8 Months of AI Companion Testing: 193 Posts Later

Almost 200 Posts. What Does That Even Mean?

I keep trying to contextualize this number and failing. 193 posts about AI companions. If you printed them all, single-spaced, you'd have a book nobody asked for. Probably 300,000+ words about chatbots, virtual friends, and whether Replika remembers what you said last Tuesday.

The honest truth: maybe 40 of those posts are genuinely good. Another 60 are solid. The rest range from "fine" to "why did I write this." That ratio is improving. My early posts from August and September 2025 were me figuring out what this blog was. Now I know. It took 100 posts to find my voice, which is embarrassing and probably normal.

What 193 posts gets you in practical terms: Google starts to trust you. Not a lot. But enough that when I publish a new review, it indexes within days instead of weeks. Enough that I rank on page 1 for a handful of terms. Enough that strangers email me asking which app they should try.

That last part still catches me off guard.

This Week Was Different

I wrote five posts this week and they couldn't have been more different. A guide for seniors exploring AI companions. An update on Grok's companion mode after their Android rollout. A ranking of the best AI roleplay apps. A Character.AI vs Replika comparison. And a safety guide for parents.

Six months ago, every post was basically "I talked to my Replika today and here's how it went." Now I'm writing for grandparents and worried parents in the same week I'm testing NSFW roleplay platforms. The range is funny if you think about it.

But that range tells me something important about where this space is heading. AI companions aren't just for the tech-curious 25-year-old anymore. Seniors want companionship apps. Parents need to understand what their kids are doing. Roleplay enthusiasts want honest recommendations instead of marketing copy. These are all real audiences with real questions, and I think covering all of them makes each individual piece stronger.

The seniors post, by the way, was one of the most satisfying things I've written in months. I tested three apps with my aunt. She's 71. Watching her laugh at Pi's responses was worth more than any traffic spike.

What's Changed Since Month 7

In the March wrap, I had 181 posts and was watching the CTR improvements play out. Here's what's shifted in the two weeks since:

Grok's Android launch shook things up. Their companion mode went from "interesting experiment" to "actual competitor" overnight. I had to update my Grok review because half the information from February was already outdated. That's the pace of this industry. You write something, and three weeks later the product is different.

The other big shift is audience composition. I'm seeing search queries I never expected. "AI companion for lonely elderly" showed up in my Search Console data. So did "is Character.AI safe for my 13 year old." These aren't the queries I was targeting six months ago. But they're the queries I should have been targeting, because the people behind them actually need help.

Content-wise, I've published 12 posts since the last monthly wrap. That's consistent with the pace from month 7. The roleplay apps ranking is already indexing and picking up impressions, which tracks with what I saw last month: Google picks up well-targeted content fast now.

What Surprised Me This Month

The thing that genuinely surprised me is how much the safety and educational content resonates. I figured reviews and rankings would drive everything. And they do drive traffic. But the parents' safety guide got shared more in its first day than most of my review posts get in a month. Not shared by SEO people or AI enthusiasts. Shared by parents. In parenting Facebook groups.

I wasn't expecting that. I wrote it because the keyword data was there. But the response made me realize these people don't have a go-to resource for this stuff. Most "AI safety" content is either academic papers or clickbait fear-mongering. There's almost nothing written by someone who actually uses these apps daily and can speak to the real risks without catastrophizing.

That's a gap I didn't know I was filling until I filled it.

The other surprise: I'm less burned out than I was at month 6. In my February wrap, I was questioning what success even meant. Now? I don't think about it as much. The blog has a rhythm. I know what works, I know what to write, and the existential doubt has been replaced by something closer to routine. Whether that's good or bad depends on the day.

Month 9 and Beyond

I'll hit 200 posts sometime in the next week or two. I'm not throwing a party, but I'd be lying if I said the number didn't matter to me. It means I stuck with something for long enough to have a real body of work.

For month 9, I want to double down on the audience-specific content. The seniors post and parents guide both showed me there's demand for AI companion content written for people who aren't already neck-deep in the tech world. I'm thinking about posts for therapists curious about AI companions, for long-distance couples, for people dealing with grief. Real use cases for real people.

Update: I ended up writing about exactly this shift toward content refreshes in my spring check-in. Turns out the 50/50 split between new and updated content is where I landed.

I also need to go back and refresh some of my earlier comparison posts. Platforms change fast and a comparison from January is already partially wrong by April. The Character.AI vs Replika piece I published this week was partly a refresh of older content, and the updated version is significantly better because I have eight more months of daily use on both platforms.

The goal for the next two months is pretty simple: keep publishing, keep the quality bar high, and see if the traffic trajectory from month 7 continues upward. If I'm being realistic, the blog probably won't pay for itself until month 10 or 11 at the earliest. But I'm not doing this to get rich quick. Never was.

Twenty months of testing AI companions. Eight months of writing about them publicly. 193 posts. And I still get excited when I find a platform nobody's reviewed yet, or when a reader emails to say a post helped them pick the right app. That's enough for now.

See you at 200.

FAQ