Depending on the structure/need of the team, I can either implement the UI myself or I will hand it off to Engineers to implement. In the latter case I prepare handoff files in Figma (to be attached to any tracking tickets, i.e. JIRA) that details all screens, states, flows, or functionality. It also includes a section for asset handoff as needed to make it easy for Engineers to know what to export.

If engineers are the ones implementing, I request a final review (screenshots or videos) of the feature in the build to insure alignment and any final polish notes.

The final step in the process is putting it into user's hands.

For larger features or onboarding I'll either run my own limited play test sessions with players I've recruited, or (more often) I'll utilize Playtest Cloud to get the game in players' hands prior to launching it so that we can catch anything major that isn't working for players. Watching the players interacting with the game is a fantastic way to figure out where areas of improvement are.

Are they able to complete tasks? How difficult/easy was it for them to do so?
Were there any times they were confused or seemed hesitant about what to do next?
What did they learn and when did they learn it? Did they actually internalize what we taught them? Will we need to reinforce their learning?
Did they seem to enjoy the gameplay/feature? If not, why not? Etc

I've found that additional feedback can be obtained via player surveys, either through socials (Facebook or Reddit, for example) for a launched game, or collected from a player's council of target players if the game has not yet been publicly released.

Finally, if the game has been released, data analysis & A/B testing are a fantastic way to maximize player understanding - I'll work with key stakeholders and data analysts to figure out areas with room for improvement and iterate for better metrics.