AI Consultant & Marketer | Get Access to Custom AI Systems That Could 10x Your Revenue, Cut costs & Save Time ⬇️
Joined August 2023
- Tweets9.6K
- Following1K
- Followers5.4K
- Likes28.4K
Claude Duet has one feature I think businesses should steal immediately: it knows when to bother the human.
that's underrated.
bad automation asks humans to do everything.
bad “autonomy” asks humans nothing.
good automation quietly handles the boring 90% and appears when:
money moves
something gets deleted
information conflicts
confidence is low
a customer is angry
a commitment is being made
the goal isn't eliminating humans.
it's making human attention expensive again.
If your business is leaking leads, drowning in admin, or relying too heavily on people remembering what to do next, that’s exactly what we fix.
Tell us what’s happening here:
tally.so/r/mZbV0a
“build me an agent” is becoming the AI version of “build me a website.”
okay.
to do WHAT?
the profitable question is much narrower:
what job does it own?
when does it start?
what systems does it read?
what systems can it change?
what counts as success?
when does a human take over?
“AI agent” is technology.
“every inbound lead gets qualified and followed up within 5 minutes” is an offer.
there's an enormous difference.
If you want us to map out where AI can actually save time, recover revenue, or remove repetitive work inside your business, fill this out and we’ll take a look: tally.so/r/mZbV0a
Claude Duet is basically Anthropic admitting that the future of AI is not another chat window.
it can monitor connected apps.
work in the background on its own virtual machine.
prepare briefings.
build things.
surface recommendations.
then come back to you when an action needs approval.
that's a much more interesting architecture:
machine handles continuity.
human handles consequence.
most businesses currently have the opposite.
humans remember everything.
software waits patiently to be clicked.
GPT 6.1 Terra just got roughly 50% cheaper while becoming substantially better at computer work.
this is the part of the AI race businesses should actually care about.
yesterday:
“having an AI click through this old portal isn't economical.”
tomorrow:
maybe it is.
invoice entry
legacy admin panels
form filling
CRM cleanup
internal tools with no API
the frontier isn't only getting smarter.
previously stupid automations are becoming financially reasonable.
that quietly unlocks a LOT more businesses than another chatbot benchmark.
Anthropic made $16.3 BILLION in six months and still managed to make AI look expensive.
its IPO filing also reveals hundreds of billions in long term compute commitments.
which is a useful reminder:
you probably don't want to be in the business of building intelligence.
you want to be in the business of pointing rented intelligence at expensive problems.
let Anthropic buy the data centers.
you figure out why a company has $400k sitting in unpaid invoices.
much cheaper business model.
If you want us to map out where AI can actually save time, recover revenue, or remove repetitive work inside your business, fill this out and we’ll take a look:
tally.so/r/mZbV0a
I’m opening a few spots for businesses that want a custom AI system built around a real workflow, not another chatbot.
If there’s a repetitive process, sales bottleneck, or operational headache you want off your plate:
tally.so/r/mZbV0a
AI agents have a hilariously normal enterprise problem.
the AI often isn't the thing stopping deployment.
the company is.
bad data.
legacy systems.
unclear ownership.
no audit trail.
no escalation rules.
no idea who is responsible when the agent does something weird.
Gartner's production research literally points to trust and integration as major blockers.
this is why an incredible demo can die three meetings after procurement.
the model worked.
the organization around it didn't.
If you want us to map out where AI can actually save time, recover revenue, or remove repetitive work inside your business, fill this out and we’ll take a look:
tally.so/r/mZbV0a
DoorDash processes roughly 3 million deliveries a DAY.
third party AI agents have placed...
a few hundred orders in total.
beautiful reality check.
people are acting like humans are about to delegate every sandwich purchase to autonomous software.
probably not.
simple tasks already have excellent interfaces.
the interesting agent opportunities are the ugly jobs involving:
research
multiple systems
waiting
follow ups
exceptions
memory
judgement
agents don't need to replace the “order burrito” button.
they need to replace the 17 step process nobody wanted to own anyway.
AI isn't making agencies useless.
it's making certain agency invoices harder to explain.
if AI makes
first drafts
variations
research
repurposing
basic editing
production
dramatically easier...
you can't keep pretending the labor itself is the moat.
the value has to migrate upward.
taste.
strategy.
distribution.
offers.
experimentation.
customer knowledge.
revenue.
AI doesn't necessarily destroy service businesses.
it forces them to become more valuable.
one reason AI pricing feels weird
companies sell the unit the infrastructure understands
tokens.
credits.
runs.
actions.
seats.
the customer does not wake up wanting 40,000 additional AI actions.
they want:
every qualified lead answered
every call captured
every proposal followed up
every customer issue routed
every important exception surfaced
sell the thing the customer can recognize in their business.
not the thing your API recognizes on the invoice.
Google's Gemini 4 Argon can autonomously find, validate and patch software vulnerabilities in its cyber defense setting.
OpenAI is building persistent agents.
Google is turning reusable instructions into Skills.
the direction is becoming pretty obvious.
we spent the first phase of generative AI asking:
“what can this model tell me?”
the next phase is:
“what responsibility can this system reliably own?”
that is a much bigger question.
because once software can own a recurring piece of work, you're no longer just selling intelligence.
you're redesigning the company around it.
automation without observability is just faster mystery.
this is going to become painfully obvious as companies deploy more agents.
if an agent sends the wrong email at 2:14 AM, I want to know:
what triggered it
what context it saw
what decision it made
what tool it called
what changed
and how to undo it
not:
“huh. that’s weird.”
people obsess over making agents autonomous.
I’m increasingly interested in making them easy to investigate when autonomy goes sideways.
coding agents are creating a hilarious new bottleneck
building internal software is getting easier
knowing what should exist is not
you can give somebody an agent capable of shipping an internal tool incredibly quickly
great
what process should it fix?
where does the process actually break?
which edge cases matter?
what should stay human?
what does success mean?
who owns it afterward?
implementation is getting cheaper
knowing what to implement is becoming more expensive