Building next gen Bus-Ops platforms by partnering with accounting and PE firms. https://xitter.cf/t.co/4IKQZnoJz7 / https://xitter.cf/t.co/3m9GPyOb5E
San Francisco
Joined September 2009
- Tweets581
- Following1.7K
- Followers223
- Likes1.8K
You guys want another conspiracy? Almost no european opponent until semis for Argentina. Coincidence?
💯
Regardless of what people say, I genuinely believe that we have not yet seen any real productivity gains from AI at the enterprise level
Almost all of the efficiency gains and layoffs that you see from companies are a result of overhiring over the last few years
These two letters, "AI," have become the scapegoat term for a lot of CEOs and CFOs who want to cut their budgets without admitting that they messed up
While some people agree with this and point to it as a bearish case for AI, I think the opposite is actually true
Here is why
Implementing AI into your business is not supposed to generate productivity gains right away, and I think a lot of people completely miss this point. If anything, AI is a technology that takes much longer than others to show any real ROI for an enterprise
Let's walk through an example to show what I mean
When you implement AI into a business, it has zero context around work processes on day 1. It doesn't know what the typical template format of the business looks like. It has no idea what a "client-ready" deliverable is supposed to look like from your company
At this stage, you still need employees to use this tool in their daily workflow and teach it how to add value to the business. Employees can pick up small efficiency gains in repetitive tasks, but the work needs to be continually audited or edited to turn it into a "finished product."
As employees implement it and usage increases, the AI gets to pick up context and stylistic preferences that are specific to the business. After enough time, that same tool learns to feed this information to become incrementally more useful each time
The two key pillars of this are (i) custom prompts built by employees who want to solve for a specific pain point in the business, and (ii) skills built by employees who are looking to automate an entire task from end-to-end
This is just the baseline today. As the models get better, all of the top platforms will add new features to help capture this process. In fact, many new startups are already budding today that aim to tackle this exact problem of extracting enterprise workflow knowledge
As this happens, AI startups to take over entire processes from humans and actually builds the "client-ready" finished products that can replicate an entire job.
And I think, this is when we will finally see a massive uptick in real AI-led productivity gains
As of today, 99% of companies are simply not there yet. Most businesses are just beginning to onboard AI tools. Others have just onboarded and have not yet built up that enterprise knowledge inside of their tools
I think the simple bottom line is that this technology is just getting started
I had an interesting conversation with an accounting firm owner today.
He told me something that most AI discussions completely miss.
Reviewing a tax return can take him up to an hour.
Could AI review 95% of it? Probably.
Could it review 99% of it? Maybe.
But he still has to review the entire file himself.
Why?
Because if there's a mistake, he's the one carrying the liability, not the AI.
That's where many AI ROI calculations break down.
Automating 95% of a workflow sounds impressive.
But if a human still needs to perform a full review at the end, you haven't removed the bottleneck. You've just changed the way the work gets done.
In industries where accuracy matters more than speed, the last 1-5% is often worth more than the first 95%.
AI won't truly transform these workflows until it can either:
1. Reach near-perfect reliability, or
2. Change how trust, accountability, and liability are handled.
Until then, we need to be honest about what AI is actually saving and what it isn't.
Dropping him out was the right call especially after Maguire’s ego-bruised statement
One of the most underrated approaches currently is to try to partner with industry veterans with deep networks who are bored with their life and looking for some energy. They have the tribal knowledge, and if you have the technical/AI edge, that can be a very productive relationship!
Main Street will be biggest beneficiary of AI
OpenAI is aggressively hiring leads to manage partnerships with private equity firms and drive efficiencies into their portfolio companies
Earlier today, Anthropic announced Claude for Small Business, which allows AI to work inside tools that are regularly used by SMBs
It is clear to me that the technology is slowly making its way down market to every corner of the economy
The adoption curve is moving beyond just the Fortune 500 or the tech startups in Silicon Valley
It is coming to your local HVAC repair shop, laundromat, restaurant, car wash and construction business, probably in the next 12-18 months
Things are about to get really interesting
Market starting to realize the SMB market is going to be as important if not more in the next decade.
Anthropic just made a massive bet on small business AI through Claude for Small Business.
This story is profound & personal for many reasons I'm about to share...
1) @tenex_labs is @AnthropicAI's official AI transformation partner for Claude for Small Business.
Some highlights of the partnership:
- We co-built the Cowork small business plugin: 31 skills built around the workflows SMB owners actually run (planning payroll, closing the month, chasing invoices, running campaigns).
- We co-created the Claude SMB Tour: a 10-city, free, half-day workshop for 100 SMB leaders per stop. Starting in Chicago today.
2) Bringing SMBs along for the AI ride is so important for our country:
- Small businesses are 44% of U.S. GDP. They're plagued by scarcity of money, of people, of time. AI can bring them abundance.
- If every SMB gets a 10% lift in profits from this technology, that's a $1.41 trillion increase in U.S. GDP.
- And the point most people miss: SMBs can move faster with AI than any Fortune 500. No crippling bureaucracy. No governance overhead. No corporate politics.
If you want to see what's possible, read the recent Fortune piece on Rick Chorney.
29-year-old high school dropout. Started a janitorial company in Abbotsford BC in 2023. Year one: 7am to 1am, 7 days a week, $14/hr.
One day he spent 4 hours figuring out how AI could help. Automated intake. AI receptionist for $99/month (replacing a $4K/month role). AI email triage cut his inbox to 20 minutes a day.
Year 1: $242K. Year 2: $1M. Year 3 (projected): $1.3M.
He works 8-hour days now. Takes vacations. Uses Claude as his business advisor.
Rick is the roofing contractor in Tulsa, the bodega owner in Queens, the local accountant in Birmingham. He is the gold standard of what we need to help every SMB in American become.
3) I was this SMB audience for a big chunk of my entrepreneurial career.
For most of @MorningBrew's life, we were a small business. Small team. No venture funding. Big aspirations. Scarcity was a breeding ground for creativity, but it was painful too. Worrying about payroll. Working insane hours. Saying no to anything that wasn't mission-critical.
When I look at what AI can do today, I can't stop thinking about how much further we could've gone, and how much faster, if we'd had this magic in our fingertips.
I want to make sure the next person building their version of the Brew gets to feel and harness that magic.
A CTO asked me an interesting question today:
“How are you solving the reliability problem with AI agents?”
He brought up something most people in AI avoid talking about:
Agents hallucinate.
They get stuck in loops.
They fail silently.
And one bad decision can ripple through an entire workflow.
Honestly, I think the market is still overestimating how “autonomous” these systems are today - especially in real operational environments.
Our approach has been much more practical.
Instead of giving agents unlimited freedom, we constrain them:
* scoped permissions
* structured outputs
* validation layers between steps
* confidence thresholds + fallback logic
* human approvals for higher risk actions
We also log workflow states so failures can be replayed or rolled back instead of silently propagating downstream.
What we’re seeing is that businesses don’t actually want AI making every decision.
They want:
* less operational friction
* faster execution
* less repetitive coordination work
* better visibility across operations
Reliability matters more than autonomy right now.
Especially in traditional industries.
To stand a chance of success in the world of generalized AI and LLM models you need to choose your domain of niche/expertise very carefully. It is not about execution anymore. Positioning and distribution will determine the scale of your success. We are early in this process ourselves. We can either spend time now to find your niche or waste time being mediocre.
Rental search in the bay area is brutal. 3 months in we are still looking. No one is buying and all those qualified buyers are in the rental market. It is mot abnormal for bay area rental to get 20 plus applications.
the biggest advantage in business used to be size. more people, more data, more budget.
now it's the opposite.
i watched a billion dollar software company quote a manufacturer 12 months to onboard. we did it in 4 with a fraction of the team.
the shift isn't AI vs no AI. it's who can move fast enough to use it before the committee finishes their vendor evaluation.
enterprise advantage is becoming enterprise drag. and the operators who figure that out first are the ones winning deals right now.
my wife started making music a few months ago. no studio, no label, just her laptop and some free tools. now our kids only want to listen to her songs in the car. AI didn't replace her creativity. it just removed every barrier between her and the thing she wanted to make.
we onboarded a manufacturing firm in 4 months. the billion dollar software company they were evaluating quoted 12 months. honestly i think big companies just sell timelines at this point. we showed up and needed to get the thing working so we did.
moved from Toronto to the Bay Area this year. thought the biggest change would be the tech scene or the weather. nah. it's the speed. in Toronto you schedule a coffee for three weeks out. here someone will take a meeting tomorrow morning if the idea's interesting enough.
had a call with a mid-market CEO last week. first thing he said was "show me something real, i've seen enough demos." so i walked him through one client. same headcount, double the output, costs down 30%. that was the whole meeting. no slides.
I was speaking with an accounting firm around AI opportunities and one major risk factor is they can't use services like OpenAI or Claude as they deal with highly sensitive docs from clients. My guess we will see very hyper-niche companies serving that space by using local models