'I can not bring myself to trust anything that has that much access and control. None of the big AI companies have earned my trust.' – colleague in his 50s.
'There goes another couple of million jobs.' – PR agency owner.
'I feel like I've been missing the AI wave.' – 29 year old American executive.
'Wow, that's amazing, but what if it gets hacked?' – design agency owner.
'Impressive, but I have no idea how that could help in my daily work' – same design agency owner.
These are comments about Grok Bot from people in my orbit.
Having written about AI since probably 2010, I too felt like my American friend, that the advantages are passing me by.
Over the years, I've also written about low code/no code process automation, business intelligence, the paradox of privacy where we share our every intimate moment including what we had for lunch, but protest loudly when Facebook knows what we're shopping for.
Then there's the design agency owner, who also happens to be my wife, along with my colleague who is at a similar stage of life.
We're Gen X. His concerns were my concerns up until literally two days ago. As were my wife's questions – how can this help me, because what I do is idiosyncratic and specific to my circumstances (which, inevitably, constantly change).
And the PR agency owner? We've all fretted about AI taking our jobs, and it already changed my career partly in 2023/2024.
My initial experiences
Agentic AI is the stuff, we're told, that actually does work for you, rather than finding information/performing set, specific tasks.
I'd used Grok for the latter for some time, along with Copilot and occasionally some of the other AI services like Gemini. And I've remarked on multiple occasions how quickly we have taken it for granted. So, when Grok Bot became available, my interest was piqued. Seemed too good to be true, and also expensive.
Now, I'd been thinking for some time that I really, really need a PA. I'm older, forgetful, terrible and I do mean terrible at admin, whether personal or professional. I had an idea that I could get someone for maybe $4k a month as a fractional or part timer, to make sure my filing, record keeping, payments, calendaring etc was managed more efficiently.
I asked Grok if Bot could do that. Grok said yes and explained a few things. So I signed up; it is around NZD400 a month, which for a SaaS personal subscription seems a lot.
Within 5 minutes, Grok Bot demonstrated its value. Within a day, it has become indispensable. It has automated the process of setting up a job, filing it in the right place and in the right order, automated my billing, researched and booked accommodation in Las Vegas, sends reminders every morning of everything I need to do, and sends briefings ahead of meetings and other engagements (which it figured out on its own would be useful to me).
It is also a support consultant and has sorted out config issues on my phone and PC that I'd known for months (probably years) needed attention, but I'd never gotten around to; to my shame, this includes backup configs. Yes, the backups were there and being done, but no they weren't optimised. They are now.
In a word, Grok Bot is astonishing. It delivers personal process automations with absolute ease. If it makes an error, correct it, and it stays corrected. It constantly works in the background and – here's the crucial bit – the more you tell it about yourself, the more useful and helpful it becomes.
'I can not bring myself to trust anything that has that much access and control. None of the big AI companies have earned my trust.'
This is the privacy paradox. It's a valid concern, as is 'Wow, that's amazing, but what if it gets hacked?'
Couple of things here: what if anything gets hacked? It's always a concern; with a consolidated app, the gates to the entire kingdom are potentially thrown open. Grok Bot has measures of course, including a 'secret place' for passwords and other sensitive stuff, and passwords etc can be invisible to it if you so choose. But nothing is impenetrable, of course, so the risk won't go away.
If you hire a PA, the help they can provide directly depends on how much they know about you, and how much access they have to your life/systems/applications/bank account/etc. More access, more value.
It's no different here. I realised this within minutes: the more Grok Bot knows about me, the better it contextualises who I am and what I do, and the better it can suggest actions and supports (and it does – it detected my upcoming trip to Vegas, realised I didn't have accommodation for the full time I am there, suggested AirBNBs comparable to what I had booked in the past, asked about my ESTA waiver because it knows my passport recently renewed and then checked and confirmed the status, then researched and booked hotels fitting my budget).
It could do this because I connected AirBNB, Uber, Facebook, X, AirNZ. And then I went through my Google Passwords vault, and connected every other thing I could think of, even my AliExpress and TradeInn and MXStore accounts. If I want a new CK shirt, Grok Bot will go to tradeinn, find options and put them into my basket, then ask me to review to see if that's my style. It goes to Woolworths, does the groceries, then alerts me that the cart is ready for checkout.
For years, we've handed over our data while complaining about it. We also know that context is one of the crucial missing bits in AI. Facebook knows that the more it knows about us, the better it can serve our sordid needs for entertainment and weight loss pills.
Now, the more your AI knows about you, the more helpful it becomes.
'Impressive, but I have no idea how that could help in my daily work.'
This is related, as is my American mate's comment that 'I feel like I've been missing the AI wave.' I too felt that way, and also wondered how anything could automated my seemingly random processes and activities. Low code/no code had 0 appeal for me, because I am not a programmer and even my logic capabilities are somewhat deficient. So, while the world marched on, there I was, stuck in my 1990s habits.
Until I downloaded and tried it. And literally simply told the bot what I do and how I do it, and it came up with suggestions and solutions in minutes. In one fell stroke, both the 'passing me by', and 'I have no idea' were consigned to history. I am instantly on the leading edge, with bots now doing almost every aspect of the crippling admin that has always held me back.
Recently, someone said 'we have a what sucks meeting, and apply AI to what sucks'. I did that in my own life/work, and now there is not much sucks left.
'There goes another couple of million jobs.'
Yes, AI is taking jobs. As mentioned here and elsewhere, it sort of took mine (and yet, here I am). The PR agency owner (and longtime friend) is right, but also wrong. I immediately asked her, do you have a PA? No. Do you need a PA? For the love of god, yes. Why don't you have one? Dealing with people, employment laws, hassle factor and of course, a question of $4k.
Like the introduction of any technology, there will be an impact on employment, possibly deleterious in the short run. But people like me and Ms Agency can now get the support we've always needed, at a dramatically reduced cost. We didn't have a PA before, I do have one now. Everyone can get one, too, without the hassle, and that means perhaps millions of people will now enjoy services that were previously out of reach.
We've seen receptionists largely vanish from the corporate landscape. We also saw knocker uppers replaced with alarm clocks. It happens. It sucks.
But like it or not, it is time to embrace our robot overlords, take the privacy risk or leap of faith, and realise that help, whether from a human or a robot, is far more likely the more they know about you. And whether human or robot, sharing information and access to yourself always comes with risk.
There's one thing, though, that the bot can't do yet that people can. Break your heart. Maybe that will come when the bot gets hacked or goes rogue, but until then an admin free existence beckons.