Best way to make money using ai?

I tried making AI videos for TikTok affiliate marketing but it wasn't successful, so I think to make money from it I need to learn many other things besides using AI.

tztz said:

A lot of faceless channels fail for one simple reason: they build for views first and monetization second.


That sounds fine in theory, but in practice it creates channels that can get traffic without creating much revenue. Then people burn out because they are chasing RPM screenshots and waiting months for ad income that may never become meaningful.


The mistake is not using AI. The mistake is using AI to mass-produce content without a business model behind it.


What usually works better in 2026 is starting with the monetization path first, then building the content around that.


Instead of asking, “What niche can get views?” ask, “What niche has a clear problem, a clear buyer, and a clear next step after the video?”


That changes everything.


A faceless channel is not the business by itself. It is a traffic and trust layer. AI is not the product either. AI is just leverage. The actual money usually comes from what happens after the view: affiliate offers with decent commissions, lead gen, digital products, simple info products, templates, toolkits, or a service funnel if you already know the market.


If all you do is publish generic videos and hope ad revenue carries the whole thing, you are building on the weakest layer first.


The first thing I’d check is whether the niche has buyer intent, not just audience size.


A lot of people here get stuck because they choose broad entertainment or random trending topics. That can get impressions, but it rarely attracts people who are ready to buy anything. A niche with a smaller audience but stronger intent is usually better. Finance, software workflows, business problems, productivity systems, recruiting, home services, B2B pain points, certification help, specialized hobbies with expensive tools, or anything where the viewer is already trying to solve something can outperform a “viral” niche with no commercial angle.


The content should lead naturally into an action, not feel like bait-and-switch.


For example, if the channel is about a specific problem, the landing page should continue that exact problem. Not a generic homepage. Not a random list of offers. Not a messy Linktree full of unrelated stuff. One problem, one audience, one next step.


That is where most faceless channel setups break.


They spend hours generating scripts, voices, visuals, edits, thumbnails, and posting schedules, but almost no time on the offer-page match. So even when a video gets traction, the traffic leaks everywhere.


Before changing five things at once, I’d separate the funnel into 3 parts:

1. Content angle – Does the topic attract the right viewer, or just a curious viewer?
2. Offer match – Does the next step solve the exact problem raised in the content?
3. Conversion path – Is the landing page focused enough to convert cold traffic?

If one of those is weak, more uploads will not fix it.


AI helps most when it reduces production friction in a niche you already understand. Script drafting, research clustering, hook variations, thumbnail ideation, repurposing long-form into short-form, voice cleanup, simple motion visuals, and testing different content angles faster — that is useful leverage.


What AI does not fix is weak positioning.


If the niche is unclear, the offer is generic, or the landing page is built like an afterthought, AI just helps you scale a bad setup faster.


Another mistake is treating views like the main KPI.


Views matter, but not in isolation. What would count as a real signal here is:

• click-through to the landing page
• email opt-in rate
• affiliate click quality
• conversion rate by content topic
• revenue per 1,000 views
• repeat viewers from the same niche cluster

Those numbers tell you whether the channel is attracting buyers or just browsers.


A 20k-view video with no downstream action is often less useful than a 2k-view video that sends qualified traffic into a good offer.


That is why I would not tell people to ignore ad revenue completely, but I also would not build around it in the early stage. Ad revenue is better treated as an extra layer, not the foundation. Nice when it comes, but too slow and too unstable to be the whole plan for most new faceless channels.


If I were starting one from scratch, the order would be:


Pick a niche with obvious commercial intent.
Find 2–3 real problems people already spend money to solve.
Map one clean offer path for each problem.
Build content that answers, demonstrates, compares, or breaks down those problems.
Send traffic to a focused page that matches the video topic.
Track which topic-to-offer pair produces actual revenue, not just engagement.
Then scale the winning angle.


That approach is slower in week one, but much better in month three.


Also worth saying: faceless does not mean low trust. A lot of channels fail because they feel disposable. Generic AI voice, recycled scripts, overused stock footage, and no real point of view usually kills retention and conversion. Even if the channel is faceless, it still needs a clear editorial style, niche consistency, and useful information. Otherwise it just looks like another churn-and-burn content farm.


So yes, AI can absolutely help build and scale faceless channels, but I would use it as an execution multiplier, not as the strategy itself.


The big picture is not “get more views with AI.”


The big picture is “build a channel around commercial intent, use AI to speed up production, and make sure the traffic lands somewhere that can actually convert.”


That is usually the difference between a channel that looks busy and a channel that actually makes money.

Click to expand...

What was said about AI not fixing weak positioning is spot on

In cold email, AI helped me most with personalization. Instead of sending the same template to everyone, I use it to generate unique drafts based on the prospect's site or recent activity. That alone lifted reply rates

But the foundation still has to be right, clean list, proper deliverability setup, clear offer. AI just speeds up the work, doesn't replace thinking through the funnel

Combining artificial intelligence (AI) with affiliate marketing could be a profitable avenue.

AI can help you generate operational strategies, while affiliate marketing can help you manage product sales and revenue collection.

Essentially, you have a product; once it's sold, you make money. AI can help you devise numerous promotional strategies.

However, various tests are needed to find the profitable combination of AI and affiliate marketing.

I'm currently testing this approach, and I believe it should be viable. However, the process is undoubtedly one of continuous failures; you need to learn from these setbacks to achieve success.

A. The goal is to sound like an experienced member adding practical perspective, not pitching a tool or making inflated claims.





B. The biggest mistake I see is people treating AI like it’s the strategy, when really it’s just leverage. If the offer is weak, the targeting is off, the angle is generic, or the funnel is leaking, AI will mostly help you fail faster and at a bigger scale.

In cold email, where AI helped me the most was personalization, not magic copywriting.

Before, a lot of people were basically sending the same template to 1,000 prospects and hoping a first line like “love what you’re doing” would count as relevance. It doesn’t. Most inboxes are full of that already. What AI does well is speed up the research-to-draft step. You can feed it a prospect’s site, recent activity, positioning, product category, maybe even a weak point in their messaging, and get a custom draft that at least sounds like it was written for that business instead of for “anyone with a website.”

That part alone can lift reply rates, but only if the rest of the campaign is already in decent shape.

A lot of people here get stuck because they focus on the visible part of outreach, the message, while ignoring the hidden part that usually decides whether the message even gets a chance. Clean list, proper domain setup, warmed infrastructure, alignment between the prospect and the offer, and a reason for them to care right now. That foundation matters more than whatever prompt stack someone is bragging about.

The mistake is not “using AI.” The mistake is using AI to decorate a broken funnel.

If your list is scraped badly, your ICP is too broad, your domain reputation is poor, or your offer needs three paragraphs to explain, AI won’t rescue that. It will just help you produce more polished-looking underperformance.

Same thing with affiliate marketing.

A lot of newer people think combining AI with affiliate means easy money because AI can generate content, landing page ideas, ad angles, comparisons, emails, scripts, social posts, and so on. That part is true. AI absolutely helps with operational output. You can brainstorm faster, test faster, and build more assets without doing every step manually.

But affiliate is still a numbers game tied to economics and market fit.

You still need to answer the boring questions:

Who is the traffic for?
Why would they click?
Why would they trust the bridge page?
Why this offer and not another one?
What payout gives you room to test?
What is the actual conversion path after the click?
Where is the drop-off happening?

AI can help generate ten angles. It cannot tell you which one has real buyer intent unless you run traffic and check the data.

That’s why I think the best use of AI in affiliate isn’t “replace the marketer.” It’s more like “reduce the cost of iteration.”

For example, AI is useful for:

coming up with multiple hooks around the same offer
rewriting pre-sell content for different audience segments
drafting advertorial structures
generating comparison-style copy variations
turning one idea into email, short-form content, and landing-page variants
summarizing competitor positioning so you can spot angle gaps faster

All of that is useful. None of that removes the need to test.

Usually this happens when people get excited by output volume. They start producing 50 creatives, 20 blog posts, 10 email variants, and a pile of “optimized” copy, but they still haven’t nailed one clear market-message match. So they end up with more assets but not more revenue.

Before changing five things at once, I’d separate the symptom from the cause first.

If reply rates are low in cold email, is the problem the copy, the targeting, the deliverability, or the offer?
If affiliate pages get clicks but no conversions, is the problem traffic quality, offer mismatch, weak pre-sell, or poor trust transfer?
If content gets impressions but no action, is the hook wrong, or is the underlying audience just low intent?

AI helps once you know what variable you’re actually testing.

What would count as a real signal here is not “the copy sounds better” or “the AI wrote this faster.” Real signals are things like:

reply rate improvement from relevant personalization
more positive replies instead of generic opens
higher click-through on one angle versus another
better EPC after changing the bridge message
lower CPA from clearer audience-to-offer alignment
more stable conversions across traffic segments

That’s the level where AI becomes useful instead of just entertaining.

I’m also with you on the failure part. That’s the reality nobody wants to hear because it isn’t sexy. Whether it’s cold email or affiliate, most combinations don’t work on the first try. Most tests are either neutral or losers. The value is in failing with structure so you actually learn something from each round.

Random failure teaches nothing. Tracked failure teaches a lot.

That means keeping variables tight, testing one meaningful change at a time, and not letting AI generate so much stuff that you lose the thread of what’s being validated. A lot of people sabotage themselves by creating too many options before they have enough data to justify any of them.

So yes, AI plus affiliate can absolutely be viable. AI plus cold email can also be very effective. But in both cases, the foundation has to be right first.

Good list beats clever copy.
Good deliverability beats fancy prompts.
Clear positioning beats “humanized” fluff.
A real offer beats content volume.
Testing beats theory.

That’s why I look at AI as a multiplier, not a savior. If the base system is solid, it saves time and increases iteration speed. If the base system is weak, it mostly helps you create higher-quality noise.

That’s the practical difference.

Andreiv_12 said:

Whats your best way to make money using ai?

I see a lot of people doing this faceless youtube channels with ai or tiktok/ig pages but the thing is how do they make their money?

Click to expand...

There’s a lot of noise around AI right now. The people actually making money with it usually combine it with something that already works (traffic, content, offers) instead of relying on AI alone. What direction were you looking at?