A small firm, doing one thing

Elstrand measures whether AI assistants recommend consumer brands, and explains what decides it.

Why this exists

People have started asking AI what to buy. Not searching. Asking. The answer that comes back names three or four brands and moves on.

If your brand is one of them, you get the sale. If it isn't, you never find out. It doesn't show up in your analytics, there's no ranking to check, and no notification arrives. The channel is invisible from the inside.

That's the gap this firm exists to close.

How it works in practice

Every month we ask the questions your customers ask, record what comes back, and work out which sources the assistants are leaning on. Then we write you a ranked list of what to publish about it.

There's nothing to configure, nothing to install, and no onboarding call. A report arrives by email. It also sits in your account on our site, behind a code we send you, so you can find last month's without digging through your inbox.

How it's run

Elstrand is small, from Singapore, and intends to stay small. That's a choice about the work, not a stage we're trying to grow out of.

The person who reads your results is the person who writes your recommendations and the person who replies when you email. That person is Javier Ler, who founded Elstrand and runs it from Singapore. Nothing is passed down a chain, and nothing is generated and shipped unread.

The cost of that is capacity. We take a limited number of clients, and when the month is full, it's full. The benefit is that your report is written by someone who has actually looked at your category rather than run it through a template.

To be clear about how that is possible at all: one person does not type hundreds of questions a month. Questions that can be asked through an official, licensed API are collected automatically, three runs each, and that is where the breadth comes from. A deliberately small sample is typed by hand into the assistants' consumer apps, because those apps layer their own shopping data and local results on top and an API will never show you what your customer actually sees. Two of the seven assistants have no API at all, so they are hand-run by necessity.

What is never automated is the reading. Every answer that reaches your report has been looked at by a person, and the judgement about what it means is the thing you are paying for. The collection is plumbing. Anyone claiming to hand-type thousands of questions a month is either not doing it or not sleeping.

What you can check before paying us anything

This is a young firm and you have no reason to take any of it on trust. So rather than ask you to, here is what you can verify for yourself:

What this site deliberately doesn't have

No testimonials, no client logos, no "trusted by" strip. Not because they wouldn't help (they would) but because the firm is new and there is nobody to quote yet. Those are the easiest things on a page like this to invent, and inventing them is the fastest way to deserve the suspicion you arrived with.

When there are clients willing to be named, they will appear here with their names, and you will be able to go and ask them directly.

What Elstrand is, precisely

Elstrand is a Singapore-based business-to-business analytics and reporting service. It measures how AI assistants answer buying questions, and sells written monthly reports to consumer ecommerce brands. It is a research and reporting firm: it does not manufacture, sell or ship physical products of any kind.

The name is shared with unrelated companies in other industries, including sellers of outdoor and home storage goods. There is no affiliation, ownership or commercial relationship between this firm and any of them.

What we won't do

Start with a free check

Ten questions, real answers, your competitors named, no charge. If it's useful, we can talk about what a full report covers.

Request a free check