This might sound like a strange phrase, because AI, at least in the form most people now recognise it, has only really been around for a few years. But there is already AI history, and some moments that have been dismissed in one region or country have been fundamental in another.
One example that I think is well worth reading about, especially if you are interested in how China came to view AI, is the AlphaGo tournament, where an AI beat one of the world’s best Go players.
To many of us who had already seen similar things happen with chess, this felt like another incremental step. Computers had been getting better and better, and sooner or later they were going to beat humans at more complex board games. Chess had already had its Deep Blue moment, so AlphaGo could easily be seen as the next stage in the same thing.
But Go is different.
If you watch some of the videos about the match, particularly those that cover Chinese reactions to it, one move stands out: Move 37. In that move, the AI did something that was classed by many observers as almost alien. Mathematically, it was simply finding a logical gap in a progression. It was a move that worked, and there was no real reason why it should not have been played. But up to that point, it looked completely weird and different.
Because this happened within a subject that was culturally important and widely understood, it helped redefine, to a certain extent, how China saw AI. From that point on, AI was taken far more seriously, contributing to the enormous effort and investment that followed.
I will put a link to a video below, but it is well worth watching.
A delivery that goes on and on, but never quite reaches a conclusion.
Explanation:
I must admit that this is partially borrowed from Rory Sutherland, who has used the word “tantric” in the context of sectors. However, you often see tantric deliveries in projects, particularly regulatory or administrative projects.
A tantric delivery is a project that continues for an extended period but never reaches its climax. It often feels this way because no one is willing to take responsibility, take ownership, or sign off on anything. Instead, the project becomes trapped in increasingly fine-grained paperwork, approvals and process steps.
Also such projects can be caused because of never-ending scope creep or a set of deliverables that change faster than they can be delivered upon
As a result, the delivery never seems to get past its final set of bottlenecks or goes live. It is always nearly finished, always close to being signed off, and always somehow still not delivered.
Disclaimer: As always these posts are not aimed at anyone client or employer and are just my personal observations over a lifetime of dealing with both management and frontline associates.
A slightly odd, old-fashioned practice that I have always favoured as a long-term contractor is that favours, praise and rewards should be reserved for permanent members of staff, not contractors.
As a contractor, you receive your remuneration directly in the form of money, so you should not really need anything additional. Yes, public praise can be nice for your reputation and can help with securing future work, but I have found that it does not need to be displayed openly. Merely having a list of people for whom you have done excellent work and who remember that tends to be the best long-term reward. Public recognition further up the ranks is not really necessary.
Although this may sound like an odd, humble brag or slightly self-effacing view, it is one that has some basis in history. During and before the First World War, some military traditions treated medals and formal rewards as things intended primarily for career soldiers, where they could support rank, status and advancement. The thinking was that such honours were, in a sense, less useful when given to conscripted soldiers who would only be with the regiment for the duration of that conflict.
I have found that the same broad idea can help with contractor and permanent staff relations. Contractors should not really be receiving extra rewards, such as public praise or internal recognition, because they are already being rewarded through their rate and contract. That recognition is often better spent on permanent staff, for whom it can support morale, career development and a sense of belonging.
This is only a personal viewpoint, but it is one that has stood me in good stead.
I am currently sitting in that little used reception area that every giant corporate building seems to have for guests who are waiting for someone to come down and collect them. You know the sort of place: odd potted plants, huge chairs, and the sense that it has been designed to be used as little as possible.
The reason I am sitting here at the start of my working day is that I left my pass for this client on my desk at home because I am an idiot.
While sitting here, I assumed I could simply go to security, explain who I was, identify myself, get a temporary pass printed, and head upstairs to work. However, I was wrong. Those days no longer exist.
Autonomy by individuals, particularly in well-established service roles or roles seen as replaceable, has become massively limited, and it has become noticeably worse over time. Gone are the days when security guards would really know you. I normally know most of the security guards, cleaning staff, and the other support people who appear in the office at the same time as I do, which is usually about 6:30 in the morning. But they are all kept in very rigid boxes, because that makes them easy to replace with no impact on the smooth running of the building facilities.
And here is the nub of it: that also makes them easy to replace with non-humans.
I think this is one of the core and genuine worries about AI. Yes, there will always be things that require a human touch, and there will be things that robots have not yet mastered. But by allowing work to become so regimented, with such fixed boundaries around its deliverables and such an absence of adaptability, we have set ourselves up to be easily replaceable by AI and similar systems.
So if your job has very strict boundaries, very strict deliverables, and does not reward innovation or adaptability, then I think you are in the area of humanity that should be most worried about being replaced. That does not matter whether you are in the service industry, deeply technical, or anything else. If you cannot display humanity or any of its advantages during your working day, and if your work is bounded by strict borders with easily quantifiable inputs and outputs, then you are at real risk.
Historically, that risk meant being replaced by a cheaper human. I suspect that, eventually, it will mean being replaced by a cheaper non-human.
Let us step back and talk about General Electric, some 30 years ago, when the first major rounds of outsourcing were being done and call centre outsourcing was brand new. I started in a call centre doing help desk and support work, in an 800-person call centre in the middle of Leeds, Yorkshire. They were just beginning the first stages of outsourcing to India.
There was joking, of course, and there were practical issues. Postcodes were new to the country taking on the work, and they had to be explained. Training had to be done. Processes had to be documented. All that kind of thing, but it was still happening, and as a freshly hired person, I was worried that this was the beginning of the end when I had only just started, but a very clever person told me something that came back to me recently. Whether you can be replaced by another person, a piece of software, AI, or anything else comes down to a few simple things.
Can your work be wrapped in strict definitions?
Is it easy to encompass and define precisely?
When you do your job, do you simply deliver the average expected result?
That average delivery might not even be fully within your control. If there are exact SLAs to deliver against, and your job is simply to meet them, then you may already be in trouble. If the answer to these questions is yes, then you are at serious risk of being outsourced. Back then, that meant outsourcing to India. Today, it can mean outsourcing to AI, automation, or any number of other things.
So, how do we prevent it? Perhaps we do not. Perhaps the real answer is that we have to work with it.
Back then, I was told to make sure I was better than everybody else. Be value for money. Because regardless of what any company might say, value for money always matters. If you come in, do what you consider an ordinary day’s work, and nothing more, then you are at direct risk from AI. AI is, by definition, an average of what it has learned. If it uses the company’s internal data, it becomes the average of what that company has historically been able to do.
So if you are doing an average job, or even just a consistently good job within a tightly defined structure, you are at strong risk of being outsourced by something that can improve on average delivery. It can improve on cost, dependency, repeatability, and consistency.
The second question is whether your job is easily defined. The stricter and clearer the definition of what you actually do, the easier it is for AI to replace you. With traditional outsourcing, work had to be defined well enough for another person or team to take it on, but there would always be variance based on human understanding and interpretation. With AI, the better the definition, the easier the replacement becomes.
So now we know whether we can be replaced. The next question is how we adapt to the companies using AI to put that into practice.
I think the answer has to be defined at an individual level, rather than by industry or even job title. I strongly suspect that AI, just like outsourcing to cheaper countries, will eventually be able to do a large amount of non-physical, interaction-based, white-collar work. We are not going to win by pretending that will not happen. The Luddites did not stop progress by burning down machinery, and we will not stop this by simply objecting to it. Progress will march on.
So how do we deal with it?
We deal with it by using the default that AI provides as a stepping stone to become exceptional.
That is going to be hard in a number of areas, particularly where there is wholesale outsourcing or where there are hard definitions of what is absolutely correct. Copy editors are having a very difficult time at the moment, because their job is to make something complete and correct. If AI can do that, it is hard to go over and above to prove your value.
But for work that involves any form of soft skill, from customer service to facilities, management, consultancy, or technical delivery, we are going to have to produce better than average. That does not mean giving every hour of your life to various corporations. It means recognising that they will be able to supply the average through AI. You will have to remove the base part of the work and focus your effort on what AI produces, then improve it, shape it, and add value beyond it.
The main difference between historical outsourcing and AI outsourcing is the turnaround cycle. It is also about whether your improvements are learnt and absorbed by the AI.
For example, if your work was outsourced to another country, the comparison was whether you were better at the job than the outsourced team. As you learned and improved, you might have been able to stay ahead. They might have improved too, but there was still a human and organisational delay.
With AI, it may only be able to produce the best average result that all the data in your company can support. But it will constantly play catch up with you. That means you will have to keep re-referencing the baseline it produces. No longer will the question simply be, “Am I the best in the market?” The question will become, “Am I a meaningful percentage better than what the AI is already producing?”
That change will be particularly relevant to white-collar workers. I think they are especially vulnerable to this form of AI learning. But, ultimately, AI is here. The people with power see it as a way of producing better value for money with more consistency.
TLDR;
You are most at risk if you are an average worker doing an average day’s work, or if you are a white-collar worker whose output is easily defined and judged as either correct or incorrect. If there is softness in your delivery, use it. If your work requires judgement, adaptability, empathy, creativity, or context, keep developing those things.
Keep ahead of AI by using it as a starting point, rather than competing with it directly.
This post comes from watching two very different vendors, on two very different projects, handle a familiar situation. A project hits issues, the mood turns, and suddenly the vendor is no longer the most popular party in the room. What stood out was the contrast in how each vendor responded, and more importantly, how one of them managed to turn things around.
There are many times in delivery work where you can find yourself on the wrong side of client sentiment, and quite often it has very little to do with the quality of your work. Budgets change, and what was previously seen as good value suddenly looks expensive. Internal restructures happen, and contracts that once made sense are now resented. People move on, political dynamics shift, and you may have been brought in by someone who is no longer in favour.
Sometimes the issue is entirely external to your delivery. A change elsewhere in the organisation can make a project seem less relevant or less innovative. It is not your fault, but it is still your delivery, and that means you will feel the impact.
In these situations, the reason matters far less than the response. As a vendor or consultant, you are being paid to navigate this, whether it is fair or not. The question is how you handle it.
One option is to double down on delivery. That might mean absorbing additional cost, adding resources, or simply pushing harder to ensure the outcome lands well. In fixed price environments, some of this should already be accounted for, but there are times when you have to take a hit. Larger organisations often don’t opt for this type of response purely on short-term financial grounds, i.e., this year’s bonus. That is understandable, but it can be short sighted. Reputation has real value, and people have long memories. It is worth weighing the reputational cost before reacting defensively. In many cases, it is better to grit your teeth and deliver.
Secondly, do not get drawn into internal politics. As a vendor, your role is not to take sides. Stay focused on both the letter and the spirit of your delivery. Turning overly rigid or contractual in tone can be just as damaging as open frustration. Clients are looking for a quality service, not just a strict interpretation of a contract.
There will always be occasional clients who push too far, but they are the exception rather than the rule. In most cases, if you provide something they can confidently take up the chain and demonstrate as progress, you will remain in a good position.
Another practical point is resourcing. Some organisations respond to struggling projects by quietly rotating out strong people and replacing them with those who will simply maintain the status quo. This rarely helps. If anything, it reinforces decline. If a project is under pressure, it is worth putting strong people on it, or at least ensuring visible senior engagement. Even if the issue is largely political, visible commitment matters. It shows that you are taking the situation seriously.
Thirdly, look to the future. Even when a project is difficult, it helps to position it as a temporary setback rather than a defining failure. Talk about what can be improved next time and how you can work better together. That sense of continuity and investment can shift the tone of the relationship. If you demonstrate that you are thinking beyond the immediate problem, clients often respond in kind.
Finally, be careful in how you offer solutions or retrospective insight. When relationships become strained, it is very easy to slip into criticism. Phrases that imply fault or hindsight superiority will only escalate tension. It is far more effective to frame things collaboratively. i.e “If we had known this earlier, we would have approached it differently, and we can take that forward to the next project”. The aim is to reinforce that you are working together, not against each other.
In practice, much of this comes down to restraint. You will sometimes be blamed for things that are outside your control. That is part of the role. The vendors who handle it well are the ones who avoid becoming adversarial themselves, focus on delivery, and keep an eye on the longer term relationship.
Once the immediate tensions pass, and they usually do, those behaviours are what determine how well you work together going forward.
