our county parks instagram sat at 1400 followers for a year and a half, i switched to one illustrated character in every post and seven months in the engagement is completely different

i run the instagram account for a county parks department in rural ohio. it is not my main job. i am an admin assistant and social media got added to my responsibilities about two years ago with no extra budget and no strategy beyond post something a few times a week. i spend maybe ten hours a week on it now. for the first year and a half i did what every small municipal account does, which is pull stock photos from unsplash, stick the county logo in a corner, write a caption that reads like a notice pinned to a corkboard, and post it into the void. the account had about 1,400 followers when i started and it still had about 1,400 followers fourteen months later. average post got 22 likes. one or two comments, usually from the same three people who work down the hall from me.

last january i pitched my supervisor on building one illustrated character who would show up in every single post. not a photo of a real person, not a stock model, just a drawn figure with a consistent look. my thinking was that every account people actually follow has a face attached to it, and ours didn't have one. my supervisor said go ahead as long as it didn't cost anything new and nobody could confuse the character with a real county employee.

i was upfront from the beginning that she was AI artwork. the very first post she appeared in, the caption explained that she was an illustrated character created using AI tools to give the account a recognizable visual identity and that she was not a real person. i pinned that explanation in the story highlights. seven months later i still drop a note about it in captions every couple weeks because new followers were not around for that original post.

the workflow is simple enough that i feel almost dumb describing it. the only software in the loop is Buffer, Canva, and APOB AI. i batch about two weeks of posts on a saturday morning, it takes around three hours, and then the schedule runs itself.

january and february were discouraging. the first six weeks after the switch i lost about fifty followers and average likes per post dropped from 22 to around 15. some of that was people who had followed for nature photography and didn't want cartoon illustration in their feed instead. some was probably the algorithm adjusting because the visual profile of the account had changed overnight. i genuinely considered scrapping the whole experiment at the six week mark.

but the comment section told a completely different story even while the numbers were dropping. the very first week she appeared, comments per post jumped from one or two to six or seven. most of them were people asking who she was. is this a new ranger. is she real. did the county hire someone. a few people sent direct messages asking if she was based on a specific person. i answered every one the same way: she is not real, she is an AI generated character, we created her so the account would have a consistent face instead of random stock photos. almost everyone who asked responded well to that. two people unfollowed and told me they do not support AI art, which i respect.

the character didn't have a name at first. i just called her the ranger in captions. sometime in late february a commenter suggested june and it stuck immediately. i didn't plan it. someone typed june and the next post i wrote her name like it had always been there. that might be the single thing that changed the account the most. once she had a name the comments shifted from what is this to oh june is at the covered bridge today, and that is a fundamentally different kind of engagement between an audience and a post.

by march the follower count started moving for the first time in over a year. i was picking up about 80 new followers a month. june posts averaged around 45 likes and about 12 comments. the posts without her, plain landscape shots, still pulled maybe 25 likes and three or four comments. for a county parks page in a county of 74,000 people that felt significant.

the growth picked up through spring. by the end of may the account was somewhere around 2,000 followers. june posts were pulling about 120 likes and 14 to 16 comments. the plain landscape posts had actually improved too, settling around 35 to 40 likes, which was better than the stock photo era but nothing close to what the character posts did. the ratio stayed remarkably stable from that point forward. character posts reliably pulled about three times the engagement of posts without her.

now in late august the account sits at 3,200 followers. june posts average about 160 likes and 18 to 22 comments. the landscape posts without her average around 50 to 55. the account has more than doubled in seven months, which i know sounds modest, but for a rural county parks page it is more reach than the department has had on any platform in its history. the parks director mentioned last month that summer program registrations were up about 20 percent over last year and that several parents specifically mentioned seeing the instagram posts. that was the first time in our office connected social media to an actual outcome.

the part i did not expect is how much of the conversation ends up being about the artwork itself rather than the parks. people notice when her jacket changes for the season. they point out when a particular illustration looks slightly off from her usual style. they comment on small details i put in the background. on a typical post maybe three or four out of eighteen comments are purely about the illustration and have nothing to do with the trail or event it was supposed to promote. one post in early may where i had her standing in front of a blooming redbud tree, sixteen of the twenty two comments were about the drawing and not about the trail. that was the moment i realized the character had become the content as much as the parks information had, and i still have mixed feelings about that.

and then there is the portion of the audience that genuinely dislikes the artwork on principle. every two or three weeks someone leaves a comment about how a public agency should hire a local artist instead of using AI, or that the illustrations look cheap, or just a flat AI garbage with nothing else attached. i do not delete those and i do not argue with them. i usually reply with something like i understand the concern and that there is no budget for commissioned illustration, which is the truth. sometimes another follower pushes back on the critic. more often the comment just sits there with two or three likes and nobody else engages with it. it happens often enough that i notice every instance and i won't pretend it doesn't sting a little. but it has never turned into anything larger than a single comment thread, and the overall engagement trend has not changed because of it.

the message that meant the most came from a local watercolor artist back in april. she wasn't angry. she said she thought the character had real personality and asked whether the county would ever commission work from a local illustrator. i told her honestly that there was nothing in the budget for it, that the tools i already use are all the department approved and that is the whole line item. she said she understood and that the approach made sense given the constraints. we have talked a few times since and she actually shared one of the june posts from her own account, which was not something i saw coming.

what i actually took away from seven months of doing this is that the consistency was the variable that mattered. not the art style, not the novelty of AI, not any single post or caption trick. people built a relationship with a recurring character the way they latch onto a comic strip or a mascot on a cereal box. june showing up reliably in their feed three times a week is what made them stop scrolling and comment and tag their friends. whether she was drawn by a human or generated by software turned out to matter less to the audience than i expected, and a bit more to a small vocal minority than i would have liked.

i expect the critical comments to keep showing up every few weeks and i plan to keep answering them the same way. the account is still growing and the department is more engaged with social media than it has been in the three years i've been handling it. that is the whole picture.