A note before the list. These aren't bad products. They are bad small business purchases right now, for specific reasons. A list of things to skip is more useful than a list of things to buy, because the things to buy depend entirely on what you do. The things to skip apply to almost everyone.
This list is updated every quarter. If you're reading this six months from now and one of these has improved a lot, the criticism may no longer apply. Use your judgment.
AI "voice agents" that answer your phone
The pitch: an AI voice agent that answers calls 24/7, books appointments, qualifies leads, and never takes a sick day. The price: $500 to $3,000 a month.
The reality in 2026: the technology is impressive in a demo and frustrating in production. The voice sounds almost human until the caller asks something off-script, at which point it either makes something up or asks the same question three times. Your repeat customers will know within two calls. Your new customers will leave a worse review than they would have if no one answered at all.
This will be a great product in three years. It is not a great product today. If you genuinely cannot answer your phones, hire an answering service. They cost less, and a real human handling 60% of calls is better than a robot handling 100%.
AI website builders that promise "a website in 30 seconds"
The website you get is technically a website. It is also generic, slow, identical to ten thousand other small business sites, and nearly impossible to edit when you decide six months later that you want it to look like yours.
The problem isn't that AI can't build websites. The problem is that the cheap AI website builders optimize for the demo — instant generation, lots of pages — not for the year you'll spend trying to maintain it.
If you want a fast, cheap website, a $99 Squarespace template still beats most of these. If you want a real website, you want a designer for three hours of work, not a tool that generates everything and means nothing.
Most "AI for [your industry]" verticalized tools
When you see "AI for plumbers," "AI for dentists," "AI for restaurants," ask one question: what is this company actually good at?
In nine cases out of ten, the answer is: marketing to your industry. Not understanding your industry. They have taken a generic chatbot or generic scheduling tool, painted it with your industry's logos and language, and added a 40% markup.
The few real exceptions are companies whose founders came out of the industry, have specific domain knowledge baked into the product, and can name three things their tool does that a generic version couldn't. If you can't get those three things in a 10-minute call with the sales rep, you're paying for a wrapper.
"AI sales rep" tools that send personalized cold emails at scale
Reasons to skip:
- They don't work as well as the demos suggest, because every prospect has gotten ten of them this month.
- "Personalization" generated from a LinkedIn scrape and a press release isn't personal. Your prospects can tell.
- The legal landscape on automated cold outreach is shifting fast. Even where it's still legal, the deliverability impact on your domain reputation is real and lasts longer than the campaign.
- You will be associated, by your prospects' inboxes, with the worst senders using the same tool. Email spam filters are increasingly sophisticated about this.
If you have a real cold outreach problem, hire one part-time SDR and give them good tools. The math is similar. The output is much better. The reputational cost is much lower.
"Enterprise AI platforms" sold to small businesses
If a tool's pricing page has tiers labeled "Team" and "Business" and "Enterprise," and the "Team" tier costs $99 per user per month, the product was designed for companies much larger than yours.
The tell isn't the price. The tell is the complexity. These tools come with admin panels, role-based permissions, audit logs, SSO requirements, and onboarding processes that assume you have an IT team. You don't. You'll spend the first six weeks configuring it, never finish, and end up using 8% of the features.
Two specific categories where this is rampant:
- AI data analysis platforms. A spreadsheet and ChatGPT will get a small business 90% of the way. The platform exists for companies with a data team.
- AI knowledge base / "company brain" tools. Useful at 500 employees. At 15, you don't have enough institutional knowledge to justify the friction of putting it all in another system.
Chatbots glued onto your website
The 2024-2026 wave of "add an AI chatbot to your site to convert more visitors" is mostly noise.
The honest data: most small business websites get a small amount of traffic that converts at known rates. Adding a chatbot does not meaningfully change those rates. What it does is add a maintenance burden (someone has to keep its training material current) and a small but real risk that it answers a customer's question wrong in a way that costs you a sale or, worse, exposes you to liability.
The exception: if you have a complex product, high site traffic, and a clear pattern of pre-sales questions that all sound similar — yes, a well-trained AI chatbot can pay for itself. That's a narrow exception. Most small business sites don't qualify.
The general pattern
If you squint at all of these, they share a structure:
- The demo looks great.
- The marketing implies your competitors are already doing this.
- The price is enough to feel real but not enough to require deep due diligence.
- The actual ROI requires conditions that most small businesses don't meet.
A useful filter when you see a new AI product: imagine the version of your business in two years that adopted this tool and made it work perfectly. Is the gain noticeable? Or is it the kind of improvement you wouldn't be sure happened?
If it's the second one, skip it. There will be a better version next year, or you'll realize you didn't need it.