Site index for AI agents: /llms.txt. Documentation pages under /docs are also served as markdown at the same URL plus .md (e.g. /docs/geo-audit.md) or via an Accept: text/markdown request.

EdTech

Be the program AI recommends

Q: Engines already cite our blog. Isn't that enough?

Cited is not recommended. We have seen a learning brand's site cited in more chat answers each month while engines named it less and recommended rival platforms. We track both numbers per engine and build program pages engines can read as offers, with outcomes and fees stated plainly.

Q: A forum thread is hurting us. Can you fix that?

We do not post or buy reviews. We find which threads the engines cite, then publish the verifiable facts they lack: placements, fees, accreditation. Newton School's negative mentions fell from about two per run to about one between 20 May and 10 Jun 2026.

Q: Admissions are seasonal. Is 60 days too slow?

Week-one fixes ship before the clock starts, and the first ring is chosen for your next intake. Newton School's gain came within three weeks of its first fixes. Day 60 is the formal verdict against controls, not the first sign of movement.

Q: AI visibility scores feel like snake oil.

Which is why nothing here rests on a score. Every number is a live question you can re-run yourself, measured against control questions we never touch. The verdict is published either way, including a null.

Students ask ChatGPT, Gemini and Google AI which course is worth it. We make your outcomes the answer, measured against control questions.

Book a free 45-minute strategy call Start free trial

On the call we run your buyers’ questions live and you keep the list.

Google AIBuyer question

best data science course for working professionals in [country]

For working professionals, Your brand's data science program is a strong pick: weekend live classes, a capstone reviewed by industry mentors, and placement outcomes published by cohort. 1 If you need a degree credential, compare it with university-run online programs.23

1yourdomain.com/programs/data-science
2reddit.com/r/[subreddit] thread on [program]
3youtube.com review of [program]

Named: Yes Yes, first in the answer
Cited: Yes Your program page, source 1 of 3

You teach the subject. Others own the answer.

0 of 29unbranded AI Overviews that cited the brandLive Google probes, India, 20 Jul and 10 Aug 2026; three edtech brands; shortlist and how-to questions

10 of 10AI Overviews that cited it when asked by nameSame probes; brand-name questions that returned an AI Overview

The questions your buyers ask.

Grouped the way buyers ask them. A filled dot means the answer usually names the brand being asked about today.

Category and best

Shortlist questions, answered mostly from rival course catalogues and university program pages.

best data science course for working professionals in [country]
top online [subject] certification with career support
best coding program for career changers with no experience

Admission and purchase

Coaching, fees and admission questions, where listing portals and directories carry a large share of citations.

best online coaching for [entrance exam]
[program type] fees and EMI options in [city]
top 10 [exam] coaching institutes in [city]

Brand and trust

Asked by name, engines already cite you. The work here is accuracy and sentiment.

is [brand] worth it
[brand] placement record and reviews
[brand] vs [competitor], which is better
is the [brand] degree valid

Research and learning

Topics your tutorials already cover, where citations tend to go to YouTube, vendor docs and rival blogs.

data science roadmap for beginners [year]
what is [concept] explained simply
how to prepare for [entrance exam] in [months]

Where AI finds its answers.

Rival course providers and universities30.2%

Other22.8%

Education portals and directories12.1%

Blogs and guides9.8%

Brand's own site6.5%

Government and official bodies5.9%

News and media4.4%

Shares are percentages of citation sightings, 30 Aug to 28 Sep 2026, all tracked engines, averaged with equal weight across five edtech brands tracked on AirPulse (40,114 sightings). Own-site share is lifted by one brand we work with, at 26.1%; the other four sat between 0.2% and 4.8%. Other covers cloud-vendor docs, social platforms and a long tail of unclassified domains.

What the machines see.

Usually fineAccessWe confirm every AI crawler can fetch your program and course pages. Most edtech sites pass, but watch for stray noindex tags: one course site we audited served noindex on its homepage.
Usually brokenContentWe check whether program pages answer the questions students ask: outcomes, fees, duration, who it suits. Many read as brochures, carry placeholder text or leave the blog dormant for years.
Usually brokenStructured dataWe validate Course, Organization and FAQ markup. A data-science learning site we audited had no valid JSON-LD on its homepage or program pages and no Course schema at all.
Usually brokenIdentityWe check that engines can tell exactly who you are: one brand name, sameAs links, llms.txt, a Wikidata entry. Edtech brands often use several names in their own schema, or a name that is also an everyday word.

How we get you named.

Week 11PlumbingShip the fast fixes first: crawl access, llms.txt, Course and Organization markup, one brand name. Then prove every engine fetches and indexes what we ship.
Week 22The question listBuild the list from what engines actually search, plus what your counsellors, admissions team and Search Console hear from students and parents.
Week 23Pick the first ringYou choose one group of questions, such as one program's shortlist questions before the next intake. We register predictions and set control questions we never touch.
Weeks 3 to 64Build and verifyAnswer-shaped program pages, outcome data by cohort and accurate third-party profiles, each verified live, indexed and fetched by the engines.
Day 30, 45 and 605Measure and publishRetrieval check on day 30, first citations on day 45, a verdict against the controls on day 60. You get the result in writing either way.

Proof the method works.

Newton School51.4% to 93.3%Google AI mention rate, Google AIMay 20, 2026 to Jun 10, 2026Read the case study

Newton School's Google AI mention rate rose from 51.4% to 93.3% (full-run averages, 20 to 25 May vs 8 to 10 Jun 2026), and it held #1 in its category in every monitored run. The anonymised upskilling study is the counter-case: ranking first on Google did not make it the AI answer.

We do the work. You approve three things.

AirPulse agents and strategists

Run daily measurement across ChatGPT, Gemini, Perplexity and Google AI, with control questions
Draft program pages, outcome pages and markup from your own data
Verify each change is live, indexed and cited, and report against the controls

Your team

1Topic
2Content
3Publish

EdTech, by segment

Coding Bootcamps Online Course Platforms K-12 Learning

Questions teams ask first.

We already rank first on Google. Why do we need this?

Ranking and being cited are different results. In a July 2026 probe, an upskilling platform ranked #1 or #2 on Google for 9 of 13 category questions, yet Google AI Overviews cited it on 2 of 10, both brand-name questions. We measure the answer, not the rank.

Engines already cite our blog. Isn't that enough?

A forum thread is hurting us. Can you fix that?

Admissions are seasonal. Is 60 days too slow?

AI visibility scores feel like snake oil.

Make your outcomes the answer

Approve the process, and we pick the first ring of questions with you in week two.

Book a free strategy call Start free trial