What is answer engine optimization (AEO)?

AEO is the work of getting a business cited inside AI answers: on ChatGPT, in Google's AI Overviews, on Perplexity and Copilot. Where SEO tries to rank your site on a page of links, AEO is about the short list of firms the AI names when a buyer asks who to hire.

How is AEO different from SEO?

SEO gets you a spot on a results page the buyer still has to read through. AEO gets you named inside the answer itself. The work is different because engines build answers from passages they trust: your pages need to be written so a machine can lift the answer out, and you need to show up in the third-party sources the engines quote.

How does Mitra measure AI visibility?

Mitra samples 240 AI answers per audit. We take your buyers' real questions, phrased the way a homeowner actually types them, and run them again and again on live AI answer engines under a fixed protocol. Every firm named and every source quoted gets recorded, and the report states exactly which engines the runs used, so the before and the after are always measured on the same yardstick. Then we report your citation share with a margin of error, which tells you more than a screenshot of one lucky run.

Why does Mitra run each question so many times?

Because AI answers change between runs. The same question can name different firms on Tuesday and Thursday, so one screenshot proves nothing in either direction. 240 runs gives you a rate with a margin of error. That's a number you can act on.

How long until my firm gets cited?

The guarantee window is 90 days, and your money follows the proof. Your first invoice goes out only after the shared scoreboard shows AI naming your firm on buyer questions you approved. If day 90 arrives and the scoreboard reads below the agreed floor, we refund every fee you paid in the window within 7 days, up to three months of fees. AI engines refresh their sources at different speeds, so we re-measure monthly on the same protocol and you see the movement run by run.

Does every firm get the citation guarantee?

No, and that's deliberate. The free audit doubles as the screen: we offer the hard guarantee only where your baseline numbers say the gap is closable within 90 days. A firm starting from zero, where the sources AI reads don't mention it anywhere, usually needs a foundation phase first: normal month-to-month work, cancel anytime, no hard guarantee until the baseline can support one. If your market isn't winnable right now, the audit says so, you keep it, and we don't take your money. The guarantee stays rare so it stays real.

Is this a one-off report or an ongoing service?

The free audit is one-off, and you keep it either way. Paid work is ongoing: we measure, do the work, and re-measure every month, the way the rest of your marketing runs. Think of Mitra as the part of your marketing team that owns AI answers. Whether your buyers ask ChatGPT or type their reno questions into Google and read the AI answer at the top of the results, the first list of firms they see is written by an AI. Someone at your firm has to own showing up in it, and that someone is us.

What does the free AI visibility audit include?

The audit tells you where you stand today: which engines mention your firm, which competitors get cited instead, which sources those citations come from, and what we'd fix first. It's a written report, and it's yours to keep. It also comes with an optional walkthrough call where we read the tally with you and point at the first gap worth closing.

What's usually wrong when a good firm never gets cited?

Often nothing is wrong with the firm at all. The usual problem is that you're missing from the two or three sources the engines lean on for your city (a comparison page here, a directory there), and your own site doesn't answer any question in its first two sentences, so the machine has nothing to quote. Fixing that is unglamorous work, but most months it's what moves the number.

Is Mitra an AI company?

No. We don't build AI models. Mitra reads what AI answer engines say about firms like yours (measurement runs on Perplexity today, with more engines on the roadmap, and every report states exactly which engines the runs used) and fixes the sources those products read. The measuring is automated and unbiased. The judgment and the work that moves your number, deciding which sources to earn and building the pages that get quoted, are our team's. The point is that when a buyer asks, your firm shows up in the answer.

An SEO agency burned us before. Why would this go differently?

Because nothing here asks for your trust. Your baseline is measured before you pay anything, and your first invoice goes out only after the scoreboard shows AI naming your firm on questions you approved. The work is white-hat only, and every citation we claim is one you can click and check yourself. If day 90 reads below the agreed floor, we refund the fees you paid in the window within 7 days, up to three months of fees, and the cap sits in the agreement, not a footnote. Agencies that burn clients depend on you not being able to verify what they did. Our whole model depends on you verifying it.

Will any of this put my website at risk?

No. We don't buy links, plant reviews, or use tricks that get a site flagged. Everything we do is work the engines already reward: real content on your pages, real listings in the sources they trust, and citations you can click and check yourself. The firms that get burned took shortcuts. We build visibility that holds up, and you can audit every piece of it. Your domain is never the thing we gamble with.

Is AI search actually worth worrying about yet?

That is exactly what the audit answers, with a number instead of an opinion. Buyers haven't stopped using Google. Google now puts an AI-written answer at the top of the results, and most people read it and stop before the links. ChatGPT, Perplexity and Copilot work the same way. If the AI answer doesn't name your firm, you're invisible even to a buyer who still Googles. We measure how often those answers name your firm today, so you decide from your own data, not a sales pitch. If the answer is you're already fine, we'll tell you that too.

What does the fee buy, in business terms?

Enquiries from buyers who asked an AI who to hire. A typical Singapore renovation runs S$30,000 to S$80,000, and the Citation Sprint is S$1,200 a month, so one job that arrives through an AI answer covers the engagement several times over. To be exact about what we promise: we guarantee citations, measured on a fixed protocol, not jobs. But that arithmetic is why firms bother.

How do I know this isn't snake oil?

Because every claim we make is a number you can check. We measure the same way before and after, on the same questions, the same engines and the same counting rules, and you see every run. On paid work your first invoice waits for your first citation, any full month the scoreboard shows your name zero times is a month we don't bill, and if day 90 reads below the agreed floor, the fees you paid in the window come back, up to three months of fees. We would rather show you the receipts than tell you a story.

One screenshot proves nothing

§1

A screenshot is one run.

AI answers are probabilistic. The same question, asked the same way, can name a firm on one run and skip it on the next. A screenshot captures one run out of a distribution.

That cuts both ways. When a vendor shows you a screenshot with your competitor in it, they’re showing you one run. A screenshot with your own firm in it is weak evidence too. Neither settles the question that matters: out of all the answers buyers get, how many name you?

§2

What a real claim looks like.

A rate with a margin of error, from a fixed protocol. For example: named in 36% of 240 sampled answers, plus or minus 6 points at 95% confidence.

The rate says how often the engines cite you. The interval says how sure the measurement is. The fixed protocol, meaning the questions, engines and counting rules stay the same every time, is what makes two measurements comparable at all. Without all three, the number doesn’t tell you much more than a screenshot would.

§3

Why 240 runs.

Because the margin of error has to be smaller than the decision riding on it. Here’s what sample size does to the interval, at a measured share of about a third:

10 runs	±28 pts	worse than guessing out loud
20 runs	±20 pts	still an anecdote with decimals
60 runs	±12 pts	getting somewhere, not actionable
240 runs	±6 pts	tight enough to act on, and to re-check

At around 240 runs the interval gets small enough to tell a firm that’s never cited apart from one that’s cited some of the time. It’s also small enough to pick up real movement between two measurements, which is the point of measuring twice.

§4

What sampling noise does.

Sampling noise is the run-to-run variation that has nothing to do with real change. A small sample can show “improvement” that is luck.

If a firm moves from 2 citations in 10 runs to 4 in 10, nothing has been demonstrated. If it moves from 0% to 18% with intervals of plus or minus 5 on both measurements, something real happened. Any provider who guarantees results should be measuring rigorously enough for the guarantee to mean something.

§5

Four questions before you pay anyone.

Q1How many runs is that number based on?
Q2What is the margin of error?
Q3Is the protocol fixed between measurements?
Q4Will you show the raw counts?

Anyone measuring properly can answer all four on the spot, because the answers are the product. If what comes back is “proprietary methodology”, treat that as a no.

Mitra holds itself to the same standard. Every audit we run reports counts, intervals and sources, the free one included: run the audit on your firm →