- Needs To Exist
- Posts
- Mercedes is buying the same playbook the plumbers are
Mercedes is buying the same playbook the plumbers are
Yesterday we spoke about the Avoca story. A $1B AI startup answering phones for plumbers and HVAC operators.
Today I want to do two things.
First, show you the same playbook running at the other end of the market. Then walk you through the 5-step check I'd run on any idea you're sitting on, and ask you to hit reply with your answer.

The other unicorn this week
While Avoca was getting written up in Fortune, Cognition (makers of Devin) was getting written up in Bloomberg. They're in talks to raise hundreds of millions at a $25B valuation. Up from $10.2B in September.
That alone is a story. Here's the part that matters for you.
Earlier this month, Cognition announced that Mercedes-Benz is deploying Devin across its global engineering organization. Starting where? Legacy modernization. Specifically, COBOL.
Yes, COBOL. The language your dad's bank still runs on.
The numbers from Cognition's own write-up are wild: 47% of organizations cannot fill their COBOL roles. 92% of COBOL developers plan to retire by 2030. There are trillions of dollars of business logic locked up in code that nobody under 55 can read.
Devin migrated a 25,000-line COBOL customs workflow at a top 10 global automotive manufacturer to AWS Lambda. A bank had Devin migrate hundreds of thousands of proprietary ETL framework files at 10x human speed (3-4 hours per file vs 30-40 hours).
Goldman Sachs. Santander. Nubank. Dell. Cisco. Mercedes-Benz. The customer list reads like a Fortune 100 page.
Now look at Avoca's customer list from yesterday: 1-800-GOT-JUNK?, Goettl, Turnpoint Services. Plumbers and roofers.
Different zip codes. Same playbook.
Both companies looked at industries Silicon Valley considered uninvestable, found the workflow that was eating the most human hours, and built the agent that does it. Avoca for the front office of the trades. Cognition for the legacy code rooms of the Fortune 500.
If you take one thing from this weekend, it's that the playbook scales. There's a version of this for almost every industry between a $50/hour HVAC dispatcher and a $300/hour COBOL contractor.

How Marketers Are Scaling With AI in 2026
61% of marketers say this is the biggest marketing shift in decades.
Get the data and trends shaping growth in 2026 with this groundbreaking state of marketing report.
Inside you’ll discover:
Results from over 1,500 marketers centered around results, goals and priorities in the age of AI
Stand out content and growth trends in a world full of noise
How to scale with AI without losing humanity
Where to invest for the best return in 2026
Download your 2026 state of marketing report today.
The 5-step Avoca check
Here's the framework I'd run on any idea you're sitting on. Takes 30 minutes.
Step 1: Inventory your access
What industries do you have insider access to? Through a job, a family member, a hobby, a former life. Make a list of three to five. Not the ones that sound exciting on Twitter. The ones where you can actually pick up the phone today and someone will answer.
This is the part most founders skip. They try to build for industries they have to research from scratch. You probably already have a closer starting point than you realize.
Step 2: Find the front door
For each industry, where does the customer relationship START?
Trades: a phone call. Restaurants: walking in or Resy. Mortgage: a Zillow click. B2B SaaS: a form fill. Fortune 500 IT: a six-month RFP.
The front door is almost always the highest-leverage place to insert an agent, because every dollar downstream depends on it not getting dropped.
Step 3: Map the system of record
What software does the industry already run on? What can't they switch off?
ServiceTitan for trades. Toast for restaurants. Procore for construction. Epic for hospitals. IBM mainframes running COBOL for the F500 back office.
That's your platform. You don't replace it. You build the agent that talks to it. Avoca talks to ServiceTitan. Devin reads COBOL inside IBM mainframes.
If you can't answer this step, you don't know the industry well enough yet. Spend an hour with someone who works in it before you write a line of code.
Step 4: List the workflows
For that industry, write down the top 10 things that get done over and over. Not the strategic stuff. The repetitive grind.
Phone calls answered. Quotes sent. Invoices chased. Schedules built. Reports generated. Code translated. Forms filled. Insurance pre-authorizations submitted.
Now star the ones that produce data each time they run. Every Avoca call generates a recording, a transcript, and an outcome. That data is the moat. The stars matter more than the count.
Step 5: Pick the boring one
The boring repetitive workflow is the gold. Three reasons.
One: nobody's emotionally attached to it, so customers will swap it out faster than they'd swap out their CRM.
Two: it generates data on every run, which compounds the system over time.
Three: nobody talks about it at conferences, which means less competition.
If your output from the 5 steps sounds embarrassingly small, you might be onto something Avoca-shaped. If your output is "a better ChatGPT for X," you're still in trouble.

Now hit reply
This is the part where the email earns its place in your inbox.
I want to read what you came up with. Hit reply with three things:
The industry you have access to
The system of record they use (or "I don't know yet")
One workflow that still runs on phone calls, spreadsheets, or paper
I read every reply. The most interesting ones I'll feature in a future edition (anonymized if you want). Some of the strongest NTE Pro write-ups have come directly from reader replies.
If your industry is one we already have research on, I'll send back the three most relevant ideas from the database so you can skip the blank-page phase entirely.
If we don't have research on it, that's even more interesting. Reply anyway.
Where this fits with what we do
NTE Pro is the pattern-recognition gym. The 7,000+ ideas in it aren't there for you to copy. They're there to train your brain to spot the workflow nobody's automated yet. After fifty of them, you walk into your friend's pediatric dental office and within ten minutes you have three Avoca-shaped opportunities mapped in your head.
WhoFiled is the timing layer. If anyone is about to raise capital to build the COBOL-modernization-for-mid-market-banks version of Devin, or the missed-calls-for-elder-care version of Avoca, WhoFiled is where you'll see it weeks before the Fortune piece runs.
Both are built for the operators reading this, not the spectators.
One last thing
The pattern this weekend was the same in both directions. Plumbers or Mercedes. $50B HVAC market or a trillion dollars of COBOL. The opportunity isn't AI capability. It's domain access plus an integration plan.
If you have access to a domain, you have an unfair advantage. Most of you are sitting on one without realizing it.
Hit reply. Let me know what yours is. I'll read every one and write back where I can help.

