Brand Mentions Are Exploding in Value - But Do Backlinks Still Matter?

Edward Sturm

Abstract

An hour with a veteran search specialist on what happens to the work when a model, rather than a results page, decides what gets recommended. The lever moves off the domain: away from links and on-page technical work, toward being a trusted entity the models retrieve — off-site mentions across the sources that co-occur with a niche, real brand signals, and relationships with the people who publish.

• The advantage stops belonging to the biggest budget and goes to whoever is tuned to how the models weight one specific corpus.
• Mentions overtook links, because a model does not need a physical link to name you — only to have seen you.
• Mass-produced content does not merely underperform; it stops the algorithm retrieving the site at all.
• The best link builders are now public-relations operators with publishing access to the sources that co-occur with the niche.
• Personal branding, payments, refunds and high-value support are the things that must never be automated.
• Across a whole career, conversion work on landing pages is the highest-return lever there is.

Everything below follows the conversation in order, from the shift itself through arbitration layers and niche-specific trust signals, into mentions, public relations and team design, and out through what should never be automated to the landing page. For anyone building a personal brand the useful part is the relocation: the work that decides whether you are recommended now happens almost entirely off your own site.

Chapter summaries

01:31 - Search feels the shift first, but the same pressures reach every industry

Search and SEO are at the forefront of AI-driven change, but the same pressures — rising cost to get clicks and traffic, and consolidation of traffic — are hitting other industries too.

"essentially, I'm an optimist in a world that is currently, I feel, on a bit of a downward trajectory."

— 01:31

The framing matters more than the prediction. Rising cost to earn a click and the consolidation of traffic into fewer places are not search-specific problems; discovery is simply where they show first, because discovery is the product. The optimist's position here is a working stance rather than a mood.

02:36 - An answer-default search experience was signposted well before it arrived

An AI-default search experience was a foreseeable inevitability, not a shock — it was signposted in the search company's own earnings calls for over a year; only the speed and aggressiveness of the rollout is surprising.

"Well, I I saw it as an given inevitability, right?"

— 02:36

It was in the earnings calls for over a year before it arrived. What surprised people was the speed and the aggressiveness of the rollout, not the direction — which is a useful correction to the story that the industry was blindsided.

04:16 - The bigger change is a search bar that takes images, tabs and files

The bigger change is the search bar becoming multimodal — importing tabs, images, files and extra context — because added context changes the AI's output substantially, and if user behaviour adapts to it, that alone reshapes SEO before any default flips.

"I really see the major in interesting thing of how they've done it is now you can impact import additional things into the search."

— 04:16

Importing tabs, images and files changes the answer substantially, and that is the part that reshapes the work before any default flips. Behaviour shifting is the real event; the interface change is only what makes it possible.

04:54 - Personalisation changes the answer per user, so one universal target disappears

Personalisation changes the model's answer per user, so optimisation stops being one universal target.

"if the context is substantially different in terms of number one you're importing a tab."

— 04:54

One universal target disappears. If the answer varies by who is asking, then optimising for a single canonical result is optimising for something that no longer exists, and the target becomes a distribution of answers rather than a position.

06:09 - The job moves to the trusted arbitration layer the interface chooses from

An optimiser's role shifts to layering personalisation on top and winning the trusted arbitration layer that decides the final recommendation.

"a lot of the additional marketing areas, like especially email marketing,"

— 06:09

Two jobs are left. One is layering personalisation on top through channels you own. The other is winning the arbitration layer — the point where the interface decides which recommendation to give — which is a brand problem dressed as a technical one.

07:10 - Search optimisation becomes every optimisation — reviews, video, episodes, all of it

Search optimisation becomes "every optimisation": specific sources, information types and formats get weighted higher, so building a trusted entity means reviews, video, podcast episodes and the like feeding the underlying knowledge base.

"it's it's definitely going from, you know, search engine optimization to every optimization, but the weighting is still different, right?"

— 07:10

The list of surfaces expands past anything a site can hold: reviews, video, episodes, changelogs, whatever feeds the underlying knowledge base. Being known in many places replaces being optimised in one.

08:02 - Every niche weights completely different vectors

Different niches weight completely different vectors: a local service is judged on local reviews, local press and charity/sponsor mentions; a content-thin vertical is judged on press releases, changelogs, a visible dev team, a community, active social and customer reviews.

"like a general business, okay. But what if you're doing something what if you're in iGaming?"

— 08:02

A local service is judged on local reviews, local press and sponsorship mentions. A thin-content vertical is judged on releases, changelogs, a visible team and an active community. The vectors are not slightly different between niches — they are entirely different.

08:23 - The trust jargon is really machine verification, and it predates the industry

The experience-authority-trust jargon is really just computer trust signals and machine verification, an idea that predates the search industry.

"certain vectors as well will have higher weighting as in certain niches."

— 08:23

The expertise-and-trust jargon dresses up something older and simpler: machine verification, an idea that predates the industry using the acronym. Naming it plainly makes it tractable, because verification has methods and jargon has opinions.

11:35 - Engagement signals have always been weighted, whatever the public position was

User engagement signals have always played a major role in trust and authority, despite the search company publicly denying it; recent additions include certain users having higher association with your content.

"you've always wanted as an SEO I think you always wanted to have a brand that people"

— 11:35

The public position and the practice were never the same thing. Engagement signals carried weight throughout, including newer ones that associate particular users with particular content, and a decade of denials did not change what the systems did.

12:15 - The answer replaces the click and keeps people inside the ecosystem

The AI-default experience will look nothing like today's search: not people searching pages and clicking through, but a summarised, cited, interactable answer — designed to keep people inside the ecosystem.

"as we move into the AI search and AI mode default experience"

— 12:15

The design goal is retention inside the ecosystem, not efficiency for the searcher. A summarised, cited, interactable answer means the page that would have been clicked is now a source. The click was the business model, and its removal is the whole shift.

13:10 - A brand with no outside signals gets harder to fake

A fake brand with no outside signals becomes harder to sustain, though the best search brands of the last decades already earned signals from many different places.

"I think it's going to be a harder to be a fake brand with no outside signals doing SEO."

— 13:10

The signals have to come from somewhere outside your own property, which is harder to fake and cheaper to earn honestly. The brands that were built well over the last decades already had them, so this is less a new requirement than a newly enforced one.

13:27 - An answer engine has three layers, and a large real-world footprint out-signals a lone producer

An AI has three layers — training data, retrieval, arbitration — and a giant firm with thousands of staff posting and doing real-world PR will out-signal you; expect some centralisation toward big brands.

"the AI can only understand so much and it's basically has three layers."

— 13:27

Training data, retrieval, arbitration. A large organisation with thousands of people posting and doing real-world public relations produces signal at every layer, which a lone producer cannot match — and that asymmetry points toward centralisation.

15:41 - Generative optimisation and search optimisation run on the same machine

"Generative" optimisation and search optimisation share the same underlying machine and the same techniques used for the last 10–25 years; new layers (personalisation, query fan-out that changes sources) sit on top.

"very, very similar, or at least very uh connected to all of the same techniques that we've been doing for SEO for the last 10 20"

— 15:41

The techniques are the ones the field has used for two decades, with personalisation and query fan-out sitting on top. Treating it as a new discipline sells courses; treating it as the same machine with new layers gets work done.

16:16 - The black box can only be reacted to, not predicted

The industry is a black box you can only react to, not predict; the advantage goes to whoever can react fastest to changes made by the search, AI, browser and OS companies.

"this is why I'm trying to change my workflows and trying to change my career path."

— 16:16

Nobody outside the companies can see the weights, so prediction is guesswork dressed as strategy. What remains is reaction speed, and it is a real edge: the advantage goes to whoever can move first on a change made by a search, browser or operating-system company.

16:58 - A tuned operator on a fraction of the budget beats the budget

You will be beaten by someone with 5% of the budget who is tuned to how the models weight your specific niche — which sources co-occur most, which token types the model prefers for that corpus.

"we're only going to be able to achieve based on how we can react to the changes"

— 16:58

Budget stops being the moat. Someone with a fraction of the spend, tuned to how the models weight one specific corpus — which sources co-occur, which formats the model prefers there — beats the campaign that understands none of that.

18:57 - Money follows understanding, and most of the budget does not understand the technology

Money follows understanding: the market is moved by investment, and big brands have failed to reach AI-model scope because agencies and CMOs running the campaigns don't understand the underlying technology the way power users who read the papers do.

"in terms of where everything is headed, where everything is changing, all of these kind of things,"

— 18:57

The diagnosis is uncomfortable and specific: the people running the large campaigns do not understand the underlying technology the way power users reading the papers do. Money moves the market, and it is currently moving on a poor model of the machine.

20:14 - Mass-produced content actively hurts the site's own retrieval

Mass-producing content with blogging software wastes money and actively hurts the site: a thousand low-quality signals stop the algorithm retrieving you for the AI to summarise at all.

"understanding of the underlying technology. And that causes a shift of where money actually gets put into."

— 20:14

It is worse than wasted. A thousand low-quality signals stop the algorithm retrieving the site at all, so the mass-production strategy does not merely underperform — it removes you from the pool the summariser draws from. Volume is not a strategy.

20:38 - A hundred well-placed forum comments beat a thousand mass blog posts

The operator who understands the algorithm knows a large forum site is the number-one sourced site — 100 well-placed forum comments beat 1,000 mass blog posts costing thousands a month.

"these kinds of blogging softwares that will just mass"

— 20:37

The comparison is exact and unflattering. One large forum is the single most-sourced site, so a hundred comments placed with judgement beat a thousand generated posts costing thousands a month. The caveat is real: manipulating that forum without experience gets the brand blacklisted.

23:25 - A small hyper-efficient team outperforms an agency of a hundred

A small hyper-efficient team — one expert designer, dev, SEO/marketer and project manager — powered by AI can outperform a 25-to-100-person agency, with full control and fast reaction.

"the real advantage of like what I was just saying there, where the little indie hacker can beat that"

— 23:25

One expert designer, one developer, one marketer, one project manager. Full control, fast reaction, and no layer of account management between the decision and the change. The claim is that this configuration beats twenty-five to a hundred people.

24:21 - Short-lived trends are worth riding, because a fraction of a wave is still enormous

Ride short-lived trends fast: a tiny reactive team can capture a fraction of a trend worth billions before it fades — 0.1–1% of a $10B six-month wave is still enormous.

"even if we look at the trend-based industry that we have been kind of living in through AI,"

— 24:21

A small reactive team can capture a fraction of a wave before it closes, and a fraction of something enormous is still enormous. Timing is the whole position — the trend does not need to last, it needs to be caught.

26:22 - Mentions overtook links as the thing the model weights

Backlinks no longer matter as much for getting mentioned in AI answers, because the model doesn't need physical links — it takes mentions from training data and retrieval; mentions went from near-worthless to hugely valuable while links were slightly devalued.

"So, 100% backlinks don't matter, I know, as much as they did before."

— 26:22

The mechanism is simple: the model does not need a physical link to mention you, it needs to have seen you. Mentions went from nearly worthless to extremely valuable, and links were devalued only slightly — the change is the promotion, not the demotion.

27:51 - Earning mentions is newsjacking, sponsorship and events — and then it is public relations

White-hat mention tactics: reactive newsjacking and launch-jacking around physical events to get users mentioning your brand, plus sponsorships, charities and events (strong for local), and beyond that it is essentially digital PR.

"If you want to do white hat, then reactive stuff,"

— 27:51

The white-hat list is short and concrete: reactive newsjacking and launch-jacking around real events, plus sponsorships, charities and local events. Past that it stops being link building and becomes public relations, which is a different skill with a different hire.

28:49 - The best link builders are public-relations operators with publishing access

The best link builders now are simply excellent at PR or hold direct relationships with many journalists; the real leverage is niching down and gaining publishing access to the top 50–100 co-occurring sources for your niche.

"if you can actually get into the into relationships with editors, journalists, bloggers, site owners,"

— 28:49

The leverage is access. Niching down far enough to hold relationships with the fifty to a hundred sources that co-occur with your subject beats any tactic, because those relationships are what turn a pitch into a published mention.

35:27 - An easy niche is one where the buyer cannot build the solution themselves

Pick a niche with easy SEO where buyers won't build the solution themselves — they search or ask an AI, find you, and you grow by iterating on user feedback.

"the fact that he quit an easy"

— 35:27

The test is whether the buyer could build it themselves. If they cannot, they search or ask, they find you, and you improve by iterating on what they tell you. It is a selection rule rather than a growth tactic.

36:29 - Search marketing is the hardest niche to market into

SEO is the hardest niche to market in — the audience are savvy marketers who see through everything and you compete against the biggest tools; the same effort in an easier niche goes far further.

"SEO is the hardest is I think it's like the hardest niche to do marketing in."

— 36:29

The audience are marketers who see through everything and the competitors are the largest tools in the category. The same effort spent in an easier niche travels much further, which is an argument about where to point the work rather than how hard to work.

41:55 - Grounding re-valued off-site presence over one domain's technical work

The optimism turned on when a major AI assistant began grounding a large share of queries (80–89% retrieving sources): backlinks, guest posts, mentions and co-occurring trusted sources are weighted far higher, while one domain's technical SEO matters less than the ability to post across many sources.

"Are you like are you feeling optimistic about SEO,"

— 41:55

When a major assistant began retrieving sources for the large majority of queries, off-site presence got re-priced upward: mentions, guest posts and trusted co-occurring sources matter more, and one domain's technical work matters less than the ability to publish across many.

43:43 - A funded company should buy owners of specific channels, not generalists

A funded company should invest in owners of specific channels — one person running just a couple of channels, not everyone running everything; the game is smaller pockets of expertise amplified by automation and tools.

"I'm excited about that as well, but that's why I think you need a you you need to have like uh"

— 43:43

The hiring implication is specific. Fund people who own a couple of channels deeply rather than a team where everyone covers everything thinly. Small pockets of expertise, amplified by tooling, is the shape that survives.

44:21 - Loosely coupled bets survive a single shock that destroys one of them

Spread bets across a larger number of loosely coupled ideas so that a single AI change that destroys one leaves the others intact.

"that's also where I kind of I I'm also kind of trying to hedge my bets more."

— 44:21

Correlated bets fail together. Spreading effort across a larger number of loosely coupled ideas means a single change that destroys one leaves the rest standing, which is the only real defence against a black box that moves without notice.

44:55 - The seemingly most stable operators are the ones struggling hardest

Many of the seemingly most stable operators are struggling hardest right now — even one with a strong personal brand, adopted systems and AI — as AI and robotics change the game and squeeze small and mid-sized businesses.

"really see a lot of people's businesses changing very rapidly."

— 44:55

It is the most counter-intuitive observation in the hour. The operators who look most stable — strong personal brand, systems adopted, tooling in place — are the ones struggling hardest, because the thing that made them stable is what is being repriced.

49:48 - Personal branding is the thing that must never be automated

Personal branding should not be automated: auto-posts and unedited AI avatars look bad, and losing even one or two subscribers who would each have spent hundreds is a large overnight loss.

"Is there anything in SEO and marketing that you feel people shouldn't automate or try to automate?"

— 49:48

Auto-posting and unedited generated avatars look exactly like what they are. The arithmetic is brutal: losing one or two people who would each have spent hundreds is a large loss for a saving measured in minutes.

50:27 - Payments and refunds are never automated, and support only for cheap products

Payments and refunds should never be automated, and customer service only for cheap low-stakes products; for high-value items, AI-answered calls lower brand perception and generate negative reviews and refunds.

"payments is a massive one. Anything like refunds and stuff like that should definitely not be automated."

— 50:27

Payments and refunds are absolute, and support is automated only for cheap, low-stakes products. On high-value items an automated call lowers brand perception and produces negative reviews and refunds — the saving is smaller than the damage.

51:06 - Good support and a good product manufacture free brand mentions

Good support plus a good product turns customers into your best brand mentions — you may not even need a paid brand ambassador on the forum; even a weak product with great support avoids negative mentions.

"I made a podcast about how important customer support is for AI."

— 51:06

Good support on a good product turns customers into the brand mentions you would otherwise pay for. The stronger version of the claim: even a weak product with excellent support avoids the negative mentions that are hardest to undo.

52:04 - Automation is priced on cost saved and never on brand spent

Firms automate for cost-saving too quickly without pricing the bigger cost to brand, customer relationships and revenue concentration — one mishandled customer cluster can lose a huge share of revenue overnight.

"I think that people tend to implement stuff as an as a cost-saving / efficiency mechanism too quickly."

— 52:04

The cost side of the ledger is short one column. Firms price the saving and never price the brand, the customer relationship or the revenue concentration — and one mishandled cluster of customers can remove a large share of revenue in a night.

52:37 - The biggest mistake is automating a success never actually achieved

The biggest mistake is trying to automate SEO — or social, or personal-brand — success you have never actually achieved yourself; a brand plus an occasional relevant post beats a complex AI system you don't understand.

"You know what I think what I see cuz you know, I like like you I also interact day-to-day"

— 52:37

The failure is trying to automate a result you never produced by hand. A brand and an occasional relevant post beats a complex system nobody in the building understands, because you cannot debug a process you never ran manually.

59:23 - Video ranks for almost everything, and it ranks within hours

Video ranks for almost everything and ranks fast — within hours, sometimes a top-three position in a couple of hours — and detailed descriptions influence AI answers.

"I I I can send you exactly how I do it cuz I have like all of my documentation,"

— 59:23

Video is a format the engine surfaces quickly and the assistants treat as a source. A colleague is moving hard into it for exactly that reason, with videos ranking within hours rather than weeks.

59:58 - One video can become the top source cited in an answer

A single video can become the number-one source cited in an AI answer for a competitive query.

"the opportunities with video, especially videos will rank in a couple of hours."

— 59:58

Video seeds the assistant's cited sources rather than only the results page. A test video became the top source in the overview for a competitive term — shown as a deliberate test for one specific query.

1:02:26 - Posting-time windows can be exploited, and the reaction can be automated

Exploit posting-time windows on a social platform: if a post succeeds in the first ~5 hours there's a ~3-hour window to push it higher; if it fails within ~8 hours, publish the next one — and automate agents to react (delay-post, or hype-post via a quote of the original).

"I've got a massive plan that I'm currently developing."

— 1:02:26

The windows are exploitable and mechanical. A post that lands in the first few hours has a short window to be pushed further; one that has not landed within eight is replaced by the next. Agents can be automated to react inside those windows.

1:05:56 - The move after any tactic question is to optimise for value rather than efficiency

The move after any tactic question is to stop optimising for efficiency and optimise for value: for an established project, ask what gives the highest percentage ROI increase.

"Follow me on every platform. Subscribe to the channel."

— 1:05:56

It is the move after every tactic question. Stop asking what is cheaper and start asking what returns more, which for an established project means finding the single highest-percentage improvement rather than the longest list of small ones. Value, not efficiency.

1:06:52 - Conversion work on landing pages is the highest-return lever of a whole career

Across an entire career of partnerships and investments, the single highest effective-hourly-rate lever is almost always conversion-rate work on landing pages — nothing else has the same longevity of direct revenue impact.

"what it is for most of the partnerships and investments that we do, it's almost always CRO."

— 1:06:52

Across a whole career of partnerships and investments, nothing has matched conversion work on landing pages for effective hourly rate. The claim is about longevity as much as size: the revenue impact keeps paying in a way other levers do not.

1:07:16 - A bad page, not a bad idea, is what most often sinks a business

Swapping a bad landing page for a good one can multiply revenue many times over (25–50x or more), and reverting stops the sales; many businesses fail not on a bad idea or product but on a bad representation of it via their pages.

"it's crazy when you have a bad landing page and you swap it out"

— 1:07:16

Swapping a bad page for a good one can multiply revenue many times over, and reverting stops the sales — which is as close to a controlled experiment as the field offers. Most failures are a bad representation of a good product rather than a bad idea.

Personal Branding Lessons

A conversation about where the leverage moved. The moves below all point off the domain and toward the places a model actually reads.

React fast rather than trying to predict the black box

Nobody outside the companies can see the weights, so prediction is guesswork wearing a strategy's clothes. What remains is speed: the advantage goes to whoever can respond first when a search, browser or operating-system company changes something. Building a workflow around reaction rather than forecasting is the practical version of that. 16:16

Place a hundred forum comments rather than a thousand blog posts

One large forum is the single most-sourced site, which makes a hundred well-placed comments worth more than a thousand generated posts costing thousands a month. The caveat is not decorative: manipulating that community without experience gets the brand blacklisted from it, or from the platform entirely. 20:38

Earn mentions through public relations rather than link building

The best link builders now are people who are excellent at public relations or who hold direct relationships with editors and journalists. The leverage is access — niching down far enough to be able to publish across the fifty to a hundred sources that co-occur with the subject. That is a different hire from a link builder. 28:49

Pick a niche where the buyer cannot build the solution themselves

The selection test is not size or margin but whether the person with the problem could solve it alone. If they cannot, they search or they ask, they find you, and the business grows by iterating on what they report back. Everything else is downstream of that one choice. 35:27

Spread loosely coupled bets so one shock cannot sink everything

Correlated bets fail together, and the black box moves without notice. A larger number of loosely coupled ideas means a change that destroys one leaves the others standing. It is the only real defence available against a system whose rules you cannot read. 44:21

Never automate the personal brand

Auto-posting and unedited generated avatars look precisely like what they are. The arithmetic decides it: losing even one or two people who would each have spent hundreds outweighs a saving measured in minutes, and the loss happens overnight and silently. 49:48

Never automate payments, refunds or high-value support

Payments and refunds are absolute — they can be gamed and they can be broken. Support is automated only for cheap, low-stakes products; on high-value items an automated call lowers brand perception and generates negative reviews and refunds that cost more than the headcount saved. 50:27

Put the next hour into the landing page

Across an entire career of partnerships and investments, nothing has matched conversion work on landing pages for effective hourly rate, and nothing else has the same longevity of direct revenue impact. Swapping a bad page for a good one can multiply revenue many times over, and reverting stops the sales. 1:06:52

Questions

Each answer ends at the moment in the recording where it is given.

How do I get mentioned in AI search answers instead of just ranking pages?

Build the off-site signals a model can read and verify: mentions across the sources it already trusts, real relationships with the people who publish, and a presence broad enough to be retrieved. The model no longer needs a physical link to weight you, so it takes mentions from its training data and from what it retrieves at answer time. The lever is being a findable, verifiable entity, not a well-tuned page. 26:22

Are backlinks still worth building, or do brand mentions matter more now?

Links still carry some weight, but they have softened, while mentions have gone from near-worthless to hugely valuable — because the model can account for a mention without a clickable link attached. Mentions filled the vacuum. The practical read is to stop treating a raw link count as the goal and start earning the kind of mentions a model reads as genuine, which is closer to public relations than to link acquisition. 27:19

Which sources should I try to appear on to influence AI answers in my niche?

Map the fifty to a hundred sources that keep co-occurring with your niche in AI answers, and get the standing to publish on them — paid, contributed or user-generated. Access to the exact places a model co-weights for your category seeds it better than chasing generic high-authority links ever will, and the leverage grows the tighter you draw the niche. Start by mapping the landscape you are actually competing in. 28:49

Does mass-producing blog content help or hurt me now?

It hurts. Software that mass-produces pages with no check on cannibalisation, relevance or tone loads a site with low-quality signals, and the retrieval algorithm reads those signals and stops surfacing the site — so the AI never sees it to summarise it at all. It wastes money and damages the presence you already had. A hundred well-placed contributions on a trusted source were described as beating a thousand such posts. 20:14

Can a solo builder or tiny team still beat a big brand at SEO?

Yes — spend loses to fit. Someone with a fraction of the budget, tuned to how the models weight a specific niche and which sources co-occur there, was said to beat a far larger budget aimed at brand awareness. A small hyper-efficient team amplified by AI can out-manoeuvre a large agency on control and reaction speed. The edge is understanding and speed, which is exactly the advantage a single person can actually leverage. 16:58

What parts of my marketing should I never automate?

Three lines were drawn: personal branding, because auto-posts and unedited AI avatars read as low-effort and cost you high-value followers; payments and refunds, always; and high-value support, because an automated response on an expensive purchase lowers brand perception and generates negative reviews. Automated support is tolerable only for cheap, low-stakes products. The boundary is drawn by price and stakes, not by task — automate the plumbing, never the trust. 49:48

How does good customer support turn into brand mentions?

A good product answered by a real person turns customers into the source of your mentions — they recommend you for free, so a paid advocate becomes unnecessary, and even a weak product paired with great support avoids negative mentions. In an era where a model rewards genuine mentions, handling the people who care most, by hand, is marketing rather than overhead. Automating it to save cost was named as the expensive mistake. 51:06

What niche should I build in if I want easy organic growth?

Pick ground where the buyers cannot do their own optimisation and will not build the tool themselves — they search or ask an AI, find you, and you grow by iterating on their feedback. Deliberately avoid the search-marketing niche itself: the audience are the savviest marketers and the competitors are the biggest tools, so the same effort returns far less. Work through the questions that separate an easy niche from a brutal one. 36:29

Does video help me rank and get cited by AI?

Strongly. Video was described as ranking for almost everything and ranking within hours, sometimes into a top-three position, with a detailed description doubling as material an AI answer will pull from — one test clip became the single top source cited for one specific query. Editing the transcript like a document and reusing one caption layer brings a finished short down to about twenty minutes, so the speed is affordable. See how to run video as a real channel. 59:23

Sources

Brand Mentions Are Exploding in Value - But Do Backlinks Still Matter?

The sections above follow the episode's own order. Each timestamp is the point in the recording where the idea is discussed, so the post reads as a map of the conversation rather than a rearrangement of it.

Quotations come from the episode transcript with spoken filler ("okay", "right", "you know", "um", "actually", "kind of", "sort of") and false starts removed. No words are added: square brackets mark anything inserted for sense, and an ellipsis marks any cut. British spellings are restored, and speech-to-text errors were repaired against context only where the intended word is unambiguous. Where a line appears both in a cold-open teaser and in its place in the conversation, the timestamp cites the second — the point at which it is actually said.

Quotes are stamped with the moment they are said rather than with a speaker's name.