This is like Thanos getting the 5th infinity stone, but in a good way.
Congrats @local0ptimist ❤️
good time to announce that i’ve joined @GoogleDeepMind as a forward deployed engineer
i’ll be partnering with startups and researchers across the ecosystem to co-design benchmarks.
if you’re building interesting evals, i want to hear about them!
I have reviewed over 75 proposals and sat in a couple dozen meetings over the past two weeks on behalf of a big name fashion brand trying to hire an AI agency to lead their AI implementation.
This will be their 4th agency since they started opening up to AI back in March 2025.
They reached out to me from a referral because they need help finding a good agency that won’t overpromise and underdeliver. I provide the guidance, oversee the training of employees, and ensure the agency is capable of supporting a full organization wide transformation process. Not just marketing, not just ops, not just one little workflow. The whole company.
I have time for agencies that are trying to do right by their clients and know when they need help. The markdown agencies selling expertise they can’t explain and expecting Claude to make every decision for them are the first to be tossed to the side.
Some gaps that keep coming up in proposals and meetings with me and the CTO as we interviewed the agencies:
- You can explain the tech stack and demo for PDF data extraction on Purchase Orders, but what happens when an approved order changes? Who needs to act on it, and what gets updated downstream?
- You quote three months to the first build and pitch another audit. What can you reuse from the last three agencies, and what still needs checking?
- You send every artifact to AI with minimal guidance, so it is easy to tell that we are communicating through Claude. But when we ask them to explain their recommendations in the meeting, they can’t. They run the meeting transcript through AI and send us next steps that are all redundant, but when we ask them in a meeting for specific insights they say they'll get back to us.
- You sell hours saved as ROI by multiplying time by salaries. Those salaries aren’t automatically going away. What work can those employees take on now, and how does that benefit the company?
- You claim to sell "outcomes", but the outcomes you describe provide observability, not actionable steps and proven systems. An AI agent that gathers all the data and lets you query on it is useless compared to a system that uses that data to get more customers and provide more throughput all on its own.
A good agency needs to be able to pick up where they were at, check what actually works, and make better decisions based on that information, and re-assess as they go.
A good agency would come prepared with the ordering, fulfillment, and sales processes they expect to see. Explain where they think the bottleneck is and ask, “Is that accurate?” We can confirm or deny, but at least it shows where they are thinking and the expertise they would provide... then the agency can explain how automating one part leads to the next improvement or depends on a previous step.
For example, you automate the orders coming in and get all of the data extracted, nice and good to go. The people working with it afterward are reconciling it, confirming it, and connecting those orders to fulfillment and cash flow.
Once we’ve proved which numbers we can trust, they can spend more of their time doing predictions, testing, analysis, and bigger scope items that can actually drive revenue.
Now we can start pulling levers. Say an item was successful at a certain time last year. Let’s put a marketing plan in place, test how it does with different traffic, see how the cost changes, and test new prices. Check the margin and whether we have the inventory to support it. Those are all steps that will eventually be automated, and a good agency would need to know how to support that pathway or at least have a strategy to explore. I've done this whole process at over 200 companies and even if you're at your 1st or 10th implementation, it is possible to know these things and come prepared with it.
A lot of bad agencies, aka "markdown agencies", are trying to get away with doing the bare minimum by pitching "AI audits", "AI Enablement training", and "maintenance fees" to help sustain the systems... but what they actually deliver are roadmaps generated by Claude, a weak AI guide as they help setup a Claude Team setup for an org whose identity and governance requirements it doesn’t meet (bad for any org that needs to take security seriously), and the maintenance fees are often met with bad service long term.
The last two agencies that have been considered for this opportunity both pitched $75k-$150k per month, but they could easily get $250k per month from an org this size if they just came prepared with the right items and momentum behind their builds.
As of today, I finally hit 250 companies where I have led, implemented, or consulted for full, org wide, "AI transformation". All while maintaining a full time job, what a blessing.
Rewarding myself with a three day weekend goodnight.
Every single company I have worked with, for AI enablement across the full org, has now moved on from overseas BPOs.
Throughput has been increased.
Redundant work decreased.
Less time spent managing.
More money coming in.
Hire local.
One thing that really stuck with me from my recent conversation with @coreyganim (he makes $1000+/hour with AI consulting):
You don't need to be an AI expert to sell AI consulting. You literally just need to stay one step ahead of the client.
It's the "Catch Me If You Can" principle.
99% of business owners aren't using AI properly and are frankly too busy to stay up to date (this space moves incredibly fast).
If you can just figure out their next step for implementing AI using the latest tools, you can monetise.
You're not selling deep expertise. You're essentially selling the willingness to sit down, audit their biggest bottleneck, and hand them an updated fix.
Think: building Claude skills, implementing Jev, Muse set up , etc.
That's why I think AI consulting is one of the best business models to start with right now. Low barrier to entry, and the demand is everywhere.
giyu_codes retweeted
Replying to @anton_onAI
You can get a majority of the answers looking at the data available.
You can still interview Dave, but inexperienced agencies will take Dave's words and hold onto them instead of thinking about the big picture.
For example an agency today was building an invoice reconciliation pipeline and they were using "Dave's" master spreadsheet as their source of reference for what is accurate. Then they went to go talk to Becky about where she gets her purchase orders from and she says they come in through email. Then they went to go talk to Andrew who provides them with receipts and bank statements for remittance.
By the time they were done, the data was all wrong because they didn't stop to think about the higher value source of truth which is:
- the vendor portal that holds all the invoices (dave's spreadsheet only accounted for 80%, which explains the finance leaks)
- the many different bank accounts that hold all the financial transaction data coming in and out (Andrew only had 50% of this up to date and accurate when compared to his spreadsheet)
- Accounting's email, which had more purchase orders but they forgot to forward 10% of all incoming purchase orders to Becky's email address.
I don't think markdown AI agencies understand how dumb they sound.
"For $3k I will build you an AI agent. I just need you to walk me through your workflows, your day to day, and how your business operates."
You think execs and business owners want that?
The difference between a 3k contract and a 30k contract is your ability to express your expertise and know what they need without them having to hold your hand.
What I really respect about consultants within the "Big 10" is how they handle criticism.
Over 50 engagements where I had to deal with them:
- 45 attacked my credentials
- 41 blamed me for the failure of their system
- 37 apologized after the fact, but not on the record
- 25 would have their management step in
Every single one of them would ask me to teach them on how to improve and they would also teach me on how I can do things better too.
Natural, emotional, responses are expected when I come in and audit the work of a third party on behalf of a company.
But I am always humbled when I look at the data I have and see that their learning trait is consistent and they are equally excited to exchange info.
That said, it's absolutely nuts to me how often they hire outside consultants lmao. It's like subcontractor on subcontractor on subcontractor.
Crazy how no other AI agency has talked about what happens after a company is fully AI native.
- The dependencies and bottlenecks are external
- Lower level roles get grouped into higher level roles with more pay and responsibility
- people spend more time talking to each other and collaborating on new ideas
- AI chatbot usage trends downwards
X is three years behind
If you want an example of how the type of people I think are just jaw dropping good at upscaling themselves, scroll through his profile.
Dude has the most interesting story and has such a unique app that he's building.
10/10 if I don't wake up to this guy having investors I'm firing another markdown agency this week.
Over the last month I have seen:
- Junior dev outperforming an ex-FAANG senior dev on deliverables and adoption rate for internal apps they built across an org of 700
- ex-FAANG staff engineer turned into COO of a non-tech company
- non-tech COO transitioned to CTO after outing the CTO for his lack of progress in the org and "failure to navigate the modern landscape of AI"
- three non-technical fashion designers lead a Claude app training session for the rest of their team while the engineering team was still planning how to roll out their one interal app properly. that internal app is now cancelled because they took too long to deliver and the Claude training session covered the use case
Life favors those who do things, regardless of slapping an "AI" label across it. Tale as old as time I guess
Long week.
- Helped a coaching client negotiate up $30k and some equity on an offer
- Fired an AI agency that did not know about TCPA compliance and the company that hired them got hit with a ton of demand letters totaling $150k. The company that hired them makes $20k a month lmao its a small business and now they're done with all AI agencies after this hahaha
- Helped vet and hire an AI agency for a CPG company making $50m a year. The company has gone through three agencies that just hooked up claude and called it a day. Somehow they found a new agency and from the 5 employees, NONE of them have strong technical background but somehow they're incredibly talented and learned how to upscale themselves on the tech side and they have had the best security and infrastructure implementation I have ever seen for an agency. They earned their $250k contract for sure
- My main job is just director of AI at a company, and I do all of the above on the side. We just finished the change management process of our internal app that facilitates the research, order tracking, and fulfillment of our main product lifecycle. All running off of Gemini for every automation ❤️
- Today I had to read 10 proposals on AI agencies selling their "Company Brain" markdown file shenanigans trying to charge $25k a month for hooking up claude and some tools for a company size of 500+. The audacity of these markdown agencies is growing.
What a blessing, considering I graduated from university 3 years ago.
Genuinely zero excuses now.
I remember when I had to use AI studio because I couldn't afford subscription plans.
Now you can get community support for your project with compute credits.
LIKE PLEASSSEE JUST GO BUILD
How it feels having to deal with these markdown AI agencies that keep spawning.
business idea: set up the second brain for small businesses that want to use ai but don't know where to start
most small businesses tried chatgpt, got generic output and gave up. the missing part is context.
sell a setup where you build their files from what they already have:
> offer.md from their price list and proposals
> customer.md from their reviews and sales emails
> voice.md from their best emails and posts
> faq.md from their support inbox
then show them how to point claude or chatgpt at the folder. charge once for the setup, and monthly to keep the files updated as their offer changes.
today someone from an AI agency told me I am like Jesse from Breaking Bad from season 4 episode 10.
Someone tell me how to feel about this
Sitting in some meetings where AI agencies are trying to sell themselves to the execs is hilarious.
"We are going to build you a second brain"
"We will start on one workflow and save you $100,000 per year"
"Our AI audit assessment is only $1k and it will be credited into the bigger project later".
Then they try to charge $50k for one workflow with a $5k monthly service retainer??
Zero sales skills. Zero tech skills.
Yuck














