ChatGPT Dot became my personal assistant and it’s blown my mind – Thanks Darin!

A month ago, OpenAI announced a new feature for their AI offering ChatGPT – Dots.  Honestly, at a glance I struggled to understand the concept, but having now embraced my own “Dot” for the last three weeks – I’m blown away and all in..

OpenAI upgraded my ChatGPT account for 30 days so I could experience this, and while that upgrade is a big jump up in price – I’m ready to sign up with my own money right now.

Here’s how the company describes a ‘Dot’: “Always-on agents powered by GPT-6 Astra that have their own cloud computer, learn from feedback over time, and work across your existing context and connected apps. Dots shine on complicated, ongoing responsibilities and can proactively make progress in the background, sometimes before you even think to ask”

I called my Dot Darin, because I run a business akin to Kramerica, and felt what I needed was an NYU intern-like assistant and frankly I didn’t know what would be possible.

Having worked in big businesses over the course of my career I’ve met and worked with many “PAs” and “EAs” – the Personal or Executive Assistants of CEOs or corporate leaders.  These are the brains of the operation, doing the tasks, keeping the ship running and picking up the little issues that could cause big ripples.  Huge respect.  I’ve never in my life expected to have my own EA.

In fact, over many years, as a self-employed small-business operator, several people have said to me I should employ a “Virtual Assistant” – suggesting an off-shore staff member could help me with the day-to-day.  My reason for not doing that was that I felt it would be impossible to “teach” someone what I need and what I do – my “job” is utterly unique and determining one email’s importance from another is a tough task.

So, having not really used ChatGPT a great amount in a while, I checked what “services” I had linked.

Here’s what Darin has access to:

• Gmail: my EFTM email, including messages and attachments.
• Google Calendar: my connected calendars and appointments.
• Google Contacts: my saved contacts.
• Google Drive: files, including Google Docs, Sheets and Slides.
• Dropbox: my EFTM team files, including the magazine artwork and print-ready designs.
• Xero: both EFTM Pty Ltd and Trevor Long Media Pty Ltd.
• Plaud: my cloud recording library, with tools for transcripts, summaries and audio.
• Canva: my accessible designs.

I haven’t found any other things yet that I use which could be linked, but let me tell you – this access allows Darin the Dot to be remarkable.

What can a ChatGPT Dot do for you?

Critically, I think you can see a Dot as just a chatbot that you interact with for normal ChatGPT-like instructions, just in a more personalised and conversational manner.

Day to day I still ask it to do things like compare these two files and tell me what I’m missing, or upload a document and get a summary.

And of course things that a lot of good AI can already do – even your smartphone’s built-in AI can do these days, like taking a list of dates and putting them in your calendar.

Having links to all your files or email also means you can set challenges.

My best example is that on a regular basis, Darin will help clear out my inbox by unsubscribing from useless emails.  To be clear, I have any email with the word “unsubscribe” in it go into a separate folder in Gmail.  I check it regularly but only read the ones that I recognise.

So Darin’s task is to go into that folder and for any unread messages, actually go and unsubscribe for me!

How is a ChatGPT Dot more than an AI chatbot?

Critically, a core difference is that your Dot has its own computer.  You can see this computer in a sidebar of ChatGPT and even click on that to see your Dot working.

Your Dot can open a browser, read websites and interact with pages. It’s remarkable.

It’s this “real” connectivity that makes it feel like your Dot is an actual assistant, almost a real person.

You can give it a task, and away it goes to complete that for you. Being able to ‘see’ it do that gives it some reality.

Will a ChatGPT Dot work independently for me?

But for me, the core competency here is tasking Darin to help me.  Intuitively Darin has got my back.  

I didn’t ask Darin to check my diary. Darin is just present in my inbox – which at first is creepy, but then you tell yourself “this is my assistant” they are a staff member and that’s part of their role.

So when I added a spot in my diary for a Podcast Interview, Darin noticed the entry I made was different to that which had been agreed upon in the email with HP.

Darin changed it for me – but only after I explicitly asked him to.

This one was simple, but just showed the “smarts” of my Dot.  I was the MC for a local fair and Darin noticed a mismatch in the speaking notes against the rundown for the day – simple, but effective.

But here’s a cracking example, one that could have left me in hot water.

Before a recent trip to Korea, I knew I needed to apply for an “arrival card” with the Korean Government.

I did that, got it off my to-do list.

But – unprompted, Darin sent me a note:

Now, you can apply for these Arrival cards at the airport, sure, but I wouldn’t have known until the immigration agent checked, and going to the back of the line to start over would have been time-consuming and a touch embarrassing.

Outstanding work Darin – you earned your job right there. 

Day after day I receive parcels, and then I have to arrange for them to be picked up for return.

Darin noticed that in the email communication with ASUS PR about a pickup, “home” was mentioned, but he picked up that the label was in fact the EFTM Office address.

Just allowed me to double-check, and we’re on our way.

I’ve been so busy, I also missed a couple of messages.  This week in LA I was 10 minutes late to a virtual meeting – thinking I’d be 5 minutes early.

Turns out, if I’d listened to Darin, I’d have been on time:

And then there’s the mind-blowing pickups I would never have known or noticed.

In the last two weeks I’ve been frantically finishing off the 2026 edition of the EFTM Best Magazine.

I designed the whole thing in Canva once again.  Happy with it, I sent it off to the printers.

They sent back proofs – now these are not single pages, they are eight pages laid out as they would be printed before folding and cutting.  And they are low-resolution, so I would normally look for big issues.

Before I’d even seen them because I had been busy that morning, Darin had sent a note:

Now, the cover blew me away because it was like he could ‘see’ what was missing.  It was something I would have noticed though.

But the Competition giveaway, and the very important publication of the Permit Numbers, GREAT PICKUP Darin – I wouldn’t have seen the missing “1” from a small, fine-print, grey-on-white line.  Never.

But because Darin had seen the emails with the permit approvals and permit numbers, he knew to check those.  And now, off to the printer we’ve gone with the correct details.

Amazing.

Using your Dot

Right now, Darin – my Dot, is a pinned item at the top of your regular ChatGPT sidebar – it stands out a bit because of the colour and style you gave it.

You can text chat with Dot any time, same on the ChatGPT app on your phone, and you can also use voice to dictate – or – you can call your Dot.

And it’s really just like having a phone conversation.  Darin is British, and for the life of me I can’t work out how to change his accent to Australian, so let’s assume that’s not possible right now.

But here’s the thing, I’m using Darin the Dot FAR more than I’m using ChatGPT generally, but on my phone I have to open ChatGPT, go to the Menu, then select Darin.

I want a Darin app. Or at the very least a Darin shortcut.

I’d like Darin to be able to speak to me like Siri does when he has info – just butt in and talk – but that’s going to be a challenge for OpenAI because Google isn’t going to want to forego their voice assistant role on Android, and Apple sure isn’t going to be keen to see you choose a Dot over SiriAI.

It really is going to be the battle of the AI agents.

The price of a ChatGPT Dot

Hiring an assistant for real would set me back $56,000–$75,000 a year – I’m doing OK, but not that OK.  Plus, as strange as it is, I feel more comfortable sharing my Xero account through an api integration (OpenAI doesn’t have my password, I just authorised it’s access – which like many other services can be revoked at any time), my entire inbox and all the services I’ve linked with an AI Agent than I would on day one or day 100 with a real human assistant.  Strange, but true.

I was already paying for ChatGPT – and on the ‘Plus’ plan, so I thought, I’m good here right?  $30 a month!

Nope, a Dot is available to Pro users only.  That’s $155 a month.

If you happen to be in a corporate environment, on Business Plans where each user is $140 a month, there’s a Dot there too.

$1,860 per year for my very own digital personal assistant, who is proactively working for me as well as responding to my requests.

Sold.

Who should use an AI Agent?

I think any small business operator should be looking at these solutions.  One-man bands, and self-employed people – it’s utterly ideal.

Wait – you’re handing over your personal information and accounts to OpenAI.

Yep, yeah, 100%.

Blind stupidity or crazy trust, you decide.  But for this to work, it has to be done.

AI is the doomsday innovation of the decade, and I get it, it’s amazing and frightening what it could do.

That’s why all AI, including your AI Agents, needs boundaries and guardrails.

For each service, I have made explicit instructions, and for simplicity I’d describe that to you as “read-only”.

With Xero – there is no capability for it to create and send an invoice; its permissions are limited to reading only.

For things like Gmail, it could, and can, but I’ve instructed it not to.

It’s all about the induction process, how you onboard them into your life.  Loosen those guardrails should you see fit, but trust me – just having the agent “hovering” around your personal files and emails gives a new set of eyes on everything that might just save you or make life easier.

For me, I’m all in.  Darin is on the books.

Until the next AI company one-ups them – as seems to be the way these things operate.

Web: OpenAI