Some days arrive one at a time. Others arrive all at once, and they do not knock. I want to share one personal story about feeling overwhelmed, because I think every business owner and every busy professional will recognise it, even if their version has fewer paws and fewer piano keys.

It was the period just after COVID, when we were all learning how to live in society again, enjoying our freedom and moving everything back into place. Here in Australia we have kerbside collection days, when you can put the things you no longer need out on the kerb and have them collected. (For my American friends: think bulk trash pickup on the curb.) I missed the day by one day. So I had to take my rubbish across to another neighbourhood and leave it in front of a friend's house, because he said he still had some space and it was okay to use it. I spent the whole day doing that and organising it. It was a very tiring day, just getting everything done.

Once I finished, the breeder of my cattle dog, Tilo, called to say that Tilo was ready and I could collect him that day. If I didn't, it would have to be the following week, because she was going away and all the dogs would be with somebody else. So I said okay. I drove two hours to her farm with my son, and it was quite an adventure.

When we came back home, my partner was standing with a delivery driver. We had just bought a piano, and the piano was arriving that day. So I had a new dog, I had just removed a lot of rubbish from the house, and a new piano was coming in, with people who had to assemble it. I had to help, because moving a piano into a house is not an easy task. Meanwhile the new puppy was inside, so I put him outside for the moment, and we got the piano in.

Then I sat down at the table, exhausted, just to eat something, and my partner told me she was pregnant. We were going to have our third child.

I felt so much emotion that day: a mix of tiredness, worry and happiness. And I thought to myself, I need to hire some people to help me out. I am one of those A-type players who believes he can do everything himself. I don't like to ask for help; I just keep things going. That day was a humbling day for me. In front of God, I asked for help for the season ahead, because I knew it was going to be a challenging one. It was a small prayer in my heart.

The baby, the puppy and the piano

I still remember that day as the most overwhelmed I have felt. After clearing out all the rubbish, that is what I got: a new piano, a new dog and a new baby, all at once. And every one of them is great. The piano is fantastic and made my partner happy. The dog has given our family joy. And of course the baby is our third child, whom I love dearly. It was a good day, yet I could not really enjoy it, because I was so overwhelmed.

That is the metaphor I want you to keep. In business, the baby, the puppy and the piano are the good things: the big client, the new hire who needs training, the opportunity you have wanted for years. They rarely come one at a time, and they never check your calendar first. The rubbish is everything you are still carrying yourself.

With that in mind, here are four lessons I took from that day, for anyone who runs a business or a busy career.

01Admit the season is coming

The hardest part of that day was not the piano. It was admitting that I could not carry everything on my own. Many owners and high performers wear "I'll just do it myself" as a badge of honour. It works, right up until three good things arrive on the same afternoon.

Asking for help is not weakness. It is the moment you stop running the business from memory and willpower, and start running it on purpose.

Insight: if you feel overwhelmed by good news, that is not a sign you are failing. It is a sign the season has changed, and your way of working has to change with it.

02Build the system before the piano arrives

Notice where my day really went wrong: I missed the kerbside collection by one day. One missed date turned a simple job into a whole day of driving rubbish across town. Most overwhelm in business starts the same way, with something small that lived only in somebody's head.

A system is simply a decision you make once, so you don't have to make it again on your worst day: a shared calendar with the collection dates on it, a checklist for onboarding a new client, a template for the proposal you write every month, a standard reply for the question every customer asks. Clear the kerb, then clear the calendar.

Insight: write down the three jobs you repeat every week and turn each one into a checklist or a template this month, before the busy season, not during it.

03Play to your strengths, delegate the rest

When the piano came in, I did not try to assemble it alone. There were people who knew how, and I helped where I was useful. That is delegation in one picture.

Be honest about the areas where you are good and the areas where you are not so good. Your strengths are where your time earns the most, and the rest deserves a person, a partner or a process. Delegating is not giving up control. It is choosing where your attention goes.

Insight: list your week's tasks in two columns, "only I can do this" and "someone or something else could". Hand off one item from the second column every week.

04Let AI take what it does better

Here is the humbling part for an A-type professional: maybe an AI is going to do a better job than you at some of the work. Not the judgement, the relationships or the care, but the sorting, drafting, scheduling and summarising that quietly eat the day. An assistant can triage your inbox, draft the first version of a proposal, turn meeting notes into actions, keep track of dates you would otherwise miss, and pull research together while you are on the road for two hours collecting a puppy.

Start with one bounded, repeatable job, check its work for a few weeks, and widen its role only as it earns your trust. You stay responsible for the result; the machine just carries more of the weight.

Insight: pick one weekly task you dread and give it to AI with clear instructions and a human review. The hours you win back are what let you enjoy the good days.

In short

Overwhelm often arrives with good news. Admit the season is coming, build your systems before you need them, play to your strengths and delegate the rest, and let AI take the work it does better. Then, when the baby, the puppy and the piano arrive on the same day, you can actually enjoy them.

"Two are better than one, because they have a good return for their labour." — Ecclesiastes 4:9

So, what is your baby, your puppy and your piano this season, and what will you hand over before they arrive? What are your insights?

When I was a boy in Brazil, my grandfather Alan used to take me to the newsstand. I went for the football stickers. He went for the world. He read voraciously, newspapers and magazines alike, and he told me that while travelling for his doctorate he had come to love global titles such as TIME because they gave him a wider view of what was happening. I did not know it then, but he was introducing me to the internet of his day: a kiosk where the news arrived fresh every morning.

Most of today's AI assistants are not newsstands. They are encyclopaedias: vast, articulate and impressively cross-referenced, but printed on a particular date. Ask one about this morning and it will answer with the confidence of a volume that went to press last year. This is the first essay in a short series on Grok at work, and it starts with the quality that Grok's makers put at the top of the pitch: the newsstand is open.

The AI coursework I have been studying explains why this matters. A language model predicts the next word; by default it neither remembers your business nor searches the web. It knows only your prompt and its training data, and when there is a gap between the two it fills it with what seems likely. Call it the context gap. The first remedy for missing recent information is simple: switch on search. The same lesson applies when choosing a tool. When you need the latest news, pick one that can reach live sources.

Grok is built around that idea. Its makers say it "searches the web and 𝕏 live, so answers always reflect what's happening right now — not last year's training data". That is a strong claim, and a useful one, provided you know how to read the paper. With that in mind, here are four things every business should understand before relying on real-time AI.

01Check the printing date

Every model has a knowledge cut-off, the point at which its training data stops. The original GPT-3.5, for example, knew nothing after September 2021. For many tasks that hardly matters. For a price, a competitor's launch, a regulation, a supplier's outage or this week's mood among your customers, it is decisive. An answer can be fluent, well structured and simply out of date.

Live search changes the question you should ask. It is no longer "does the model know?" but "is this question about now?" If it is, the assistant must go and look, and you should be able to see that it did.

Insight: before you trust an answer, ask whether its truth depends on the date. If it does, insist on a live search and a dated source.

02Know which shelf you are reading

A good newsstand has more than one shelf, and so does Grok. Its developer documentation lists built-in tools for web search, which searches and browses pages in real time, and X Search, which runs keyword and semantic searches across posts, looks up users and fetches whole threads. Searches can be narrowed to particular sites or handles, to date ranges, and can even read images and video.

The two shelves answer different questions. The web is where the considered record lives: official announcements, price lists, filings, reviews. X is where the conversation lives: complaints, praise, rumours and the first reports of events still unfolding. Sentiment analysis, gauging how people feel about a brand, is already a mainstream business use of AI. Real-time access turns it from a quarterly report into something closer to a pulse.

For a marketer, that means watching how a campaign lands on the day it launches. For a café owner on a busy Brisbane street, it might mean noticing that a road closure or a heatwave is being talked about before the lunch rush. For a sales team, it means arriving at a meeting knowing what the client said publicly yesterday.

Insight: decide which shelf you need before you ask. Use the web for the record, X for the reaction, and narrow both with dates, domains or handles so the answer stays on topic.

03A headline is not a fact

Speed has a price. The newest information is also the least checked, and a trending post is not the same thing as a true one. The practical advice that applies to any assistant is simple: ask it to show its sources before summarising, ask for direct quotes with citations, ask what the conflicting evidence is, and for important decisions open the links yourself.

Grok's design leans the same way. Its product page promises "live citations from primary sources", and its developer cookbook includes a fact-checking pattern that splits a post into individual claims, checks each against the web and X, discards any source the searches did not actually return, and answers "not enough evidence" when nothing supports a claim. That last option deserves to be copied. A good research assistant, like a good journalist, is allowed to say it does not know yet.

The company's own terms are blunter still. Output "may not always be accurate", they warn, and users are responsible for evaluating it, "including using human review and supervision".

Insight: treat a live answer as a lead, not a verdict. Click at least one citation, prefer primary sources, and reward an assistant that admits uncertainty.

04Not every question needs today's paper

Real-time knowledge is a tool, not a virtue. Drafting a policy, explaining a concept, summarising your own contract or brainstorming product names does not require this morning's news, and dragging live search into such tasks can add noise. Searching also costs something: X Search, for instance, is billed for each batch of posts it fetches, on top of ordinary usage. For questions about your own business, the better source is usually your own documents, as the first essay in this Journal argued.

The real advantage comes when freshness is combined with routine. The developer cookbook shows a daily briefing that gathers chosen topics from X and the web each morning, remembers what it has already covered and stays within a budget. Grok Bot, launched in beta in August, goes further: always-on agents that work in your tools and "only come back when something needs your approval". That is the subject of the next essay.

Insight: use live search where the answer can change by the hour, use your own documents where the answer is yours, and automate the morning reading once you trust it.

In short

Most AI assistants are encyclopaedias with a printing date. Grok's distinctive promise is a newsstand: live answers from the web and X, with citations. To profit from it, check whether your question is about now, choose the right shelf, treat headlines as leads rather than facts, and save the newsstand for the questions that need it. Freshness is an advantage only when it is paired with verification.

"Output may not always be accurate. Output from our Services is not professional advice. You should conduct your own thorough research and should not rely on Output as the truth." — from Grok's consumer terms of service

It is a candid line for a company to publish about its own product. It is also exactly what my grandfather would have said about any newspaper: read widely, then think for yourself.

So, what is the question in your business that cannot wait for next year's encyclopaedia? What are your insights?

Picture the owner of a small events business on a Monday morning. She asks a chatbot to help her organise a conference for 120 people, and within seconds it produces an admirable plan: confirm numbers, fix a budget, shortlist venues, check dates, book catering. Then it stops. The plan sits politely on her screen, and the real work — ringing venues, comparing quotes, chasing replies, paying the deposit — is still hers.

For most of its short public life, generative AI has behaved like an oracle. You bring it a question, it gives an answer, and then it waits for the next question. The next wave is different in kind, not just in degree. It does not simply tell you what to do. It goes and does it. It is the difference between asking for directions and handing over the keys.

IBM's course on generative AI draws the line crisply. Generative systems are reactive: they wait for a prompt, produce content from patterns learned in training, and their work ends at the point of generation. Agentic systems are proactive: they usually start from a prompt too, but then pursue a goal through a series of actions with minimal human intervention. The oracle answers. The operator acts.

That sounds like a technical footnote. It is not. Handing over the keys changes what these systems are worth, what can go wrong and, above all, where people stand. With that in mind, here are four things every business leader should understand before giving the keys to an agent.

01An agent works in a loop, not a line

A chatbot's work is a line: prompt in, content out. An agent's work is a loop. The IBM material describes four steps — perceive the environment, decide on an action, execute it, learn from the result — and then round again. IBM's introductory AI course defines agents in the same spirit, as software that interacts with its environment and performs tasks on its own.

What makes the loop possible is that the large language models that write fluent prose can also plan. Asked to organise that conference, an agent uses the model to talk itself through the job: first the requirements (size, duration, budget), then venues that fit, then which of those are free on the day. Researchers call this "chain-of-thought" reasoning, and the course describes generative AI as the "cognitive engine" behind an agent's decisions. In 2022 a team from Princeton and Google showed, in a method they called ReAct, that alternating reasoning with actions — think, search, check, think again — produced more grounded results than reasoning alone, because the model could test its thinking against the world rather than against its own memory.

Insight: a loop compounds. A good step makes the next one easier; a wrong fact early on gets acted upon later. Judge an agent on the whole journey, not on one clever answer.

02A good driver needs a reliable map

An operator is only as good as the facts it works from: your price list, your refund policy, last week's emails. A model on its own knows only what it absorbed in training, which may be out of date and is never specific to your business. The remedy, which IBM's introductory course covers, is retrieval-augmented generation, or RAG: before answering, the system fetches the relevant documents and writes with them in hand.

The technique was set out in 2020 by researchers at Facebook AI Research, University College London and New York University. They highlighted two practical virtues. Knowledge can be updated by swapping the documents rather than retraining the model, and the retrieved sources can be inspected. For a business the second matters more than the first. An agent that can show where it found a fact is an agent you can audit.

Insight: before you connect an agent to anything, decide which documents are the source of truth, and ask it to cite them every time.

03The human moves from the driver's seat to the control tower

Here is the crux. The IBM course contrasts two examples. In the first, a YouTuber uses generative AI to review scripts, suggest thumbnails and compose background music, but a person reviews and refines every output. In the second, a personal shopping agent hunts for stock across platforms, watches prices, handles checkout and arranges delivery, asking its owner for input only when it needs to.

In the first case the human is the author and editor, seeing everything. In the second the human becomes an exception-handler, seeing only what the system chooses to escalate. That is a profound change in the job description. Editors catch mistakes because every draft crosses their desk. Exception-handlers catch only the mistakes the machine recognises as mistakes.

Accountability does not move with the keys. When an agent pays the wrong supplier or sends the wrong message to a customer, "the AI did it" is not an answer a customer, an auditor or a regulator will accept. You can delegate the task; you cannot delegate the responsibility.

Insight: write down the stop signs before you write down the goal. Payments, anything sent to a customer and anything deleted should need a human "yes".

04Start with one journey, not the whole fleet

None of this is a reason for timidity. The dull middle of most jobs — chasing, comparing, copying between systems, following up — is exactly where agents shine, and every hour they take back is an hour for judgement, relationships and creative work. The course's own conclusion is measured and, I think, right: the most powerful systems will be neither purely generative nor purely agentic, but intelligent collaborators that know when to explore options and when to commit to a course of action.

The risks are real, though, and the researchers say so themselves. The ReAct authors warned that connecting a language model to real-world actions brings dangers, such as looking up private information or taking harmful actions, and they deliberately limited their experiments to safe environments. Businesses should do the same: begin with one bounded, measurable task, run it with tight permissions, read the logs, and widen the agent's licence only as it earns trust.

Insight: the most valuable skill of the agentic era is supervision: setting clear goals, reading what the agent did and spotting the odd escalation. It can be taught, and teams that learn it early will move faster.

In short

Generative AI gave us a brilliant oracle. Agentic AI offers an operator that perceives, decides, acts and learns in a loop. To make the most of it, give it a reliable map in the form of your own documents, mark the stop signs before it sets off, and start with one journey. The keys can be handed over. The responsibility stays with us.

"The AI generates possibilities, but the human curates them." — from IBM's Generative AI: Introduction and Applications course

The agentic era asks us to keep the spirit of that line even when we are no longer curating every step.

(MCP: Model Context Protocol, the standard that lets AI tools plug into other apps.)

So, which job would you hand over to your first agent, and where would you put the stop sign? What are your insights?