
Yoast SEO: A Practical Guide to Setting It Up Once and Writing for People
Yoast SEO helps most when its site-wide settings are right. See which settings to choose, what the traffic lights really mean and the mistakes to avoid.
From F&B to fintech, clinics to law firms, startups to enterprise. If your customers search on Google, we make sure they find you first, not your competitors.
One specialist team, focused only on the organic rankings that put you in front of ready-to-buy Singapore customers.
A clear, sequenced path from audit to rankings. You always know what we’re doing and why it matters for your leads.

Quick answer: The phrase ai based seo describes services where a vendor uses machine learning somewhere in the delivery. In practice it covers four very different things: bulk content production, genuine data automation, optimising to appear in generated answers, and plain relabelling. Buyers should ask which one they are being sold before comparing any quotes.
If you are shortlisting agencies in Singapore right now, at least one proposal on your desk will carry this label. The difficulty is that ai based seo is a description of a method rather than an outcome, and methods are exactly what a buyer cannot inspect from the outside. This post is written for the person signing the contract, not the person doing the work. It sets out the four distinct things vendors mean when they use the phrase, how to tell genuine automation from a familiar retainer with a new cover sheet, the questions that reliably separate the two, and what any of it should cost in SGD. It is not a technical guide and it is not an argument that the technology is empty. Our own SEO services page describes where we use it and where we do not, which is the same disclosure we would ask any vendor for.
The phrase is doing four different jobs in the market, and a proposal that does not tell you which one is on offer cannot be evaluated against another proposal at all.
The first is production. The vendor uses language models to draft content at volume, which lowers their cost per article. This is real and it is legitimate when disclosed. The question for you is not whether a model was involved but what the human contribution was, because a model trained on the average of what already exists produces, by default, the average of what already exists.
The second is analysis. The vendor uses machine learning to do things that are genuinely impractical by hand: clustering 20,000 queries into intent groups, parsing server log files to see what crawlers actually fetch, classifying thousands of product pages by attribute, detecting anomalies in ranking data before a human would notice them. This is where the real value sits, and it is the least commonly sold, because it is infrastructure rather than deliverables.
The third is the target rather than the tool. The vendor is selling work aimed at making you visible inside generated answers rather than only in the blue links. That is a legitimate objective, though it overlaps heavily with conventional technical and editorial quality.
The fourth is the label. The scope is a standard retainer, the word is on the cover, and the price carries a premium. Most agencies using this phrase are selling the first or the fourth. The second is where a buyer actually gains something, and it is the one you have to ask about directly because nobody leads with it.
These four have different success measures, different risks and different appropriate prices. Comparing a production pitch against an analysis pitch on price alone is comparing a printing quote with an engineering quote.
You do not need to be technical to recognise genuine automation, because it has a distinctive shape: it produces things that would be impossible at your budget by hand.
Real automation works on data you own. Crawl data, log files, Search Console exports, product catalogues, internal link graphs. A vendor doing serious analytical work will reference your data in the pitch, not a generic industry benchmark.
It produces artefacts you can inspect. An internal link opportunity map covering every page on a 4,000-page store. A crawl budget analysis showing which page types a crawler spends its time on. A query cluster map that groups your search demand into decision stages. Ask to see a redacted example from a live account. A vendor who has built this can show it within the hour.
It scales in a way manual work does not. The tell is that the output is the same quality at 50 pages and at 5,000. If the pitch describes a process that obviously breaks at scale, the automation claim is decorative.
And it comes with a stated failure mode. Everyone building on these systems knows where they fail: fabricated facts, confident nonsense on niche regulation, silent drift when a model updates. A vendor who cannot tell you how they catch errors is telling you they do not check.
By contrast, repackaged basics have their own signature. Deliverables that would have appeared in a proposal three years ago under different headings. Heavy emphasis on volume of output. A proprietary visibility score with no disclosed methodology. And a premium attached to the framing rather than to any named deliverable.
| Claim in the pitch | Genuine version | Repackaged version | What to ask |
|---|---|---|---|
| We use AI for content | Models draft, specialists supply first-hand data and verify claims | Model drafts, junior publishes, volume is the selling point | Who verifies facts, and what do they know that a model does not |
| We use AI for keyword research | Clustering and intent classification across the full query set | Tool exports with a new label | Show me the cluster map from a live account |
| AI powered technical audits | Log file analysis, crawl comparison, regression alerts | Standard crawler report, automated summary | Which of these findings needed a model |
| AI visibility tracking | Documented sampling, reproducible method, tied to enquiries | Proprietary score, undisclosed method | How is the score calculated and can I reproduce it |
| AI driven strategy | Forecasts built from your own historical data with stated assumptions | A generated slide deck | What data trained the forecast, and what is the error range |
Six questions, asked in a meeting, will tell you more than any amount of reading. We suggest them to prospects who are evaluating us alongside others, because a vendor who cannot answer them is a risk regardless of who wins.
What would be different if you removed the phrase from this proposal? If the deliverables are a technical audit, content, internal linking and reporting, you are buying a search retainer. That may be exactly right, and it should be priced as one.
What is the human review ratio, and who does it? Ask for a name and a background. The gap between useful and generic output is entirely in the review, and if the reviewer is a junior with no domain knowledge, the output will be structurally average no matter what produced the draft.
What does the model never touch? A credible answer exists. In regulated Singapore categories it should include claims about outcomes, anything touching MAS-regulated financial promotion, medical claims subject to advertising rules, and legal advice framing. A vendor who says the model handles everything has not thought about your liability.
What happens to my data? If your customer records, enquiry logs or internal documents are being fed into third-party tools, you need to know which tools, under what terms, and whether that is consistent with your PDPA obligations. This is a governance question, not a technical one, and it is increasingly the one that decides procurement in larger local firms.
Show me one deliverable from a live account, redacted. Not a sample from a template library. The difference between a real internal link map and a slide about internal linking is visible in five seconds.
What will you stop doing? Adapting to generated answers almost always means retiring content as well as adding it. A proposal that only adds has not analysed anything. We have found that the consolidation half of the work is where most of the gain sits, and it is the half that sells badly because it produces fewer line items per invoice.
Price is where the label does most of its damage, so it is worth stating the market plainly.
A competent monthly retainer for a Singapore SME generally runs from around SGD 1,500 for a small single-location business to SGD 5,000 or more for a competitive multi-service or multi-location operation. One-off SEO audits commonly land between SGD 2,000 and SGD 6,000 depending on site size and technical complexity. In our experience those ranges have not moved because of anything in this category, and we have seen no proposal yet that justified a departure from them on methodology alone.
A premium charged specifically for the methodology, on top of an otherwise ordinary scope, is the clearest single signal to slow down. Automation lowers a vendor’s delivery cost. There is no coherent argument for it raising your price while the deliverables stay the same. Where a higher price is genuinely justified is when the scope is larger: a log file analysis programme, catalogue-scale entity work, or a custom data pipeline are real engineering and should cost accordingly. The test is whether the extra money buys an extra deliverable you can name.
There is a second pricing pattern worth knowing. Some vendors use lowered production costs to quote unusually cheap content-heavy retainers, sometimes well under SGD 1,000 a month for a high article count. That can be fine for a business that genuinely needs breadth of coverage. It is a poor fit for a business in a small, trust-sensitive Singapore category, where five specific pages outperform forty generic ones and where a factual error in published content carries professional consequences. In those categories, a small number of carefully sourced pages tends to do more than a large number of fast ones, and our pricing page sets out how we break scope down so the comparison is like for like.
The useful question is not which vendor is best in the abstract, it is which shape of vendor fits your situation. There are three and they suit different businesses.
A full-service marketing agency gives you one contract covering search, paid, social and creative. The trade-off is depth: search is one of several disciplines competing for the same senior attention, and the person on your account may be a generalist. That is the right choice when your total marketing budget is modest and coordination across channels matters more than depth in any one of them, and we recommend it openly to businesses in that position rather than stretching a small business SEO scope to cover work it was never built for.
An independent consultant gives you direct access to an experienced practitioner and usually the sharpest diagnosis per dollar. The trade-off is capacity and continuity: one person can only carry so much execution, and holidays, illness and a better client all create gaps. That is the right choice when you have internal execution capability and need direction rather than hands.
A search specialist firm gives you depth and bench strength in one discipline. The trade-off is that you are contracting for search specifically, so paid media, creative and social sit with someone else and somebody has to hold the joins. We are a specialist by deliberate choice, because search rewards accumulated pattern knowledge across many accounts in the same small market, and a generalist cannot build that while also covering four other channels. If your priority is a single supplier across every channel, a full-service agency is the better structural fit, and we will say so in a first call rather than after a contract.
That is the resolution to the trade-off rather than a hedge. Decide which of the three problems you actually have, then judge candidates within that category on the six questions above. Comparing across categories on price produces the wrong answer every time. For a sense of how depth in one discipline compounds over an engagement, our industry SEO overview shows how the same underlying method adapts across categories.
Some signals are not negotiating points, they are exits.
Guaranteed rankings or guaranteed positions in generated answers. Nobody controls either. A guarantee means either a meaningless metric will be reported or something risky will be done to your site.
A proprietary score presented as the primary success measure. If the methodology is undisclosed and the number is not reproducible by anyone else, it cannot be audited and it will only ever move in one direction in your reports. Clicks, enquiries and revenue are measurable, comparable and yours.
No access to your own accounts. Search Console, analytics and any tool tracking your site should be owned by your company with the vendor added as a user. Vendors who keep this in their own accounts are creating a switching cost, and you will discover it at exactly the wrong moment.
Content published without a named reviewer in a regulated category. In healthcare, legal and financial services in Singapore, published claims carry professional consequences that sit with you rather than with your agency. Ask who reviews and put it in the contract.
And an unwillingness to say what they would not do. Every credible practitioner has a list of things they decline. A vendor who agrees enthusiastically with every suggestion you make in a sales meeting is selling agreement, not judgement. In hotel SEO work we routinely tell operators that a third of what they were planning to publish will not earn its keep, and that conversation is usually the most valuable one in the engagement.
Field notes: When a prospect asks us to judge a proposal, we look past the label to the work. In our finance case study, the programme for an independent CFP-licensed financial advisory firm in Raffles Place was not sold as anything new. It was author profiles with credentials, the MAS licence displayed on every service page, technical fixes, a Google Business Profile, and Singapore-specific content on CPF, SRS and HDB topics, with average content length rising from 380 to 2,200 words. Over eight months, monthly organic visitors went from 320 to 1,155 and monthly organic leads from 3 to 31. Every line of that scope could be named, reviewed and checked by a person with sector knowledge. That is the test we would apply to any vendor proposal: if the line items cannot be described in plain terms, the framing is doing the selling.
The category is real and the marketing around it is not, and a buyer only needs to separate the two well enough to compare like with like. Strip the phrase out of every proposal on your desk and read what remains. If two proposals then describe the same work, the price difference is a marketing charge.
Our practical advice is to decide what you are buying before you evaluate anybody. If you need breadth of content at low cost, judge on quality control and who reviews. If you need analysis at a scale humans cannot reach, ask for the artefact and ask to see it from a live account. If you need to be visible where answers are generated, ask what the measurement plan is before the work starts. And if you need all three, expect to pay for scope rather than for a label.
Singapore’s market is small enough that reputation circulates, which works in a careful buyer’s favour: ask for two references in your own sector and call them. The businesses that do well here are not the ones that picked the most advanced-sounding supplier. They are the ones that knew which problem they were solving, bought exactly that, and measured it in enquiries rather than in dashboard scores.
If you want to see how that discipline reads over a full engagement in a regulated category, the finance SEO results write-up follows one end to end.
Our about page sets out who does the work here, which is the same disclosure we would ask any vendor on your shortlist to make.
Both are present in the market, which is why the phrase is unhelpful on its own. There is genuine engineering work in clustering, log analysis, entity extraction and anomaly detection that no team could do manually at SME budgets. There is also a large volume of conventional retainer work being relabelled and repriced. The practical test is whether a proposal names a deliverable that could not have appeared in a good retainer three years ago. If it cannot, you are looking at the marketing version.
Not for the method itself. Automation lowers a supplier’s cost of delivery, so there is no coherent reason for it to raise your price while the deliverables stay identical. A higher price is justified when the scope genuinely grows, for example log file analysis, catalogue-scale structured data or a custom reporting pipeline, because those are engineering commitments. Ask what specific extra deliverable the premium buys and insist on seeing it named in the scope.
Read three published pieces end to end and look for anything only your business could have supplied: your prices, your process, your cases, Singapore-specific regulatory detail, named people. Generic content is fluent and says nothing checkable. Then ask who the named reviewer is and what their sector background is. Fluency is not the tell, because these systems write fluently by design. Absence of first-hand information is the tell.
Ask which third-party tools your data passes through, whether any of it is used to train external systems, where it is stored, and how that aligns with your obligations under Singapore’s Personal Data Protection Act. If customer records, enquiry logs or internal documents are involved, get the answer in writing and in the contract. Most vendors have a reasonable answer available. The ones who have never been asked are the ones to watch.
They can be, for businesses that genuinely need breadth of coverage across many low-competition topics and can tolerate variable depth. They are a poor fit for small, trust-sensitive Singapore categories where a handful of specific pages do the commercial work and where a published factual error carries professional consequences. Judge the offer on the quality control process rather than on the article count, and ask what happens when a piece is wrong.
No, and the guarantee itself is the warning. Nobody controls which sources a generated answer draws on, the selection varies by phrasing and by session, and there is no submission mechanism. What a credible vendor can commit to is the work that makes a page usable as a source: crawlability, clear extractable statements, specific verifiable facts, consistent entity information. Outcomes should be described as probability improvements, never as guarantees.
Reduce both to a list of deliverables with quantities, then delete every adjective. Two proposals describing a technical audit, twelve pages of content and monthly reporting are the same proposal regardless of their vocabulary. Once they are side by side, the differences that remain are usually seniority of the people assigned, the review process, and what each one is willing to stop doing. Those three predict outcomes far better than the methodology described.
It reports clicks and enquiries segmented by query type and page type, against a baseline set before the work started, with a plain-language account of what was done and what it was expected to move. Rankings appear as context rather than as the headline. Proprietary scores, if present at all, are explained. The strongest signal of a good pack is that it tells you what did not work that month, because every month contains something that did not.
Less than the seniority of the person actually on your account. Singapore is a small market where the same practitioners circulate, so a boutique with two experienced specialists frequently outperforms a large firm that assigns a junior. Ask who does the weekly work, not who attends the pitch. Then ask for two references in your own sector and call them. In a market this size, references are unusually easy to verify.
Ask what they would stop doing on your site. It is difficult to answer without having looked properly, it forces a judgement rather than a sales response, and it reveals whether the vendor sees their job as adding deliverables or as improving outcomes. Almost every Singapore site we audit carries content that has never produced an enquiry, and a vendor who does not raise that is not reading your account closely enough.
If you have proposals in front of you and you want an unsentimental read on them, send them over. We will reduce each to its actual deliverables, mark which items would appear in any competent retainer, and flag anything we think is genuinely worth paying extra for, including where the honest answer is nothing. There is no charge and no obligation to work with us. Speak to us and you will get the same assessment we would give a client.
Natalie leads SEO strategy at Singapore SEO Agency, helping local and regional businesses build organic search programmes that drive qualified leads. She specialises in technical SEO and content-led authority building for Singapore SMEs.
Get a free SEO audit for your Singapore website — we'll show you exactly where you stand, what's holding you back, and what it would take to rank on page 1.
Get Your Free SEO Audit →
Yoast SEO helps most when its site-wide settings are right. See which settings to choose, what the traffic lights really mean and the mistakes to avoid.

Which free keyword research tool should a Singapore clinic, firm, contractor or tutor use? Combine five free tools to find your first 30-50 keywords. See how.

SEO vs SEM is usually the wrong question. Learn what the terms really mean and how to use ads and Search Console data to decide which searches to earn or buy.

Google Search Console login problems usually start with verification and ownership. Learn how to get in, fix access errors and offboard agencies safely.

The click through rate formula is clicks divided by impressions. Learn what each platform counts, the averaging trap and how to set it up in Google Sheets.

A click through rate means nothing on its own. Learn to compare CTR by position, query type and SERP features, and see what low CTR is really telling you.
Fast, no obligation. We reply within 24 hrs.
Singapore’s specialist SEO agency for SMEs. We rank your business on Google — and only Google. No distractions, just results.
© 2026 Singapore SEO Agency. All rights reserved.