Adam Giuriceo
Marketing Ops · Revenue Ops · GTM Engineering
8 years building campaign, lifecycle, and lead management systems on Marketo and Salesforce — supporting $20M+ in influenced pipeline while making funnels faster, cleaner, and easier to trust across global B2B teams.
This site is a working showcase — automations I've actually built with n8n, MCPs, Clay, CRM APIs, and LLMs, written up the way I'd walk a hiring manager through them.
What I’ve built
The Duplicate Contact Consolidator
Finds duplicate Salesforce records, asks for approval in Slack, runs Salesforce’s real merge through Apex, then re-checks HubSpot for one record per email.
The Self-Reported Attribution Reconciler
Tracking credited search with $677K of pipeline; buyers credited it with $123K. This build reads what buyers said and shows where the two disagree.
The Single-Thread Risk Radar
Flags renewals that rest on one relationship or none, catches departures nobody logged, and names who to bring in, quoted from the notes.
How I build
Most AI demos stop at “the model wrote something.” These builds follow five rules instead:
- Code does the math; the AI reads and labels. Every number in a post is computed, not generated.
- Every quote is checked. When Claude pulls words from a note or a ticket, code confirms they appear in the source word for word, or drops them.
- Risky changes wait for a person. Merges, lead reassignments, lifecycle fixes and outreach go out only after an approval click in Slack.
- Tested against an answer key, with traps. Where test data is seeded, it comes with an answer key and decoys built to fool a careless model, and the card reports how the build did.
- Honest about what’s real. Every card says which data is live Salesforce or HubSpot and which is seeded for the test.
Outside of Work
Reading, writing, hiking, photography, and film — mostly in that order on any given weekend.
Reading
Untold Story
Dick