partnerships, 12 use Zapier workflows to turn emailed reports into usable data.
When a report arrives, Zapier picks it up, identifies the format, converts it to Gourmet Ads' common reporting format, and saves it to Google Drive. From there, a cron job maps the data into their reporting system every 12-24 hours.
That workflow gives the leadership and publisher teams a daily view of sell-through and partner discrepancies. In advertising, even the timing of a report can make the numbers look wrong. A New York partner reporting in GMT time can look 30% off mid-month, then settle within a couple points by the end of the month.
Without that automated feed, the ad operations team would wait days to understand whether a margin issue was real or just a reporting discrepancy.
That was the old version of the work: get the data where it needed to go.
The reusable MCP pattern
Gourmet Ads started using Claude to build internal reporting for sales and marketing teams. But the team still needed a better way to connect data from different systems for analysis.
Zapier MCP changed the question from "how do we build this?" to "what can we build next?"
Salesforce stays the sales system of record. Atlassian Confluence becomes the team-facing operating surface. Zapier MCP connects Salesforce, Google Analytics, email parsing, and other tools into reports the team can actually use.
The result is a short list of recommended actions, not another dashboard to interpret.
The weekly report that changed the rhythm
The clearest example started with a missing weekly marketing report.
Gourmet Ads Marketing Manager, Justin Mosgrober, previously spent about two hours pulling together a weekly report. Benjamin already talked to Justin four or five times a week, so the report rarely surfaced anything surprising. Then Justin went on three weeks' leave.
"I thought, well, maybe I can build a report that's amazing," Benjamin said.
The first version analyzed media kit requests, one of Gourmet Ads' main top-of-funnel signals. Agencies and brands often begin by asking for a media kit, so Benjamin wanted to understand where those requests came from and what happened next.
The report now lives in Atlassian Confluence so the whole team can see it. It pulls from Google Analytics, direct partnerships, beta tools Benjamin is testing, Salesforce signals, and other business systems. What started as conversion reporting has grown into a six-part weekly operating report.
The most important prompt is simple: give the team 5 action items for next week, and make sure each one takes less than 20 minutes to complete.
"We are now looking forward to this report each week," Benjamin said. "Instead of it just taking up hours of Justin's Friday, now it has so many gems we can use."
The findings that made the report matter
One week, the report found that 6% of site traffic appeared to be fake. That meant 6% of the analytics data underneath the report was suspect.
"We immediately excluded the domain from our reporting and our web server," Benjamin said. "The report helped unearth that."
Another week, the report found that a key media kit link in everyone's email footer pointed to a page that had been removed 4 years earlier.
"Everybody's had it for years," Benjamin said. "It just went to an error page. Two-minute fix."
That kind of issue hides in plain sight. The link was not buried in an obscure workflow. It sat in the email footer people used every day. A manual report could have missed it indefinitely because nobody was asking the system to look across the business and point to small fixes with outsized impact.
The report also flagged that Healthy Ads, a Gourmet Ads sub-brand, was publishing less content than Gourmet Ads. Organic search is one of the company's most important business channels, and the marketing team already uses tools to identify content gaps. The AI report points the team toward the places where attention is most likely to pay off.
Why the action list works
For Benjamin, the value is not that the report is automated. The value is that the report finds work worth doing.
The action items are intentionally small. A broken link can be fixed in 2 minutes. A fake traffic source can be investigated. A content gap can become next week's assignment.
The report gives the marketing team a short list of impactful fixes they can make, not a strategy deck to interpret.
The same pattern for sales
That same pattern is moving into sales.
Benjamin showed his team a daily sales board. At 7:30 a.m. Eastern, it gives a salesperson a snapshot of the day inside Atlassian: who needs a call, which records are missing key data, which leads have fallen quiet, and which Salesforce records need attention.
One example showed a lead that had gone 35 days without engagement. The name linked back into Salesforce, turning the report from a readout into a work queue.
"Our daily sales report is everything Salesforce should be but isn't," Benjamin said.
The architecture is straightforward: Zapier MCP reads from Salesforce and surfaces the daily board in Atlassian Confluence. Google Analytics MCP pulls in analytics data. Zapier Tables are not part of this workflow.
That matters because Benjamin is not trying to replace Salesforce, Google Analytics, Atlassian, or the tools his team already uses. He is connecting them so the team can see where to look, what changed, and what needs attention.
What changes for a small team
Benjamin said work he can now test from his desk would normally be "a three-day engineering project." His engineers still matter. But MCP gives Benjamin, as President and product lead, a way to test ideas before asking engineers to spend days building them.
One example is Gourmet Ads' "Next Email" project.
About 18 months ago, the engineering team built a simple tool that let sales staff click a Salesforce URL and generate a next email from a hosted page powered by the ChatGPT API. It worked, but it never became a "wow" experience.
When Claude launched, the team built V2. The new version is powered by the Claude API and uses more known data, including examples of good and bad emails Gourmet Ads has already written.
"The new version came to life," Benjamin said.
His personal output is up too.
"My personal output is probably 30 to 40% more," Benjamin said. "And we still haven't got a lot pulled in."
Gourmet Ads is still adding key metrics to the weekly report, and Benjamin plans to build more daily reporting for other roles, including finance.
That is the real shift. Every week, the business tells the team what needs attention. Every week, they get 5 recommendations they can act on quickly. And every time one of those recommendations catches something the team would have missed, the report gets harder to ignore.
Try the weekly growth report pattern
Gourmet Ads' workflow can be generalized into a reusable MCP pattern: connect your source-of-truth CRM to the place your team already reviews work, then ask your agent for a short list of actions small enough to complete this week. Get the skill in the Zapier GitHub.
The pattern:
Input: CRM records, analytics data, email/request signals, and any team-facing reporting surface.
AI/MCP workflow: Read the source systems, identify anomalies or stalled work, rank the highest-impact fixes, and write the results into a shared workspace.
Output: A weekly or daily report with 5 recommended actions, each scoped to a small fix.
The important part is not the report format. It is the constraint: the recommendations have to be specific enough that someone can act on them.
At a glance: Salesforce stays the system of record. Atlassian Confluence becomes the work surface. Zapier MCP connects the tools and turns the report into a work queue.
Install Zapier MCP · Browse MCP templates
Benjamin Christie is President of Gourmet Ads, a digital advertising business helping food brands reach household grocery buyers and home cooks online.
