Cory Mitchell

Ex-Buy Side Analyst Turned Operator | Building a Scalable Commercial Cleaning Company | Bringing Lean to Service Businesses

[Cory Mitchell @cjazmitchell]

Started in equity research. Moved to the buy side.

Landed on a board after a proxy fight.

Got let go.

Bought a janitorial business.
Got defrauded. Nearly went under.

Took it from ~$1M to $2M+ and growing.

I thought analyzing businesses meant I could run one. I was wrong.

[Cory Mitchell @cjazmitchell]

@blueprintsmb22 @SMB_Attorney I was nearly fully drawn on my line by month 2 because of compliance processes with property managers. Seller days receivable was 15 on net 30 invoices. I had to get compliant as a new entity. Didn’t get any cash in the door for 70 days.

[Cory Mitchell @cjazmitchell]

@m3_melody @KennyCap_Phd Ytd interesting… positive correlation with a lot of supply that’s come to market the last 12 months. Nashville has shown it can absorb. Any reason to be worried now?

[Cory Mitchell @cjazmitchell]

@BigDemoPrez I’m a dummy

[Cory Mitchell @cjazmitchell]

Chris is right. These services are more in the commoditized end of the spectrum. Easier to see downward pricing toward the low cost producers without a brand or sources of differentiation that would allow the producer to see a step change in margins with lower costs.

[Chris Hoffmann @STLChrisH]

I think this is right (ish)— but this shift will be accompanied by significant downward pricing pressure on agencies.

Companies can build these tools internally (and are), and agencies will be looking to capture share by lowering prices to a degree that reflects their new cost structure.

[Tyler Todt @tyromper]

A guy in my sister's class (5 years ahead of me) joined us for dinner one night.

He was there once.
Then never again.

I didn't know why until 30 years later.

This guy grew up in a troubled home.
No dad. Mom was a mess.

They were super poor.

One day my dad went to pick my sister up from school & saw 5 kids picking on him.....

He was cowering in a corner.

My dad ran over immediately.

[Cory Mitchell @cjazmitchell]

@eduschesne1

[Cory Mitchell @cjazmitchell]

Ideas for janitorial:

1. Permit & License Pull
AI agent monitors your county/city building permit database daily. Every new commercial tenant buildout, renovation permit, or certificate of occupancy filing = a business about to move in or expand. The agent auto-identifies the property manager or business owner and triggers an outreach sequence before competitors even know the space exists.

2. Google Reviews Scraper
AI scans Google, Yelp, and Bing reviews for commercial properties — specifically hunting for complaints about cleanliness, odors, dirty restrooms, or pest issues. When it finds them, it identifies the property manager and queues an outreach: “We noticed your tenants have mentioned cleanliness concerns — we’d love to show you what BHS can do.”

3. Job Posting Intelligence
AI monitors Indeed, LinkedIn, and ZipRecruiter for companies posting janitorial or facilities roles. A business hiring an in-house cleaner is either about to be disappointed or is already unhappy with a current vendor. That’s a warm lead. The agent flags them and initiates outreach timed to when they’re most frustrated.

4. Health & Fire Inspection Reports
Many municipalities publish health inspection violations and fire marshal reports publicly. AI scans these for commercial properties that received citations related to sanitation, cleanliness, or biohazard conditions. Property managers with a recent citation are motivated — they need a solution fast.

5. Commercial Real Estate Leasing Activity
AI monitors LoopNet, CoStar activity (via public listings), and local commercial RE broker announcements. Every new lease signed = a company moving into a space that needs janitorial services lined up. The agent identifies the incoming tenant, maps them to a decision maker on LinkedIn, and triggers outreach.

6. Facilities Manager Job Change Tracker
AI monitors LinkedIn for facilities managers, property managers, and office managers who just changed jobs. A new FM at a building is one of the highest-probability moments to win a contract — they’re evaluating everything the previous person set up. The agent flags the job change and triggers outreach within days of them starting.

The attorney’s model works because he found a data exhaust stream the competition ignores. For janitorial, the equivalent streams are permit filings, inspection records, lease activity, and job postings — all public, all ignored by most cleaning companies, all rich with intent signals.

The best part: tools like Clay, Apollo, and Phantombuster combined with a simple AI layer (GPT or Claude API) can automate most of this for a few hundred dollars a month

[John LeFevre @JohnLeFevre]

I'm not sure he wants me to mention him by name, but a close friend of mine runs one of the largest personal injury law firms in NY, NJ, and FL.

He now has an AI agent that scans EVERY police report to identify "victims" of accidents and crimes, filters it down to candidates for claims and settlements, and automatically sends them solicitations for representation.

While knowing that most cases settle and the law firm gets 40%.

And now he is licensing this process to other law firms.

[Cory Mitchell @cjazmitchell]

@chipperPat @Baup0stGOAT With @chipperPat on this one!

[Cory Mitchell @cjazmitchell]

We just won our first class A rfp. Over 500k sq ft.

And what I’m most excited about isn’t the size of the account.

It’s the chance to roll out technology that almost nobody else in our space is using.

Property managers don’t just want a building cleaned. They want visibility. Accountability. Proof. Confidence that standards are actually being met when no one is watching.

That’s where this gets interesting.

We’re implementing new tech on this account that should give customers a level of transparency and operational control they typically don’t get from a janitorial vendor.

If we execute this the way I think we can, it won’t just help us serve this property better.

It will make us different.

And in a crowded market, different wins.

The goal is simple: build a case study,
show property managers something they haven’t seen before, and create a reason for more of them to want to work with us.

Big opportunity.

[Cory Mitchell @cjazmitchell]

We just won our first class A rfp. Over 500k sq ft.

And what I’m most excited about isn’t the size of the account.

It’s the chance to roll out technology that almost nobody else in our space is using.

Property managers don’t just want a building cleaned. They want visibility. Accountability. Proof. Confidence that standards are actually being met when no one is watching.

That’s where this gets interesting.

We’re implementing new tech on this account that should give customers a level of transparency and operational control they typically don’t get from a janitorial vendor.

If we execute this the way I think we can, it won’t just help us serve this property better.

It will make us different.

And in a crowded market, different wins.

The goal is simple: build a case study,
show property managers something they haven’t seen before, and create a reason for more of them to want to work with us.

Big opportunity.

[Cory Mitchell @cjazmitchell]

@AizikZimerman Congrats! All organic?!

[Cory Mitchell @cjazmitchell]

@pjmcgeary Great for rankings and seo. Miles and miles ahead, but as a consumer 98 reviews is a large enough sample size to get confidence in the star rating.

[Cory Mitchell @cjazmitchell]

@AizikZimerman We have an offshore recruiter as well and we’re 10% your size!

[Cory Mitchell @cjazmitchell]

@ClintFiore How often would you go to a place like that? I’m guessing about as often as you eat that type of food at the state fair.

[Cory Mitchell @cjazmitchell]

@chipperPat It’s being talked about more and more in smb twitter!

[Cory Mitchell @cjazmitchell]

If janitorial multiples were 3-4x where are they going if it’s one of the safest from ai disintermediation?

[SMB Attorney @SMB_Attorney]

Helpful if you’re considering buying a business:

Andrej Karpathy just dropped a project scoring every job in America on how likely an AI will replace it from 0-10

> Scraped all 342 occupations from the Bureau of Labor
> Fed each one to an LLM with a detailed scoring rubric
> Built an interactive treemap where rectangle size = number of jobs and color = how exposed that job is to AI

The key signal in his scoring: if the work product is fundamentally digital and the job can be done entirely from a home office, exposure is inherently high.

The scale:
0-1: Roofers, janitors
4-5: Nurses, retail, physicians
8-9: Software devs, paralegals, data analysts
10: Medical transcriptionists

Average across all 342 occupations: 5.3/10.

The entire pipeline is open source. BLS scraping, LLM scoring, the visualization. All of it. Much respect for the sensei this is scary and awesome

[Cory Mitchell @cjazmitchell]

@remotecleanguy Don’t forget the 2am calls because the alarm is going off!

[Cory Mitchell @cjazmitchell]

I think this also has to do with trust. Boomers trust their doctors and pop whatever pill they get pushed on them. Also has to do with heathspan vs lifespan. Younger generations, especially after Covid, are taking health into their own hands and are more focused on healthspan.

[Jonah Lupton @JonahLupton]

Younger generations want to prevent diseases and bad health outcomes. They are being more proactive.

Older generations just waited until something went wrong then took a pill.

I’m obviously generalizing to make a point but I do think healthcare costs have peaked as a % of GDP not to mention GLP-1s are bringing down obesity related diseases and costs.

Leaner people are healthier people.

[Cory Mitchell @cjazmitchell]

@Codie_Sanchez Number two hits hard