If AI Can Replace Workers, Why Is It Hiring Consultants?
Last week, I noticed that Anthropic was hiring for a partner success manager.
The very first line of the description was:
Consulting and systems integration firms are racing to build Claude practices.
From there, the posting went on to explain how important it is for Anthropic to support consultancies, systems integrators, and implementation partners as they bring Claude into the enterprise.
And all I could think was:
Wait.
If the AI is good enough to replace engineers, why does it need a small army of consultants to help companies use it?
On one hand, we keep hearing that AI is going to automate software engineering, data work, analytics, customer support, operations, and almost every other form of knowledge work.
It will end work as we know it!
And yet, every major AI company seems to be investing in partner ecosystems, systems integrators, implementation teams, enterprise success teams, and consultants in one way or another.
I had a similar thought about Salesforce recently.
If your AI is so good, why do companies still spend millions of dollars paying consultants to set up Salesforce, customize workflows, clean up CRM data, integrate systems, train teams, and keep the whole thing from becoming a very expensive mess?
Shouldn’t AI be able to just… do that?
But of course, anyone who has worked inside a real company knows the answer.
The hard part is rarely clicking the buttons.
The hard part is understanding the business, the process, the edge cases, the incentives, the data, the politics, and the mess that already exists.
So why?
Why do the LLM labs want to partner with consultancies?
My thought?
AI is powerful enough to change work, but not simple enough to magically reorganize companies by itself.
AI As A Utility
We are having our Prometheus moment. We’ve been given fire, and we are all sitting around wondering what the best way to use it is.
Most of us are using it to keep warm and make incremental improvements to our lives. We are using it to code slightly faster, summarize documentation, help teams move through information, and, of course, send out spam marketing even faster.
But eventually, someone realizes fire can do more than keep you warm.
Someone is going to figure out that you can heat water and take the compressed steam and move tons of goods and people across a continent, among other things. In the same way, some consultancies and individuals will find brand new uses for LLMs that change the way we work.
Those are the great consultancies the LLM labs want to work with.
The ones that are asking...
“What kind of work can exist now that this capability is available?”
Someone Needs To Build The Appliances
In a recent article Joe Reis referenced the fact that electricity is the right analogy for AI not the dot-com boom. He had a slightly different angle but I think this framing is useful.
Because simply having access to electricity did not transform the world by itself.
Electricity became valuable when people built things that used it.
You know, the things we take for granted now…Light bulbs, refrigerators, washing machines, elevators, factories, etc.
And here is the thing, there was a time where electricity was seen as a novelty, and until people created these appliances, it essentially was.
A company can give every employee access to AI and still have no idea what to do with it beyond:
“Write this email.”
“Summarize this meeting.”
“Help me with this spreadsheet.”
Useful? Sure.
Transformative???
That’s why I believe Anthropic wants to partner with consultants.
They want to.
Increase token consumption
Support consultants who find ways to drive value with AI in new ways that take things to the next level
And a list of other benefits I’ll discuss in the following section.
But before that, I believe it’s important to go over a couple other things they’ve stated.
They know that:
A focused fraction of them will become great — the kind of partner whose architects deliver Claude work customers love, who specializes deeply enough to be the obvious choice in their lane
Some of these SIs, and consultants will be able to message the value of AI in new ways and drive new forms of value where AI becomes more than just an add-on.
Quick pause: If your team is trying to figure out how to integrate AI in a way that actually drives business value, and you’re not sure where to start, feel free to reach out to Dorian or me. At CodeStrap, we help enterprises move beyond chatbots and demos by turning messy business processes into reliable, observable AI workflows. Alright back to the article!
These consultants and SIs will find ways to integrate AI in totally new ways. Maybe there are manual processes that will get replaced.
That is why I like the appliance analogy.
As a funny aside, there was a point before electric vacuum cleaners where there were hand-pumped vacuum cleaners. Which is hilarious to think about now…needing to work so hard to barely vacuum any dirt whatsoever.
I think that’s also something worth considering. Many of these technologies are built on top of each other or perhaps in entirely different skill trees.
Then someone, one day comes along and realizes, wait, we could improve the vacuum experience by making it electric.
Final Thoughts
Plenty of CEOs and VPs I talk to want AI integrated yesterday. They know that if they can skate where the puck is going, they’ll be far ahead of the competition. But we are still in the early days of figuring out exactly how to make this tool useful.
Really useful.
I imagine like most technological shifts, this will be slow. We’ll make some incremental progress. Then get stuck.
Then someone else will come along and see a few different tools that exist and see how they could all be used to improve a problem we are currently facing.
Until then, thanks for reading!
Not every query needs a cloud data warehouse. 99% of BI queries are small, scanning less than 100GB. Running small queries on Snowflake is like taking a helicopter to go get groceries. Greybeam acts as an intelligent routing layer, matching every workload to its most efficient compute engine:
Cost optimization: Only pay for the warehouse for the 1% that needs it
Faster queries: They run 3x faster on the right engine
Data portability. Manage your data as Apache Iceberg tables, readable by any engine
Stop spending blindly. Start saving on average 86% of your compute costs immediately. No migration. No code changes.
[ Try one month free ] - Book a chat
Thanks for supporting the Seattle Data Guy!
Video Of The Week - The Data Engineering Job No One Wants To Do - Backfilling
Articles Worth Reading
There are thousands of new articles posted daily all over the web! I have spent a lot of time sifting through some of these articles as well as TechCrunch and companies tech blog and wanted to share some of my favorites!
Why Is Data Modeling So Challenging
Learning about how to data models from basic star schemas on the internet is like learning data science using the IRIS data set.
It works great as a toy example.
But it doesn’t match real life at all.
Data modeling in real life requires you fully understand the data sources and your business use cases. Which can be difficult to replicate as each business might have its data sources set up differently.
For example, one company might have a simple hierarchy table that can be pulled from Netsuite or its internal application.
Whereas another one might have it strewn across 4 systems, with data gaps, all of which need to be fixed to report accurately. So you’ll never really understand the challenge you will face when data modeling until you have to do it. Then you’ll start to understand how you should data model and all the various trad-offs you’ll have to make.
In this article I wanted to discuss those challenges and help future data engineers and architects face them head on!
So let’s dive in.
The Data Engineer’s Guide to ETL Alternatives
ETL alternatives help teams move and transform data using approaches optimized for cloud data warehouses, real-time analytics, and AI workloads.
Today, most data integration strategies fall into three approaches — ELT, ETL, and CDC — each optimized for different latency, transformation, and operational requirements. Choosing the right approach depends on how fast data needs to move, where transformations should run, and how the data will be used downstream.
This guide to ETL alternatives is continuously evolving. Conversations can be had on the Estuary Slack workspace and Estuary LinkedIn pages.
This guide is designed for teams facing three common data integration use cases:
Companies focused on lowering latency when loading a cloud data warehouse
Companies implementing real-time analytics using specialized databases
Companies starting to implement their first AI projects
End Of Day 219
Thanks for checking out our community. We put out 4-5 Newsletters a month discussing data, tech, and start-ups.
If you enjoyed it, consider liking, sharing and helping this newsletter grow.