Mark Zuckerberg has set out an extraordinarily ambitious vision for artificial intelligence. In his essay, The Future is for Everyone, the Meta CEO argues that we are approaching a world in which people have access to AI systems with capabilities beyond human intelligence. His central argument is optimistic.
“Invention, not automation, will be the greatest contribution of superintelligence.”
Zuckerberg believes AI will become something far more significant than a productivity tool. He envisages personal AI agents working continuously on our behalf, helping with everything from careers and finances to health, education and entrepreneurship. In his words:
“Everyone will have an exceptionally capable personal agent.”
Whether that future arrives exactly as Zuckerberg predicts is open to debate. The debate his essay has generated raises an important question for businesses adopting AI today.
If AI can give us an answer, how do we know whether we should trust it?
And perhaps more importantly:
Do we understand where that answer came from?
Zuckerberg frames the defining question of the AI era largely around access.
“The defining questions of our age are who will have access to superintelligence and what will we direct it towards.”
It is an important question, but inside businesses, there is another one that deserves just as much attention: What information will that intelligence be working with?
AI does not operate independently of information, it interprets data and it identifies patterns in that data, then it draws conclusions from data. Increasingly, it will recommend, or autonomously take, actions based upon that data. Which means the quality, context and trustworthiness of the information beneath AI becomes critical. The more responsibility we delegate to AI, the less tolerance we have for uncertainty in the foundations supporting it.
Much of what Zuckerberg describes already exists in today's generation of AI assistants. ChatGPT, Claude, Gemini and other tools can already explain complex subjects, research questions and help people learn. There is an uncomfortable flip side, the same personalised assisstant capable of helping someone understand the work, can simply do the work for them.
That distinction matters. Receiving the answer and understanding the answer are two completely different things. Exactly the same problem is emerging inside enterprise IT.
Imagine asking an AI assistant:
“Which servers support this critical business application?”
The AI produces an answer immediately, What happens next? Where did the information come from? Was it taken from the CMDB? Was it discovered directly? Was it inferred from relationships? When was the underlying information last validated? Did another operational system disagree? Has anything changed since the data was collected? If an experienced engineer believes the answer is wrong, can they interrogate the evidence behind it?
That is the difference between information and operational trust.
The enterprise technology industry has spent decades becoming better at discovering infrastructure. Modern discovery platforms can identify enormous amounts of information across increasingly complicated technology estates, but visibility alone does not automatically create truth.
An organisation may have several different platforms describing the same infrastructure. A discovery platform might identify one thing, the CMDB might record another. An asset management system might contain something different again and, a spreadsheet maintained by an infrastructure team may tell yet another story. So which answer is correct?
This is precisely why TekWurx talks about Operational Trust.
Discovery tells you what a system has observed. Operational trust helps you establish whether the information is accurate, how it relates to other sources and whether you have enough evidence to confidently act upon it.
There is a temptation to believe AI will somehow resolve messy enterprise data, when AI makes the quality of that data considerably more important.
Gartner found that 63% of organisations either do not have, or are unsure whether they have, the right data management practices for AI.
More significantly, Gartner predicts that through 2026 organisations will abandon 60% of AI projects unsupported by AI-ready data.
The organisations achieving successful AI outcomes appear to understand this. Gartner reported in 2026 that organisations with successful AI initiatives invest up to four times more in foundations including data quality, governance, AI-ready people and change management than organisations experiencing poor outcomes.
The AI race therefore isn't simply about who deploys the most sophisticated model. It may ultimately be won by organisations with the most trustworthy foundations beneath those models.
There is another Gartner prediction that should make technology leaders pay attention. By 2028, Gartner predicts 50% of organisations will adopt a zero-trust posture for data governance because of the proliferation of unverified AI-generated data.
The principle is simple: Do not automatically trust information because it exists, verify it. That means understanding its source, context, and how it was created. Then understanding whether it has changed.
That philosophy should sound familiar to anyone working in enterprise IT because it is the challenge organisations already face with operational data.
As AI becomes increasingly embedded within IT Operations, we believe organisations should expect more than an answer. They should be able to establish:
1. What is the answer?
The obvious starting point. What infrastructure exists? What supports this application? What is likely to be causing this incident?
2. Why is that the answer?
What evidence supports the conclusion?
3. Where did the information come from?
Discovery? CMDB? Cloud platform? Monitoring system? Asset repository? Another operational source?
4. Can we challenge it?
If an engineer believes something is wrong, can they trace the conclusion back through the underlying operational evidence? This matters enormously. AI should augment human expertise rather than obscure the information experts need to exercise judgement.
At TekWurx, we describe operational excellence through four connected pillars.
Visibility tells you what exists.
Trust establishes confidence that the information is accurate.
Traceability allows you to understand where that information came from, how things are connected and why they matter.
Control gives you the confidence to act.
AI potentially makes all four dramatically more valuable. An AI agent operating against incomplete, duplicated or contradictory operational information can produce a convincing answer that is still wrong. Once AI begins acting autonomously rather than simply advising humans, the consequences become greater again.
Most large organisations don't have a shortage of operational data. They have too much of it.
Discovery platforms, CMDBs, ITSM systems, cloud platforms, monitoring tools, endpoint management platforms, asset registers and spreadsheets all contain different views of the technology estate.
uControl doesn't ask organisations to rip those investments out and start again. Instead, it creates an operational trust layer across them.
uControl can ingest operational information from multiple sources, normalise inconsistent data, reconcile conflicting information, identify trusted sources, model business services and applications, and detect operational drift. The result isn't simply another answer. It is trusted operational context. For AI, that distinction could become fundamental. Instead of asking an AI system to interpret a collection of fragmented operational records, organisations can provide it with reconciled, contextualised and continuously validated operational information.
There is enormous excitement surrounding artificial intelligence. There should be. Zuckerberg believes AI could give individuals what he describes as unprecedented powers of invention and discovery. Others are considerably more cautious.
The Guardian notes that Zuckerberg's optimistic vision comes amid concerns about labour disruption, surveillance, security and the enormous power increasingly concentrated around companies developing advanced AI.
404 Media offers a much more sceptical interpretation, questioning whether a future presented as being “for everyone” may also happen to align extremely well with Meta's commercial interests.
The wider argument will continue. At TekWurx, our position is considerably more practical. AI is coming into the enterprise. In many organisations, it is already there. The job of technology leaders isn't simply to decide whether AI is good or bad. It is to make sure that when AI is used to make operational decisions, those decisions have trustworthy foundations. The greatest danger may not be an AI that cannot answer our questions. It may be an AI that gives us a confident answer that nobody knows how to challenge. To discuss your achieving operational trust in your IT environment, talk to our team.
Zuckerberg's vision is ultimately built around giving people extraordinarily powerful intelligence. That could transform how we live and how we work. Intelligence and trust are not the same thing. The workplace of the future cannot simply become somewhere we ask machines questions and accept whatever comes back. We need evidence, provenance, context, traceability, and we will need experienced people who are prepared to say:
“I don't think that's right. Show me why.”
That is why operational trust matters. As AI becomes more powerful, we believe it will matter more, not less. AI can give you the answer. Operational Trust gives you the confidence to act on it.