Financial services
Fintech
Technology firms delivering financial services directly.
Fintech firms sell to people who read before they buy: a CTO comparing gateways, a head of credit reading a fresh circular, a first-time investor, a policyholder holding a rejected cashless request. All four ask an engine first, and it answers with one firm's explanation. In a category built on documentation, the best-documented product gets named.
Where the answer is being lost
Fintech products are now evaluated before anyone opens the site.
A head of finance at a subscription business asks "Which payment gateway has the best success rate for Indian recurring payments" and gets three names, a reason for each, and a note about mandate failures. She has not opened a pricing page. The same happens to the investor asking how to start small and the policyholder told to pay first and claim later. Fintech writes its best material for people who have already signed up. The docs sit behind an integration, the console behind a login, the claims flow behind a policy PDF. Each of those explains the product properly, and none of them is reachable by the engine answering the question that decides the shortlist. So the demo request goes to whoever wrote in the open.
How we win this
The programme for fintech
Documentation is the asset
Payments and lending teams already write more reference material than any agency will: error codes, webhook behaviour, mandate flows, underwriting logic. Most of it sits inside a console, a login or a JavaScript app, so it may as well not exist. We move the explanations that answer buying questions into crawlable pages, structured the way an engine quotes them.
Dated pages that explain, never advise
We explain mechanics and never recommend: how a systematic plan is executed, what a waiting period does to a claim. Assured, minimum, target and risk-free returns are prohibited outright; past performance runs only where PaRRVA has verified it, and a settlement figure only with a published source. Regulated pages age, so each carries a date and an effective-from line, and monitoring tells us which version the engines quote.
Written for the approval queue
Advertising material at a SEBI-registered firm goes live when compliance has cleared it and a SEBI-recognised supervisory body has given prior approval. The approved version sits in a file for five years. That is a lead time, not a formality. We plan to it: batches sized to one approval cycle, every claim carrying its source, a calendar built around the queue. Insurance material runs the same way under IRDAI.
Own the failure path
Nobody searches when a product works. They search when the mandate fails, when the portfolio turns, when the hospital refuses cashless, when the app will not complete a transfer. Those questions are specific, they are asked at the exact moment somebody is willing to switch, and almost nobody answers them properly. That is the ground we take first.
The mix that carries it
Content
Answer and comparison pages
Cost, process, eligibility and comparison pages built for direct extraction, not for a reader who scrolls.
Content
GEO blogs and authority content
The definitive written answer to the questions your buyers put to an engine, structured so it can be lifted and attributed.
Foundation
Entity and schema engineering
Structured data and entity definition so engines know exactly what you are, where you operate, and what you are credible in.
Foundation
Technical fixes
Crawlability, render, speed and the machine-readability faults that keep an engine from reading you at all.
Measurement
AI Presence tracking
Standing measurement of inclusion, share of answer and competitor movement as models update.
The constraint we work inside
Two regimes govern this work. The investing side sits under SEBI's Advertisement Code for investment advisers and research analysts, issued on 5 April 2023 and consolidated into the Master Circular for Investment Advisers in June 2025. Insurance material sits under the equivalent IRDAI advertisement regime. The advice boundary is one rule among several: education must not become recommendation. The others bind whether a page recommends anything or not. No promise or guarantee of assured, minimum, target or risk-free return. No reference to past performance or risk-return metrics unless the figure has been verified by PaRRVA, the Past Risk and Return Verification Agency, which is a gate rather than a ban and the only lawful route to a number on the page. No superlatives, so no Best, no No. 1, no Top Adviser, no Leading. No use of the SEBI logo. Nothing that trades on an investor's inexperience, which is the clause that bites hardest on exactly the first-time-investor material this programme is built from. And advertising material needs prior approval from a SEBI-recognised supervisory body before it runs, with the approved copy retained for five years, so the publishing calendar is built around that queue rather than around a writer's availability. Regulated content also ages, so pages carry a date and are revised as circulars land. All of it makes the writing narrower and, usefully, more quotable.
Specialisations
4 total
The pitch is different for each one, because the buyer, the trigger and the rules on what may be published are different for each one. Open the one that is yours.
A CTO rebuilding checkout for subscriptions asks "Which payment gateway has the best success rate for Indian recurring payments", and the three gateways named in that answer are the three that get evaluated.
The question deciding this today
“Which payment gateway has the best success rate for Indian recurring payments”
- Who they sell to
- Merchants and platforms moving money between payers and sellers
- Who signs
- Founder, CTO, head of finance
- What starts it
- New checkout build, failure rate problem, expansion into a new market
- Cost of staying invisible
- Transactions failing at the point money should have moved
Two things decide whether that answer can name you, and both are checkable before a word gets written. First, whether your success-rate evidence exists anywhere a stranger can reach, or only in a sales deck and a sandbox that needs an account. Second, whether your documentation renders without JavaScript. Where it does not, the error-code reference, the mandate-debit flow and the retry logic are invisible to the thing your buyer just asked. What gets quoted in their place is an old competitor post, a review-site roundup, and whichever gateway wrote a plain page about UPI Autopay failures.
What we would run
- 01Technical fixes
We take the developer documentation out of the client-rendered app: server-rendered reference pages, one URL per error code and per payment method, sitemapped, with the sandbox examples visible without a login.
The CTO's question is answered from documentation. If a machine cannot fetch yours, it quotes whoever left theirs in the open.
- 02Answer and comparison pages
One page per failure mode and per method: recurring mandates on UPI and cards, what a specific decline code means, what happens on a bank outage, how tokenisation changes a saved-card charge.
These are the pages a head of finance sends to a CTO. Method-level specificity is the only thing a large content team cannot copy quickly.
- 03GEO blogs and authority content
Long-form engineering writing under a named author: how you handle retries and idempotency, what your reconciliation file contains, how a merchant migrates gateways without dropping live mandates.
Migration is the trigger. A merchant with a failure-rate problem is already leaving somebody, and wants proof you have moved a book before.
- 04Entity and schema engineering
Entity work on the company and each product: your authorisation status and regulator, the methods and currencies you actually settle in, the SDKs and platforms covered, the regions you collect in, and every pricing tier written as machine-readable data. A pricing table published as an image is a blank space to an engine.
An engine shortlisting gateways needs to confirm you are licensed and support recurring debit. Facts you never stated are treated as absent.
- 05AI Presence tracking
A standing set of gateway questions tracked across the major engines: recurring payments, international collection, specific decline codes, plus which rivals are named beside you and which pages the answers cite.
This market moves on launches and outages. When a rival's new page starts taking a method question, we would rather know inside a fortnight than at renewal.
What we would not recommend
- Reddit. Developer forums are a genuine channel here, but vendor answers about your own success rates get removed and read as pitching. We stay out.
- Reviews and testimonials. Success-rate testimonials are performance claims about money moving. Easy to write, impossible to substantiate at method level, and not what a CTO trusts anyway.
- Instagram. Nobody picks a payment stack from a feed. The buying committee here is technical and reads reference material, not carousels.
What a lead looks like
A head of finance at a subscription company, three weeks into a failure-rate problem, having read your decline-code page and your mandate-migration piece. She already knows how you handle bank outages. She wants to talk about her card-on-file volumes, her current provider's exit terms, and how long a switch takes.
What we measure
- Named for recurring-payment questions
- Decline-code pages cited by engines
- Docs fetchable without JavaScript
- Demo requests citing a specific page
- Share of answer against named rivals
What changes
A head of finance books a call already knowing your recurring-payment failure handling, and wants to talk volumes. A head of credit asks whether your stack covers the obligations he read about on your page. An investor opens an account without a sales conversation at all. A policyholder arrives at renewal because you were the one who explained reimbursement. Fewer enquiries begin with what do you do, more begin with can you handle this.
Start here
See who gets named in fintech today
We put your buyers' real questions to the live models and come back with the businesses they name, the sources behind those answers, and the gap between that list and yours.