@channel Question on my mind as I'm attempting to build AI workflows & use Claude code. As AI evolves β do you think GTM teams actually need to know how to build AI agents or just know how to use them? π¨Build β if you can't build it you'll always be dependent π₯οΈ Use β I don't need to know how the car works to drive it π€ Both β but most teams aren't there yet β Neither yet β the tools aren't ready for mainstream GTM π§ Depends on the role π§΅ your thoughts and how you're building/using ai
I think no matter the role people should speak to the work vs just pumping out product; otherwise why bother doing it if you're not learning.
Patrick A. agreed. do you think bigger companies foster those types of cultures?
Jared R. They used to, but the lay off trend, doubtful now.
It's not what they should, rather what an employer expects.
shopify famously said 'figure it out' Patrick A. - have you heard of any other companies fostering an even more productive culture
Jared R. I have not, but I haven't really put effort in to find out; I do know team leads are also getting let go for cheaper hires because 'I can give them rewrapped AI tools' vs using tools as a multiplier. Figure it out could also just be code word for 'results/outcomes' vs caring how it gets done - which, eh, is a double edged sword. I have been reading too much Plato lately though. Ill ask around though, this is a great question.
I believe its a split across the team. 20% Of the team some should now how to build and adjust agents 80% Of the are just going to be users/consumers of the output from the agent
What roles for each group maurice o.
My experience at TR leading a sales team tasked with 'demonstrate X amount of AI utilization a month across XYZ types' is that most GTM teams are drowning in tool sprawl already β "learn to code AI agents" is something that needs to be considered carefully with a serious opportunity cost lens. What I think is probably a realistic short term path is having a dedicated person with both GTM experience and an AI agent creation skill set to take on that burden, that may be an existing revenue generator with a curious mind and a bit of pre-existing dev experience, or may be a dedicated hire. Otherwise, the big risk is you end up taking resources whose sole function is to generate revenue, typically with relatively short horizons (annual max), and distracting them with something that may significantly reduce their output in the short/midterm. But here's what I've noticed: the teams that understand at least the basics of how these things work make way better decisions about which tools to buy and how to implement them. They ask the right vendor questions, spot the BS demos, and don't get oversold on capabilities. What's your take β are you seeing teams that jumped straight to using AI tools without understanding the fundamentals struggle more with implementation?
I just restructured my team: Builder:
Agentic Engineer
Consumer:
RevOps analyst
Sales executives leadership
Marketing Leadership
etc...
β³theyβre not ready yet, 98% of the workflows are used to build scrapers of public data and are creating a massive mess in the CRM systems. I see more and more teams drowning in super complex setups that they think saves them time, but they end up going back to easier tooling.
I think the technical skill gap between build vs use an existing solution is going to disappear When it comes to using these AI coding tools, you just need to start playing around, come up with some tools and be patient enough to stick around until you figure things out
My thesis is that people will be the systems thinkers and have to teach agents how to build processes and monitor those processes. I am building a claude template for this, the goal is to automatically pick up what the user is doing and codify that into skills and workflows that can be shared with teammates in a safe, governed way. Check it out, I'd love some feedback: https://github.com/jaronsander/agent-gateway
