Test 01
Data Readiness Panel
Can an agent actually do useful work against this company's data today? If not, what are the three highest-leverage cleanup projects to get there in 90–180 days?
An Outsider Labs advisory · 2026
Your company needs a harness, not a chatbot.
We run the 8-test diagnostic that tells executive teams exactly where they stand on AI — and what to build next.
A short conversation we keep having with operators.
Names changed. The pattern doesn't.
Outsider Labs
So what business harness are you running?
Business Owner
What are you talking about?
Outsider Labs
I'm asking how your teams are set up. Are they running on a real AI workflow? One framework that reinforces your business SOPs, plugs into your customer software, and gives the whole team deep AI insights. The kind of thing that 10x's how they work.
You're doing this, right?
Business Owner
Uh, sort of. We've got ChatGPT. The team's chatting with it.
Outsider Labs
No, no, no. That was 2022.
The companies that win from here are going to invest in a business harness. Think of it like an executive team turned into software. A set of rules, goals, and tools all wired together on top of AI. Your best people use it as the dashboard they work from every day.
It might take three years to hit its apex. But starting now is what keeps you pointed the right direction. By the time it matters, your legs are already strong and you're pacing ahead.
If you don't? I think it's a BlackBerry moment.
Business Owner
Huh. Maybe we should talk.
Outsider Labs
Yeah, we can. But we're booked into 2027. We're a small elite team. We only work with a handful of companies at a time. What we can do is run a high-level audit and get you scheduled.
The Outsider AI Audit
The 8 tests you need an outsider to run.
An 8-test diagnostic for operating businesses with $5M–$100M in revenue. Two to three weeks. One kickoff. Async data collection. One findings call. A written report sealed by Harry and Anthony. Built to be read by a board in 60 minutes.
— Foundation
Can this company actually run agents against its own work?
Test 02
Agent Governance Panel
Seven control questions every agent deployment needs to answer before anything goes into production.
Test 03
Role + Agent Ownership Map
Which human roles are most exposed to AI displacement, and who is the named owner for every agent the company runs?
— Defense
What does AI take from this business if you do nothing?
Test 04
Customer Disruption Map
How exposed is this company's revenue to AI disruption of its own customers' industries?
Test 05
Pricing Power Audit
What does this company's revenue look like in 24 months when customers' willingness-to-pay compresses 40–70% due to AI?
Test 06
Sellability Marker
Where does the business trade on the AI-readiness premium-to-discount spectrum, and what are the three moves to migrate up before exit?
— Offense
Where does AI hand you margin and throughput, right now?
Test 07
Margin Expansion Map
Where can AI cut cost or expand throughput in the next 12 months — function by function, and counterparty by counterparty?
Test 08
Workforce Architecture
What does the next 12 months of hiring look like through the AI lens — what to hire, what to pause, what new roles nobody is creating yet?
Two partners. One financial lens, one technical. Every audit sealed by both.
Partner · Financial & Transaction Lens
Harry DeMott
Thirty years across equity research, hedge funds, and venture capital. Nine years at First Boston (now Credit Suisse) covering media and technology. General Partner at Raptor Ventures. Board member at Pandora from its earliest days through 2011. Co-founder and CEO of Proper. Princeton, BA Economics. NYU, MBA Finance.
Harry leads the financial and transaction lens of every audit — customer disruption, pricing power, and sellability.
Partner · Technical & Operational Lens
Anthony Batt
Entrepreneur and executive with deep roots in digital media, SaaS, and operational technology. Built and scaled venture-backed companies. Served as a CTO-for-hire across multiple businesses. Founded and led teams at Buzz Media and Proper. Knows what a functioning data and AI stack looks like from the inside.
Anthony leads the technical and operational lens — data readiness, agent architecture, and margin expansion.
We specialize in Anthropic's Claude and Google's Gemini — the frontier systems we build these harnesses on. We work with a handful of companies at a time. If you think you're a fit, write us.
Think you're a fit? Write us.
We're a small team and only work with a handful of companies at a time. The audit is the front door. Three sentences about your business is enough to start.
- Fit
- Operating businesses, $5M–$100M in revenue.
- Frame
- Two to three weeks. One report. Sealed by both partners.
- Bench
- Booked into 2027. Audits slot ahead of full engagements.
- Stack
- Anthropic Claude · Google Gemini. Frontier only.