I enjoy using an array of devices, tools and services in my day to day life when working with Microsoft 365, Teams and AI, and sometimes I like a product so much that I’ll decide to write a review on it. These days that increasingly means AI tools, models, agents and platforms.
Vendors will also reach out to me and ask if I’d like a loan/demo/free product (or free credits, or early access), and will sometimes ask if I can write a review of it.
Having been inspired by Greig Sheridan’s review policy, I thought it was about time I wrote my own policy.
It’s important to me that my reviews always remain impartial and truthful, regardless of whether the product I’m reviewing has been supplied to me for free, or if I’ve purchased it myself.
Also, I should point out that my definition of a review is my evaluation of a product, to figure out if you should buy it or not. This means a review will have my experiences, my opinions and my recommendations.
My general reviews policy
- I will make the ultimate decision on whether I choose to review a product.
- I will make the decision on the medium used for the review (e.g. blog post, video review).
- The review is my own personal opinion, and does not necessarily reflect the opinion of my employer or any past employer.
- I have not, and will not, accept payment in return for a product review.
- I will always disclose it clearly, at the start of the review, if a product was supplied to me by a vendor (as a gift, loan, demo unit, free credits or early access) or if I have a business relationship with the vendor, for example through my employer or its partnerships. For video reviews, I will say so in the video and in the description. If I bought the product myself, I’ll say that too.
- If a physical product is given to me by a vendor, I may keep it, and I’ll say so in the review. If it’s on loan, the vendor will pay the costs involved in returning it.
- The company or vendor providing the product for review does not get an early review or copy-approval rights concerning the review.
- Reviews are a snapshot in time, and generally speaking won’t be updated should new functionality or features be provided at a later date via a service or firmware update. For fast-moving AI products, I may add a dated note if something significant changes.
- The company or vendor (or anyone else) is able to point out technical errors in a posted review, and these will be corrected where applicable.
An example of a technical error is:
What I said in the review: The camera supports HDMI and USB
Correction: The camera only supports USB
Reviewing AI products and services
AI tools behave differently to most products, so I hold myself to a few extra rules when I review them:
- I test AI products with my own or synthetic data. I will never use customer, employer or other confidential data in a review.
- I’ll state what I actually tested: the product, model or version, licence or plan tier, and the date. AI results change between runs and between releases, so this matters.
- AI output isn’t deterministic. I’ll share what I saw, and where it makes sense I’ll repeat a test, but your results may differ from mine.
- Preview, beta and early-access features will be labelled as such, and may change or be withdrawn after I publish.
- I’ll be clear about what I tested myself and what is a vendor claim. I won’t repeat benchmarks or marketing claims as if they were my findings.
- Where relevant, I’ll cover privacy, security and governance, such as where data is processed, whether it’s used for training and what admin controls exist. I’ll share what the vendor documents and what I observed, but this isn’t legal or compliance advice, so please check with your own security and legal teams.
- I write my own reviews. If AI helps me draft, edit or create images, I’ll say so, and the opinions and conclusions remain mine, checked by me.
If you’d like to get in contact with me, you can do so here: https://blog.chiffers.com/contact-craig/
References: