Who Is Really Getting the AI Productivity Gains?

Most of us feel overawed where to place our AI bets. I’ve definitely felt uncertainty making some calls, so I empathize a lot with fellow C54 members being hit from all angles… However, I luckily have a fool-proof methodology that will make us all millions more dollars!!!

At least that’s what ChatGPT said I should write about. 

The reality is very different – overwhelmed by how AI improves the speed of content and data-crunching use-cases. Underwhelmed by the quality and consistency of anything requiring true rigor and precision.

No-one has all the answers on what to do, but I’m immersed in the tech, software & AI tools available to Pro Serv firms, so I wanted to share what I’ve seen working, and what is flattering to deceive.

Vibe-coding: Time suck or productivity hack?

You know $CAC doesn’t matter and money is sloshing around tech, when 50% of the Superbowl ad’s featured some form of AI or vibe-coding tool…  Standalone vibe coders, big SaaS with integrated AI, LLM’s themselves poking fun at each other…

But do you need everyone in your firm to be able to vibe code an App? These AI tools are not being built with interoperability in mind, so more data silos are being created…Software vendors & LLM’s will want to lock you on their platform too, mind.

And even if you did… is it making life better? 

EBITDA, gross margin & revenue growth have, on average in boutiques, barely accelerated since 2023. In the past 2-3 years, the average firms financial performance has plateaued, yet 95% have tried to leverage AI… so whilst we all feel more productive, who has benefited from the gains? Not Owners or Founders. It’s going to the employees.

TLDR – stick with me

In smaller firms, where data doesn’t need to flow or be accessed by many people – vibe-coding is a great productivity driver. SMB’s whose delivery is largely contingent on generalized content or structured data, can deliver more, faster. 

In larger firms, where there’s already too much fragmentation of process, workflow and data – vibe-coding is a genuine nightmare. It just means you have a higher volume of nonsense, being delivered at warp-speed.

… and I appreciate this is where I may run into a debate with Matt Alexander, who predicted on LinkedIn the demise of CRM’s in <5 years… whereas I believe big software is best-placed to take more of our $tech-spend dollars.

You, me, anyone in C54, we don’t actually need software… or AI… or vibe-coding, for functional purposes… 99% of the value is it embeds standardization, process & data integrity. Dave from Sales, vibe-coding a CRM on Monday morning after being hammered for having terrible pipeline management, and blaming it on a missing field, is not thinking about any of those things. 

Guess what tech the AI & vibe-coding businesses use themselves internally too? They’re all on big-brand CRMs to make sure every sales-rep follows the same process. They’ve got an Ops platform to build products & deliver projects. Their accountants aren’t on Loveable – they’ve got popular finance or ERP systems, which will forever be needed to send bills & collect cash. They’re not reliant on vibe-coding, because they’ve moved beyond scale. 

$50 licenses vs. $100k salaries: placing smarter bets

I might be wrong, but I imagine a disproportionate amount of your current time & effort is going into thinking about, signing off & implementing tech, compared to, let’s say, your biggest cost base – your people? 

At C54’s Reunion in October, Greg introduced Paychex, who told us all the average salaries for Pro Serv professions across all US States was around $six figures per year. (there was no Canadian data – don’t shoot the messenger).

For the same expense as one employee, you could have 2000 months access to your favorite $50 a month software product. That’s 166 years worth of access. You can grow > scale > exit 10 different firms in that timeframe!

So, the point is… don’t overthink where you place some tech bets. You cannot believe the huge failures & colossal wastes of money, that we’re going to see in AI over the next few years (I’m talking about following the big tech’s billions, not the $10k-$100k most of us have as tech budgets)…

You won’t get right every time, processes will have to change, and the firms who come out the best are probably those that make more nimble decisions & are prepared to cut, switch & pivot, when you outgrow or need different tech. 

Getting off the fence – conclusions

I actually think it’s pretty simple… and it applies to AI, tech, people, in fact anything you spend your $$money, time & brainpower on… Does it create Enterprise value, higher EBITDA or improved productivity? Everything else is just noise.