I followed along on all of the comments and, not to belabor the matter, but maybe I can help clear things up. Maybe. Which should probably be capitalized. As in a big maybe.
When doing digital typesetting for the web, almost all web browsers can accept the UTF-8 character set by default, which can encode over 1.1 million character code points. Technically, "Unicode" is the character set and UTF-8 is the "character encoding." But that's a hair we don't need to split. See also ISO/IEC 10646, and listings of the UTF-8 character codes.
For example--and this is important in just a moment--as represented by decimal values, diacritical marks live in UTF-8 at the range of 768 through 879; mathematical operators are at 8704 through 8959; Cyrillic characters are 1024 through 1327. And...general punctuation can be found at 8192 through 8303.
So we have no problem using true typographer's punctuation like left and right single quotation marks ( ‘ ’ UTF-8 characters 8216 and 8217); or left and right double quotation marks ( “ ” UTF-8 8220 and 8221); or bullet points or em-dashes or even footnote daggers ( • — † respectively, UTF-8 8226, 8212, and 8224). In fact, they're fine to use in webpage content because they carry no special programming instructions to the HTML that underlies the rendering of the page.
A URL--Uniform Resource Locator, a webpage address--is universally valid only if it uses the ASCII character set, which technically is the first 128 characters in UTF-8...but in practice we can use only the 95 characters that are printable in a URL. For example, in URLs we can't use ASCII non-printing entities like spaces or tab characters or line-breaks.
So it isn't a glitch at Ancestry. To be strictly compliant with the World Wide Web Consortium, WikiTree shouldn't be allowing a typographer's apostrophe, UTF-8 8216, in a generated URL; see the W3 School's "HTML URL Encoding Reference." No UTF-8 character with a decimal value higher than that of a tilde ( ~ 126; 127 is "delete") should be used in a URL. We can generally get away with it for most printable characters, but there are web browsers/readers that will choke on it.
When constructing URLs--at least whenever possible--it's best to stick with that basic ASCII character set. If that can't be done, strict URL encoding then requires the characters above 127 in UTF-8 be represented by one or more instances of what's called an "escape" code, in this case a percent sign, followed by two hexidecimal digits. We've all seen URLs with %20 in them. That %20 represents a space, which the URL can't contain. A typographer's single right quotation mark (an apostrophe) escapes as %E2%80%99, as you and Jim have noted. Left and right double quotation marks escape as %E2%80%9C and %E2%80%9D respectively.
Here's a super-simple URI/URL validator where you can paste in an address and it will automatically translate it to the appropriate escape sequences: https://quuz.org/url/uri-validate.html. There are integrated functions in PHP--the extremely common coding language that drives the WikiTree database--and others that can be used to automatically convert characters in a URL to be properly "escaped" to make a strictly conformant URL. That's why, typically, if you create a web page that's titled "Hello There World" the resultant address for that page will be "Hello%20There%20World" automatically.
We see space characters like that escaped in WikiTree Free Space page names, but evidently we're not doing the same with non-ASCII characters like that pesky apostrophe. I can definitely understand the reasoning: if a profile is for a person named Jørgensen it won't be very tidy if the WikiTree ID is Jørgensen-000 but the URL contains J%C3%B8rgensen-000.
I told you the "clearing things up" part was a big maybe. Ahem. But at least that's the why behind Ancestry.com needing the Heimlich maneuver when you enter the non-compliant URL. I've never tried entering such a URL elsewhere--social media or Groups.io, for example--but it wouldn't surprise me if some of those wouldn't reject it also. I've run into genealogists who still run Windows XP and a decade-old web browser, and they might not be able to use those URLs either.
That little URL validator I linked to above is one alternative. Copying the URL from the address bar and pasting into that will give you the converted, strictly compliant version of the URL, and that can be used anywhere.
Something else I find interesting is that if I go to the aforementioned O'Shea profile and copy the URL directly from the address bar of my latest version of Chrome running on Windows 10, I can't even paste it in here as it shows in the address bar. I automatically get the converted version: https://www.wikitree.com/wiki/O%E2%80%99Shea-1059. Which would be a compliant URL that Ancestry would accept. Out of curiosity, what operating system and web browser are you running?