The range
Reports vary enormously, across traditions, across practices, and across the same person on different days. Some common territory appears often enough to be worth naming.
People describe a loosening of the ordinary sense of self, sometimes to the point where the boundary between self and surroundings stops being obvious. They describe emotion arriving with unusual force, often grief, often for something long past. They describe a vivid return of memory. They describe a sense of contact with something larger, framed through whatever language they already had for it. They describe the body doing things it was not asked to do: shaking, weeping, heat, cold, nausea.
They also, frequently, describe very little. A quiet hour, some tiredness, and a mild disappointment. That is a real and ordinary outcome, and people need to be told so beforehand, or they will conclude something is wrong with them.
The difficult passages
Difficulty is not the same as harm. Many of the experiences people later describe as the most valuable were frightening at the time.
What commonly comes up: fear that does not attach to anything in particular, the conviction that something has gone irreversibly wrong, a sense of dying, confusion about where or who one is, and shame surfacing about something the person has not spoken about.
What supports someone through it is ordinary presence. A calm voice, a reminder of where they are and that it will pass, and no attempt to fix or interpret. What does not support them is being told to surrender, being told this is what they needed, or being left alone because someone decided it was their process.
Why the accounts are hard to give
People frequently cannot describe what happened, and this is not evasiveness. Ordinary language is built for subjects, objects, and time. Some of what gets reported does not divide up that way, so the account comes out as metaphor, as religious language, or as a shrug.
This has two consequences worth holding. First, the account you receive is a reconstruction, assembled after the fact, and it will change with each retelling. Second, the words the person reaches for come from the frameworks already available to them, including whatever the facilitator said beforehand. That is a good reason to say very little beforehand.
Listening without leading
The most useful thing a practitioner can do with an account is receive it without adding to it.
Questions that open: what do you remember. What stayed with you. Where were you in your body. What was the hardest part. What has been different since.
Statements that close: what that means is, what you met was, your ancestors were telling you, this is why you needed it, now you need to. Each of those hands the person an interpretation at the moment they are least equipped to refuse one. See what ethical integration means.
Accounts that warrant concern
Most reports, however strange, are not clinical concerns. A few are.
Pay attention to persistent derealisation or depersonalisation lasting beyond a couple of weeks, fixed beliefs that are not moving and that alarm the people around the person, a sense of being watched or interfered with, sleep that has not returned, and any thinking about self-harm. Those are reasons to involve a clinician promptly rather than to keep talking it through. See scope of competence and referral.
On extraordinary claims
People sometimes come back with claims that cannot be evaluated: about reality, about other beings, about what is going to happen. A practitioner does not have to agree, and does not have to argue.
The workable position is to take the experience seriously and stay agnostic about the metaphysics. What matters practically is what the person does next. A claim that leads someone to be kinder and more honest is a different situation from one that leads them to stop taking their medication or to abandon their family. The second is not a theological disagreement. It is a safety concern.