Who writes this, and why you should believe any of it

Most marketing blogs do not tell you who wrote them or what they have actually done. That is convenient for the writer and useless for you. So here is the whole thing, including the parts that do not help me.

What I do

I build AI and automation systems for businesses. That means working with large language model APIs, automation pipelines and data tooling day to day, not reading about them.

InnoBotZ applies that to one narrow problem: AI assistants have become a front door for local recommendations, and most HVAC owners have no idea what those assistants say about them. So I built a measurement system and a fix list for it.

The measurement part is a pipeline that asks 20 real customer questions for a company's trade and city, several runs each, across four AI assistants, then scores every answer for whether the company was named, in what position, which competitors appeared, which sources the assistant cited, and whether anything stated about the company was factually wrong. That is 240 answers per audit. It is the free AI Shortlist Audit.

What I have not done

This is the part that matters more.

• I have no HVAC clients yet. InnoBotZ is a new service. There are no case studies on this site, no client logos, no testimonials and no results, because I do not have any to show.
• I will not invent them. If you ever see a testimonial or a client result anywhere on this site, it is a mistake and I want to hear about it at info@innobotz.com.
• I am not a licensed HVAC contractor. I write about marketing, AI visibility and the numbers behind them. I do not give technical HVAC advice.
• I cannot guarantee a position in any AI answer. Nobody can. There is no paid placement inside an AI answer and no ranking data for sale. Anyone promising you a spot is selling something that does not exist.

I would rather tell you all of that than have you find out. It is also why the audit is free, the 14-day Fast Start is guaranteed, and I only take on four new companies a month.

How the numbers on this blog get checked

Every post here is built on published 2026 research, and every figure names its source in the same sentence. My process, in order:

1. Find the original. Not a blog quoting a blog quoting a study. The actual publication, with its sample size and date.
2. Check the number says what the claim says. A surprising amount of marketing content misquotes its own sources.
3. Record the sample size. "816 contractors and $14.9 million in spend" tells you how much weight a figure carries. "Industry data shows" tells you nothing.
4. Delete what I cannot verify. If I cannot find the original, the number comes out of the post. There are several figures I wanted to use that are not here for this reason.
5. Label estimates as estimates. Where a figure is a vendor estimate or one owner's self-reported numbers rather than a study, the post says so.
6. Sign it. My name is on every page and I am the responsible person for all of them.

Where sources disagree, the posts show the disagreement rather than picking the flattering number. Published HVAC lifetime value figures, for example, vary by more than five times, so the posts say that and show you how to calculate your own.

Where AI fits in writing this

I sell an AI service, so I am not going to pretend a machine was not involved. The first draft of each post is written with AI assistance from the research above. Then I check every fact, number and source by hand, rewrite what is wrong, and publish under my own name.

One hard internal rule: once I have reviewed a page, no AI edits it again. If a page needs changing, my review runs again on the changed text.

Every post carries a visible disclosure saying this, directly under the title. The full statement, including how this meets Article 50 of the EU AI Act, is on the AI transparency page.

Why HVAC specifically

Three reasons, and they are all arithmetic.

• The tickets are large. Average HVAC tickets run about $1,400 on service and $7,800 on an install. One extra install a month changes a company's year.
• The shortlist is short. An AI answer names two or three companies, and 72% of consumers look at three or fewer businesses before deciding. Being absent is not a small disadvantage.
• The timing is unusual. 45% of US consumers used a generative AI tool for local business recommendations in 2026, up from 6% a year earlier. A seven-fold change in one year means almost nobody has adapted yet, which is the only kind of window a new company can walk through.

I work with one HVAC company per service area, because I cannot honestly help two competitors fight for the same shortlist.

Contact, and how to tell me I am wrong

If a number on this site is wrong, email info@innobotz.com with the page address and the line. I will check it against the original source. If I got it wrong, I fix the page and note the correction date. No argument, no delay.

• Email: info@innobotz.com
• Phone: +36 20 242 8206
• LinkedIn: Levi Kotka

Start with the blog if you want to see how I think, or the free audit if you want your own numbers.