Welcome to the Sellforte MCP Use Case Library. Each post takes one question that marketing teams really ask their AI assistant through Sellforte MCP, which is now in general availability and accessible to all marketers worldwide. It strips out anything customer-specific and walks through what the assistant does with it, step by step. We run every question on our demo account so the numbers you see are illustrative, not anyone's real results. The idea is simple: steal the prompts, try them on your own model, and see what you get back. We hope you find this series inspiring!
-Juha, CEO at Sellforte
Sooner or later someone needs a list of your ad accounts: an ad platform asking which accounts to include in a measurement study, a data provider setting up a new feed, or an agency asking for access. The IDs are scattered across platform interfaces, and the person who knows them all is usually on holiday. If your accounts already flow into Sellforte, you can just ask for the list.
List the ad platforms, countries and ad account IDs in my Sellforte data that had spend between 2026-05-01 and 2026-10-31, one row per account, so I can prepare a data access request.
Which channels inside those accounts had no spend in October 2026?
The assistant first checks which datasets have an account dimension. In Sellforte, that's the campaign-level performance data: it carries the ad platform, the ad account name and the account ID. The MMM data, on the other hand, knows platforms and countries, so the assistant uses it to confirm that both lists cover the same markets.
It then groups spend by ad platform and ad account for the six months you asked about. On our demo account that came back as 22 rows across 13 ad platforms, two countries (US and Germany) and $7 446 000 of spend. Here's the digital part of the list. We've shortened the account IDs to their last four digits for this post; the assistant returns them in full.
| Ad platform | Ad account | Account ID | Spend 2026-05-01 to 2026-10-31 |
|---|---|---|---|
| Google Ads | E-com DE | …0534 | $1 933 352 |
| Google Ads | E-com US | …3292 | $858 962 |
| Meta Ads | E-com DE | …7562 | $774 623 |
| Meta Ads | E-com US | …7639 | $549 989 |
| Adtraction | E-com DE | …9156 | $224 071 |
| TikTok Ads | E-com DE | …1365 | $134 145 |
| Adtraction | E-com US | …4910 | $98 591 |
| TikTok Ads | E-com US | …4545 | $72 222 |
| Criteo | E-com DE | …2490 | $42 101 |
| Microsoft Advertising | E-com DE | …0175 | $23 505 |
| Awin | E-com DE | …9301 | $23 156 |
| Criteo | E-com US | …3615 | $18 525 |
| Microsoft Advertising | E-com US | …3826 | $10 342 |
| Awin | E-com US | …7165 | $10 189 |
| PriceRunner | E-com DE | …4651 | $4 693 |
| PriceRunner | E-com US | …7465 | $2 065 |
| Pinterest Ads | E-com DE | …2015 | $1 863 |
| Pinterest Ads | E-com US | …7265 | $820 |
Illustrative demo data, spend rounded to the dollar.
That's 18 accounts on nine digital platforms, one per country each. The other four platforms, TV, direct mail, email and print, have no ad account at all, so the assistant lists them separately with "N/A" as the account ID. They still matter for the request, because they're part of the spend the model explains: direct mail ($922 857), TV ($885 333), email ($846 493) and print ($8 102) in the same window.
The follow-up is where it gets useful. A data request for an account that went quiet months ago is wasted effort, so the assistant breaks October 2026 down by account and channel. On the demo data, every account was still active, but several channels inside them weren't:
Total October spend across all platforms was $1 071 131 (illustrative demo data).
This one isn't about insight. It's about not wasting a week on admin. Data and access requests stall when the list is incomplete or out of date: a missing account means a second round of emails, and an old account means someone sets up a feed nobody needs.
Pulling the list from the same data your model runs on also gives you a quiet consistency check. If an account you expected isn't in the list, either it had no spend in the window or it isn't connected, and both are worth knowing before you send anything out. One thing to keep in mind: an account with no spend in the window you pick won't show up, so choose a window that matches the request.
Want to see this on your own data? Book a demo, or read how to connect Sellforte MCP to your LLM client. You can also catch up on the Sellforte MCP launch post.
New use cases land in this library regularly, so check back soon.