Table of Contents
If you are a senior manager with enough tech cycles under your belt, you might be feeling it.
You might not be able to fully explain it because this time is different: it has new elements, and it feels bigger.
But it is happening. You know something is wrong with your engineering team, and you need to make changes fast. Otherwise, they are heading for a reputation crash (fueled by overpromise and underdeliver).
The problem? You brought an AI toolset that seems like a no-brainer — a productivity booster that can’t fail. This overlooked new element in your fragile ecosystem, your P&E team, is overheating its engines.
What is going to happen?
Back in 2000, I was hired by a local company to build an ERP from scratch (from scratch!).
I was told to use whatever the latest technologies were — cutting-edge architecture, everything new (my boss and his client did not know what they were actually asking).
So, I spent a year building the perfect tech stack, architecture, and back-end. My boss wanted to kill me, and the client tried to kill me — I was high on pride.
Now, when you bring a new, incredible, disruptive technology into your team’s hands, what do you think will happen? They will be hyped, show you how great it is, and spend more time with it than actually building your product.
Sooner or later, you will lose patience.
The Breaking Point
At some point, you would say, “Enough, let’s focus on the product,” -> and this is the turning point.
Would you throw away the AI boosters or keep part of them? Would you do an audit and verify with your most senior engineers which ones should be kept and which ones should go?
In my story, the client ended up pulling the plug. My boss did not have enough technical knowledge to continue with my work (despite the fact that I’ve documented everything and my code was pretty clear — I’m still proud of it).
My boss back then had two choices to turn the project around:
- Learn the technology himself, simplify it (identify the positive ROI pieces), and cut scope to ship.
- Hire an external consultant who can do #1 for him.
And that’s exactly your options right now.
Throwing away AI to reclaim your productivity
You are well-connected; you might know someone who has been playing with AI, telling you the wonders of it. Or that person might know someone who is helping startups implement AI. It doesn’t matter if that person is living next to you or across the globe.
You might already have an AI expert, but soon you will discover that an AI expert means different things to different people: it could be an expert in AI tools for a given space, a prompt engineer, or maybe a data scientist or LLM builder.
All I can tell you is that the actual AI experts are already hired by the big companies, so the chances are that this person has been following the trend and sees others (or himself) take a dent in it (like me).
What you won’t find easily is someone who actually implemented AI tools in startup engineering teams in a way that you see significant (x2) speed and delivery improvement.
Someone who knows how to implement AI from customer feedback, to requirements, to UI/UX design, architecture/infra, and coding implementation.
The market is too new, too immature to have that kind of expert.
What should you do then?
Oh, you won’t like the answer, I’m sorry — you are already in the problem, so you are remediating the situation at this point, not entering a greenfield state.
Let’s assume you are in the early stages of using the AI toolset for 1-2 months. If you can find a colleague or someone you trust who has actually implemented AI tools in the development workflow, schedule a meeting with that person.
If you don’t, take your most senior engineer — trust him completely to find the right AI tool (there is only one!) and use it for a sprint in a specific low-risk feature.
Ask him to write down prompts he/she used, workflows, ideas, and anything that came from that experience. (Remember, we are dealing with new tech, so apply it in your particular case here.)
That senior engineer is creating the playbook to use AI correctly in your P&E team.
Don’t expect 2x productivity. In fact, expect the senior engineer to be at half velocity. Then repeat the same for another sprint. You and the senior engineer should validate the speed this time.
Did it increase 2x? OK, a playbook is ready to be tested with another engineer.
My company, InTheValley, can help you
Before we rolled out our best practices to our startup clients, we played with AI/LLMs in pet projects for almost two years.
Right now, when a Remote Engineer from InTheValley joins an existing team, it feels natural: they don’t preach AI, they pollinate the experience naturally, and they don’t try to disrupt any existing team processes.
They are just 2x more productive because they use AI tools as they use any other tool.
Have you seen an engineer using Visual Studio for the first time? After a while, they can fly. It is the same thing with AI: once the engineers know how to use it, they can fly.
You’d have to see it to believe it!
- You are in a Reorg Bubble about to Explode - 04/22/25
- The Wrong AI Tool Is As Bad As the Wrong Hire - 04/15/25
- COVID, Remote Work & AI - 04/08/25