An ad-sales firm for podcasters and YouTubers
We met the salespeople where the value was concrete on day one, while the data foundation took shape underneath. By the end, the firm's most valuable asset, who they know and what those relationships prefer, lived in a system they own.
Before
The client sells advertising on behalf of more than a hundred podcasters, YouTubers, and newsletter writers. Before we started, their data lived across hundreds of Airtable tables and Google Sheets, with thousands of automations layered on top, Airtable used as a notepad, Sheets used the same way, and nobody clear on which version of any record was actually true. It was costing them speed on RFPs, confidence in conversations with creators, and probably some deals they didn't even know they'd lost.
The first bet
Help the salespeople win an RFP without leaving their desk
We started where they could feel the value on day one. When a brand sends an RFP, the system reads it, finds the matching shows in the roster, and drafts a personalized vetting email to each creator the salesperson can send with one click. When the replies come back, the system reads those too and logs each creator's preferences forever. The data work happened underneath, where they didn't have to worry about it.
Where the leverage shows up
What the salespeople didn't have time to do, the system does in the background.
Always the right data
Before
Before, getting a list of available creators meant cross-checking three different Sheets and hoping nothing had changed since yesterday.
Now
Now the system pulls from the canonical roster every time, automatically. The salesperson never has to second-guess it.
Dozens of vetting emails at once
Before
Vetting used to mean writing one email at a time, by hand. A single RFP could eat a whole afternoon.
Now
A personalized email goes out to every relevant creator in parallel, drafted and ready to send with one click.
Custom rationalization that sells
Before
There was rarely time to write a tailored pitch, so most emails leaned on the same boilerplate.
Now
The system writes a real argument for why each creator fits this brand, written one creator at a time.
Custom research per creator
Before
Research was whatever the salesperson happened to remember, plus five minutes of Googling between calls.
Now
Each email comes with fresh facts, recent context, and the angle on why this creator is the right fit for this brand.
The data work happened underneath. The salespeople just got their afternoons back.
What we built next
A single source of truth
Behind the scenes, all the brand records, creator records, and inventory now live in one place. The platform-specific Airtables still exist, but they feed the main system, and nobody on the team has to think about that anymore.
A roster the brands can browse
Brands and ad agencies can log into a private gallery and see the full roster, filtered by what they're looking for. We built it almost by accident, because the data was already there.
The RFP agent in Slack
The Anthropic-managed agent is now the first place every seller goes when an RFP arrives. The client reports that each seller uses it several times a day to add an RFP, find inventory, or take the next action without leaving Slack.
What was hard
Put the value where sellers could feel it
The first interface was the RFP response, not a data-cleanup dashboard. Sellers got a faster way to respond while the consolidation happened underneath the workflow they already cared about.
Move the system into the place work already happened
The managed agent came after the records and actions were dependable. Slack then became a useful interface to the system, not a chatbot sitting on top of disconnected data.
What's now possible
The Slack agent is now the first place every seller goes when an RFP arrives, and each seller uses it several times on a typical day. It works because the roster, inventory, relationships, and creator preferences underneath it are already dependable. Those assets now live in software the firm owns, not in a vendor's database priced per seat.
“We'd been trying to find a solution for that for years — this is perfect.”
Client executive
Real quotes from real clients. We anonymize on the site until each one says yes to being named. Most are in the queue. References available now.
What we built