L

Personalisation vs Inconsistency

Both are the output changing. One of them traces back to something you did — and if you cannot find that thread, the real cost is that you now have to check everything.

Sort six changes See what each one is

Two Mondays. The report changed both times.

One change you can explain in a sentence. The other has been explained by nobody, including the tool.

Change A · personalisation

You picked charts over tables twice. The third report arrives as a chart.

You can say out loud why this happened, and you could undo it if you wanted to. That is the whole test.
Change B · inconsistency

Last month the report was always a PDF. This week it is an Excel file, then a Google Sheet. No setting was changed by anybody.

There may well be a reason inside the system. From where you sit there is no thread to pull, which makes it random in the only sense that matters.
Both are the output changing. Only one of them has a cause you can reach.

What each one actually is

One follows your lead and stays followed. One moves for a reason that never reaches you.

The difference is not how much it changed — it is whether the change traces back to you. Personalisation is a change you could have predicted from your own past choices, and could undo. Inconsistency may have a perfectly good cause inside the system, but if it never reaches you, it is unpredictable from where you sit.

And the damage is not the wrong format on one Tuesday. It is that you go back to opening every one to check — which is precisely the work the tool was supposed to take away, quietly handed back.

When output changes, ask one question: can I name something I did that would produce this? If yes, it is personalisation and you are in control of it. If no, it is worth raising as a defect rather than absorbing as a quirk — because the alternative is checking everything forever.

Sort six changes, then see where each came from

Tap a change, then tap whether you can trace it to something you did. Then see where each one came from.

The same shifting output, different desk

Pick a situation. In every one, the tool had been reliable long enough to be trusted.

In every one of these, the tool had been reliable long enough to be trusted.

The summaries that stopped being summaries

Two weeks of tidy bullet points, set deliberately. Then one Tuesday, a full transcript, and no setting had been touched.

Test the distinction

Five questions. Nothing is scored.

Question 1 of 5
Multiple choice

The words this pair actually contains

Five terms, not two. Tap one.

Questions people ask

Open all 10 questions
How can I tell if an AI’s changes are personalisation or inconsistency?
  1. List any settings or feedback you gave recently.
  2. Check if the AI’s changes match those actions.
  3. If yes, it’s personalisation. If not, it’s inconsistency.
  4. Ask for support if the pattern makes no sense.
Why does my AI assistant sometimes switch formats or styles without warning?

Unpredictable changes usually mean inconsistency, not personalisation. The system may have a bug, or updates may have changed its behaviour. If you did not request or set the change, report it as an issue.

What are examples of personalisation in AI tools at work?
  • Email summaries that match your reading style.
  • Reports delivered in your favourite format.
  • Chatbots using your preferred greeting.
  • Document templates that reflect your past edits.
Can personalisation ever feel like inconsistency?

Yes. If personalisation is not explained or if settings are unclear, you might not realise why the output changes. Good AI tools show you what’s being personalised and let you control it.

How does inconsistency in AI cost time or money at work?

When outputs change without warning, you spend extra time fixing formats, double-checking content, or explaining errors to colleagues. Inconsistency can break workflows and damage trust in the tool.

What should I do if my AI assistant’s behaviour becomes unpredictable?
  1. Check your settings or recent feedback.
  2. Restart or update the tool.
  3. Document the changes you notice.
  4. Contact support with specific examples.
Is personalisation always helpful in AI tools?

Personalisation usually helps, but only if it matches your real needs. If the system guesses wrong or personalises too aggressively, you might get results that do not suit you. Look for options to adjust or turn off personalisation.

How do AI designers prevent inconsistency while offering personalisation?
  • Clear user settings and controls.
  • Transparent feedback systems.
  • Testing for stable behaviour.
  • Regular updates with user input.
Can two people using the same AI tool get different outputs?

Yes. Personalisation means the tool adapts to each person’s choices, so Priya might get charts while James gets tables. If the differences are random, though, that’s inconsistency and should be fixed.

What’s the fastest way to restore consistency in an AI tool?

Reset your settings to default and check if the behaviour stabilises. If random changes continue, report the issue. Consistency comes from clear settings and reliable software, not from guessing what the AI will do next.

The cost is not the wrong format. It is that you go back to checking every one — which is the whole benefit of automation, handed back.

Copyright © Pawan Nayar · LLOS.ai · 2026 — Personalisation vs Inconsistency: a change you can trace, versus a change you cannot.Original pedagogy, voice, and design — all rights reserved.