Underhyped:
Lead Repository: building a repository of all leads, campaigns, messaging in a real relational database in Supabase/Postgres
Why: Imagine being being able to have an agent review every new campaign in context of
- everything you've tested previously,
- what A/B testing splits were most effective
- every single data point you capture feeds into the next iteration
How much better do you think each new campaign iteration would be?
MCP for GTM:
When done correctly for a specific problem, the iteration tempo unlock is crazy.
Being able to control your smartlead/emailbison then:
- have an agent review campaign results,
- plan a new iteration
and scaffold out a new one in 2 minutes vs 2 hours is real value.
Overhyped:
Generic off the shelf Agents:
If it is is not fit perfectly to how you really work (aka create value), what's the point?
It will just become:
- something else to fix
- eat api credits
And automate the wrong thing.
Effective application of agents requires deep understanding of the real problems at hand.
down to literally the exact unique way you work.
Which brings me to the next section. It depends:
Extremely complicated n8n workflows:
The ones you see people sharing out on Linkedin are overhyped.
But if you build one yourself (or I guess adapt one of those linkedin ones tbf)
- for a very specific task
- fully aligned with how you creates value
Then its on the money.
Why: You should always pick the simplest tool for the job to accomplish a given task.
Throwing a 90+ node workflow at a problem you don't have causes more problems for you than it solves.
ID a real problem you have, and start with the simplest tool and step up in complexity only as needed to better solve the real problem.
This post was originally shared by on Linkedin.