
From Our Sponsor:
You’re optimising for AI visibility. But your analytics can’t even see AI.
You’re doing the hard work. Tweaking content, building authority, showing up in ChatGPT and Perplexity results. But here’s the blind spot: when someone clicks through from an AI-generated answer, Google Analytics has no idea what just happened.
It logs it as “direct.” Same bucket as someone typing your URL from memory. No source. No attribution. No trail.
Which means right now, you have no way of knowing whether your AI visibility efforts are actually driving revenue or just driving vanity metrics in an AI chat window.
Stape’s free Website Tracking Checker shows you in seconds how much signal your current setup is actually capturing. Enter your URL, get a health score, see what’s leaking.
Guide: How to Start With Your First Personal AI Agent

Source: Instinct
At an AI conference I attended recently, one name kept coming up in conversation after conversation: Instinct. People were talking about how it now runs a big chunk of their personal lives, from booking appointments to chasing refunds to sorting travel.
It’s made a lot of noise, so I decided to dig deeper. I read the terms, the privacy policy and the early reviews. The result is a simple myth-busting guide for anyone who wants to experiment with Instinct: what it is, how to get started, and what you’re actually agreeing to.
This isn’t just for ecommerce operators. It’s for anyone curious about testing a personal AI agent.
Understanding Instinct AI
Instinct is an AI personal assistant that you text through iMessage or WhatsApp, or call. Unlike ChatGPT or Claude, which mostly answer questions, Instinct acts on your behalf. It books, buys, cancels, chases refunds and fills in forms.
Why everyone’s talking about it: it’s the first agent that non-technical people can actually use. There’s no app to learn and no setup beyond entering your phone number. It’s also proactive and persistent: it will chase an unresponsive company for days until it gets an answer. The traction backs this up. Instinct raised a $250M Series B at a $2.5B valuation in August 2026, is reportedly in talks for a round at around $10B, and has more than 100,000 reported users. It’s currently invite-only and free during the beta.
Here’s how it works:
Its own identity. Each Instinct gets its own phone numbers and email address, so it can contact businesses as your assistant. You can also CC or forward it into email threads.
Connectors. It links to Google Workspace, Outlook, Notion, Slack, GitHub, Granola, Linear, and Link for payments. Availability varies by account.
The Vault. This stores your logins (including 2FA), payment cards, addresses and phone numbers.
Computer use. A cloud computer operates websites for you, such as checkouts, booking forms and government portals.
Memory. It remembers context across conversations and follows up without being asked.
What You Can Use It For
Instinct is built for personal life admin. Here’s what early users are handing over to it.
Bookings. It handles restaurant tables (including card-on-file holds), haircuts, massages and even local council recycling slots. It can also cancel and rebook when plans change.
Travel. It does flight check-ins, loyalty programme sign-ups and visa-on-arrival forms. In one case, it pulled three months of travel history from email and Airbnb receipts to complete a visa form.
Chasing. This is where it shines. One user had it email a company’s executive team every day until someone called back. Another had it fight a ticketing platform’s no-refund policy, and it won.
Money admin. It can find subscriptions to cancel, chase refunds, negotiate renewals such as car insurance, and follow up on invoices.
Shopping. It researches options, compares prices and places orders.
Coordination. It can plan a party end to end, including the venue, schedule clashes and catering quantities. With the new Trusted People feature, it can also sync itineraries and calendar entries with a partner’s or family member’s Instinct.
Calendar. It adds bookings automatically, blocks time for tasks and flags clashes.
What You’ll Need
An invite link from an existing user (I’ve included one)
To be 18 or older
iMessage or WhatsApp
A virtual card with a hard spending limit
A personal email and calendar (not your work accounts)
Step 1: Get a Legitimate Invite
Only sign up at instinct.com or app.instinct.com. Type the address in directly, because knockoff sites already exist.
Never pay for an invite. You’d be letting a stranger set up an account that will hold your email and cards.
Never share your sign-up SMS code. That code is your account login.
Step 2: Sign Up
Go to instinct.com and click Text Instinct to get started.
Enter your phone number and tap Send code. This also opts you into marketing texts. Replying STOP later turns those off, but it stops your SMS login codes too.
Enter the code, then your name and email.
Add your invite link if prompted. Missing this step is the most common reason sign-up stalls.
Choose iMessage or WhatsApp, then scan the QR code.
IMPORTANT: Step 3: Lock Down Privacy Before Connecting Anything
Turn off “Improve Instinct for everyone” under Preferences → Data controls. It’s on by default, and the opt-out only applies from the moment you switch it off.
Bookmark app.instinct.com/workspace. That’s where you delete indexed data.
Opt out of arbitration by emailing [email protected] within 30 days of signing up, if you want to keep your right to go to court.
Step 4: Connect Apps Selectively
Start with your calendar and a personal email account only. Data from Google Workspace isn’t used for training. Data from Outlook, Notion, Slack and other non-Google connectors can be.
Step 5: Set Up the Vault With Guardrails
Cards: use a virtual card capped at around €100–500 a day. Never use your main debit card or an uncapped credit card.
Address and phone: add these to reduce back-and-forth.
Logins: store low-risk accounts only, such as food delivery. Never store banking, email, marketplace or business accounts.
Vault data isn’t used for training, but chat is. Don’t send passports or IDs in messages.
Step 6: Set Standing Rules and Start Small
Give it standing instructions such as:
“Always confirm before paying more than €X.”
“Never accept non-refundable terms without asking me.”
Then give it one small first task, like booking a slot. A few habits help:
Talk to it like an assistant, not a search engine.
Be careful with emoji. A thumbs up on a confirmation request counts as a yes.
Be patient. If it goes quiet for 5–15 minutes, it’s usually still working.
Step 7: Check Its Work
The terms say Instinct’s own records of its actions “may not always be accurate.” Review confirmations, calendar entries and card charges weekly.
Tailor the Setup to You
These guardrails aren’t one-size-fits-all. Adjust them to how you’ll use Instinct and how much risk you’re comfortable with.
If you are... | Use it for | Keep it away from |
|---|---|---|
An individual | Bookings, travel forms, refunds, cancelling subscriptions, planning events | Your main bank card, and sensitive documents sent in chat |
A business owner | Your personal life admin | Business email, Slack, Notion and client data. The licence covers personal use only. |
An ecommerce operator | Your personal life admin | Seller Central, Shopify admin, supplier inboxes and ad accounts |
Start with a tight spending cap and a short list of connected apps, then loosen only once you’ve seen how it behaves.
For ecommerce operators, there’s a specific reason for caution. The terms make you responsible if the agent breaks the rules of the platforms it touches. Amazon has already blocked a rival agent, Muse, from shopping on its site.
Myths vs Reality
With great power comes great responsibility. An agent that can book, pay and sign up on your behalf can save you hours, but only because you’ve given it your inbox, your logins and your cards. That access comes with trade-offs most people don’t read about before they say yes.
Before you hand it anything important, consider:
What happens to your data after you disconnect an app.
Whether your data is training their models.
Who is legally responsible when the agent gets something wrong.
What protections you have if you’re outside the US.
Here’s how the common assumptions stack up against what the terms actually say.
Myth | Reality |
|---|---|
“Disconnecting an app removes my data.” | Instinct can keep using indexed data until you separately delete it. Users reported getting email summaries after disconnecting Google. |
“Opting out stops all training.” | The opt-out only applies going forward. It covers model training, not “other types of improvements,” and data flagged for safety review can still be used. |
“If it makes a mistake, Instinct is responsible.” | You carry the consequences. Its liability is capped at $100, and you indemnify the company. |
“It just clicks buttons for me.” | You appoint it as your legal agent. Terms it accepts on your behalf bind you as if you’d agreed to them yourself. |
“My EU data rights apply.” | The privacy policy has no GDPR section, and the service is governed by California law. |
“It only does what I tell it.” | The privacy policy itself warns that malicious emails or webpages may try to hijack agents through prompt injection. |
It’s also worth knowing that your data is shared with unnamed AI providers, business partners and any future acquirer. Instinct can also terminate your account for any reason.
How to Leave Cleanly
All three steps are needed:
Disconnect your apps.
Delete your indexed data at app.instinct.com/workspace.
Delete your account.
The Bottom Line
For booking, chasing and form-filling, Instinct genuinely saves time, and it’s the easiest way to see what a personal AI agent can do. But it works because you hand over your inbox, logins and cards, and the terms put almost all the risk on you. Treat it as a life-admin assistant with guardrails, and keep your most important accounts out of it.
Do You Love The AI For Ecommerce Sellers Newsletter?
You can help us!
Spread the word to your colleagues or friends who you think would benefit from our weekly insights 🙂 Simply forward this issue.
In addition, we are open to sponsorships. We have more than 70,000 subscribers with 75% of our readers based in the US. To get our rate card and more info, email us at [email protected]
The Quick Read:
Adobe forecasts $275.1B in US online holiday sales for 2026, up 6.7%, with Cyber Monday topping $15B. AI traffic to retail sites is set to jump 130%, peaking at 159% on Thanksgiving.
Prime Big Deal Days drives $9.86B in US online sales, per Adobe. Numerator finds over a third of shoppers did half their holiday shopping early, squeezing Black Friday and Cyber Monday.
ChatGPT ads draw 1,400+ advertisers, but brands aren’t betting holiday budgets on them. Measured says CPCs run about 6x Google’s and early results are poor. A testing ground, not a channel.
YouGov finds 34% of likely Black Friday shoppers plan to use AI, up from 16% last year. But only 5% would let it complete a purchase, even with approval. Trust stops at the basket.
Kevin Indig argues 2027 plans built on traffic targets are doomed, with zero-click searches at 68% and AI underattributed up to 10x. Budget for mentions, citations and community instead.
Claude for Google Workspace enters public beta on paid plans, adding a sidebar in Docs, Sheets and Slides that edits files in place. Your connectors and skills come along too.
Major AI labs now invisibly watermark generated text, driven by EU rules.
GEO is becoming a real revenue line for publishers, with Future signing 30+ clients and Ziff Davis bundling it into branded content. Measurement remains the unsolved problem.
AI agents will drive $100B in sales within 12 months, and 32% of those surveyed already let AI act without final approval. Retargeting and cart emails lose grip.
The Tools List:
📖 NotebookLM - Source-grounded research assistant that answers only from your uploaded docs.
🗣️ Cartesia - Ultra-low-latency text-to-speech and voice cloning for real-time apps.
🌐 CodeRabbit - AI code review that leaves context-aware feedback on pull requests.
🧑🏫 Gamma - Generate polished presentations, docs and sites from a prompt.
About The Writer:

Jo Lambadjieva is an entrepreneur and AI expert in the e-commerce industry. She is the founder and CEO of Amazing Wave, an agency specializing in AI-driven solutions for e-commerce businesses. With over 13 years of experience in digital marketing, agency work, and e-commerce, Joanna has established herself as a thought leader in integrating AI technologies for business growth.
For Team and Agency AI training book an intro call here.




