Few problems here.
First one is the regex.
Firstly, you don’t really want to think of the regex as if it was matching an infinitely long string of a.b.c.d.e.f.g.h.i.j.k.l.... instead you just want to think of a single case x.y. Those cases can be right next to each other.
So get rid of (\.[a-z])*.
Secondly, you don’t have capture groups. Or well, you do have one in there, but I’m assuming you accidentally did it. If you’re not yet familiar with Regex capture groups, I’d recommend learning them, they’re quite useful in certain cases (like here!).
But anyway, to create capture groups, you just put ( ) around the part of the Regex you want to capture.
So you want to capture the characters before and after the . (or well, actually only the latter one, this approach will have a problem, more on that later). So your Regex would now look like this:
([a-z])\.([a-z])
Upon a match, the hotstrings() function would output two variables, $1 and $2 (that’s all they are, names of variables).
When you refer to the variables, $1 gives you the character before the ., and $2 gives you the character after the ..
So now we get onto the second problem, referring to the capture group variables.
a := $1
b := $2
Send, a{U+22C5}b
Here you create the variables a and b for no reason, though that’s not an issue of course, but how you try to refer to the variables a and b is a problem.
You’re using a send command, so you’re in legacy AHK syntax. In legacy AHK syntax you refer to variables by wrapping them around in %%.
So your send command would look like this:
Send, %a%{U+22C5}%b%
But lets not write legacy AHK (even though the hotstrings() function totally is legacy AHK).
To switch over to modern AHK (expression syntax) we would do specify a single % followed up by a space. And then we can do this:
SendInput, % $1 "{U+22C5}" $2
Also skipped defining the useless variables a and b and switched over SendInput due to it being the recommended faster and more reliable send mode.
And now would have an almost working script like so:
hotstrings("([a-z])\.([a-z])", "word")
return
word:
SendInput, % $1 "{U+22C5}" $2
Return
It just would have the problem of chaining multiple a.b.c.d.e.f.g... not working very well. But that’s fine, since the Regex could do with more improvements.
We want to use a positive lookbehind and capture only the character after the . like so:
(?<=[a-z])\.([a-z])
Also, I’d say it would be fitting to replace [a-z] with \w (match any word character). So the Regex and the whole script would be:
hotstrings("(?<=\w)\.(\w)", "word")
return
word:
SendInput, % "{U+22C5}" $1
Return
And now it should work just as requested.
And if my talks about legacy vs modern AHK confuse you (that’s to be expected if you don’t know the difference), I’d recommend giving e.g. this a read:
https://www.autohotkey.com/docs/Language.htm
CLICK HERE to find out more related problems solutions.