Think for a moment in English.

• I sing /sɪŋ/
• I sang /sæŋ/
• I sung /sʌŋ/
• a song /sɒŋ/

Note what’s happening here: the basic meaning of the word is dictated by the consonants, that stay the same across multiple words. Then you change the vowel to convey further meaning: present vs. past vs. participle vs. noun.

In English this is a bit of an exception, but your typical Semitic language (as Arabic and Hebrew) does this all the time, typically following certain patterns. For example, extending OP’s example:

You do see some affixes here and there, like that -tu in katabtu. But the workhorse of the morphology are those vowel changes.

And since this system was already present in Proto-Semitic, you can even find cognates across Semitic words, and they’ll conjugate? decline? in similar-ish ways.

Words in semitic languages, unlike indo-european languages are conjugated with a system of roots and templates.

Roots are three (or even four) letter words, that are not meant to be used by themselves since they are equivalent to the infinitive in IE languages. So K-T-B would be “to write” and nothing else. No tense, no gender, etc etc.

Templates fill these in, by applying the root to a template. They specify the tense, gender, x-person etc.

So K-T-B (to write) + _A_A_TU (I did this thing in the past) = KATABTU

tl;dr: roots are verbs and templates are context for them