Aions Partner · Guided

We find where you lose. And stay with it until the re-test.

Aions Partner is the guided programme for companies that want results, not another tool. We derive your decision questions, audit your shop up to a test purchase by an agent and define the actions. Then we coordinate implementation and measure again afterwards, with the same measurement.

Get in touch

Free readiness check

No account, no subscription, no cost.

Price on quotation

See the case studies

Partner project · Office chairs SwitzerlandPhase 3 of 4

01Questions and baseline✓ completed

02Diagnosis✓ 14 findings

03Implementationin progress

04Re-testplanned

Action	Implemented by	Status
Price and availability in the markup	Your agency	✓ confirmed
Deliver content without JavaScript	Implementation partner	in progress
Delivery time to Switzerland in the markup	Your agency	in progress
One return period on every page	Your team	open
Make the shipping method selectable for agents	Implementation partner	in progress

Test purchase by an agent

Stopped at the shipping method

For a Swiss address, the checkout only shows ‘collection’. The agent gives up.

Preview · sample data, fictitious brands

Who it's for

When you want the result, not the tool.

No in-house team for it

You know that AI recommendations are becoming important. But nobody in your company measures, analyses and follows up on actions.

A category that needs explaining

Insurance, technical products, selling to businesses: here, conditions and details decide. Standard questions fall short.

You have an agency, but no one steering

Your agency implements. You need someone who defines what needs doing and then measures whether anything has changed.

What you get

Everything from one source. From the first question to the re-test.

Question design: we develop the question set for your category with you in a workshop, split into core questions and exploratory questions.
Research: who wins which kind of question, who narrowly loses, which sources count.
Readiness with test purchase: audit per crawler user agent, manual content review, contradictory purchase conditions. Plus a test purchase by an agent up to the point of payment.
Tailored recommendations: prioritised, with acceptance criterion and responsible party, explained in a results meeting.
Coordinated implementation: by implementation partners or your agency, with a status for each action.
Re-test: the same questions, the same systems, the same duration as the first measurement. With a report, even if nothing changes.

How a Partner project works

Four phases. One measurement before, the same one after.

01Questions and baselineQuestion set in a workshop, then daily measurement on the systems your customers use.Responsible: Aions, with you and your agency

Measurement before

02DiagnosisReadiness per crawler, purchase conditions, test purchase by an agent. Result: prioritised actions with acceptance criteria.Responsible: Aions

03ImplementationAions coordinates, tracks the status and checks each action against its acceptance criterion.Responsible: Implementation partners or your agency

04Re-testThe same questions, the same systems, the same duration as the baseline, with intervals and a report.Responsible: Aions

The same measurement after

Always measured with the same method, whoever implements

Preview · sample data, fictitious brands

01 Questions and baselinewe develop the questions with you in a workshop and measure daily on the systems your customers use.
02 Diagnosisreadiness per crawler, purchase conditions, test purchase. Result: prioritised actions with acceptance criteria.
03 Implementationby implementation partners or your agency. Aions coordinates and checks each action against its acceptance criterion.
04 Re-testthe same questions and systems, the same duration as the baseline, with intervals. We show what has changed and what has not.

You receive a requirements catalogue with an ID, a classification and an acceptance criterion for each requirement. Your implementer works through it, and Aions checks acceptance.

Request a full example

Requirements catalogue · version 2.5 · extract1.1 Roles

Client: makes the commercial and legal decisions and supplies the facts: returns, shipping, payment, identities and product master data.
Implementer (in this example the client’s agency): implements pages, templates, the shop data model and the markup. Visible content and structured data come from the same source.
Aions: specifies the measurable requirements, checks the implementation against the first measurement and the technical criteria, and measures again.

3.2 Returns, shipping and payment

Finding from the first measurement: Return, shipping and payment information is contradictory. The technical audit finds no machine-readable return or shipping rules. The client first settles the values and where they apply.

ID	Requirement	Classification	Benefit	Acceptance criterion
POL-01	The standard return rule is marked up on the shop operator (MerchantReturnPolicy). Product-specific deviations only where genuine exceptions exist.	COREbinding	high	A complete standard policy on the returns page, linked to the shop operator. Country, category and return window match the approved values.
POL-02	The standard shipping logic is marked up on the shop operator (ShippingService): delivery area, cost, free-shipping threshold and delivery time. Offer-level details only for deviations.	COREbinding	high	Cost, threshold, countries and delivery times match the shipping page and the checkout, tested below, exactly at and above the threshold and for each delivery area.
POL-04	Information pages, the checkout, legal pages and the markup use the same approved values for returns, shipping cost, free shipping, delivery area and payment.	COREbinding	high	A comparison of the relevant information pages and a sample of five product pages shows no factual deviations.

CORE: binding for the agreed scope, open items block acceptance · RECOMMENDED: recommended, does not block acceptance · Benefit: estimated contribution to data quality and machine readability (high, medium, supporting), not a promise of visibility

Extract from a real catalogue, anonymised · Full example on request

Only in the Partner programme

We let an agent shop with you.

A browser agent goes through your shop up to the point of payment. It shows where it gets stuck. Typical hurdles: bot protection, captcha, mandatory account, cookie banner, Swiss address, shipping and payment methods.

Buying through AI agents is still uncommon in Switzerland today. That is why we mark this stage as preparation.

Only with your authorisation. Stops before payment, no test orders. If bot protection blocks the agent, that is itself the finding.

Test purchase log · product page to paymentAgent in the browser

passed: Product found via search4 s passed: Variant chosen, added to basket9 s passed: Guest checkout without accountpossible passed: Swiss address acceptedpostcode checked stopped: Shipping methodonly ‘collection’ shown not reached: Paymentnot reached

Finding NO_SHIPPING_OPTION_CH · Action: deliver the shipping options for Switzerland server-side · Acceptance: the agent reaches the payment page

Stage ‘purchasable’, marked as preparation · stops before payment, no test order

Preview · sample data, fictitious brands

Implementation

Implementation: by our implementation partners or your agency, coordinated by Aions.

Aions does not implement itself. We define what needs doing, coordinate the implementation and check each action against its acceptance criterion.

If your agency has the necessary experience, it implements. On request, we suggest implementation partners from our network. If you come to us through an agency, that agency implements. You sign the contract directly with the company that does the work.

We measure before and after with the same questions, systems and rules, whoever implements. Denominator and interval remain visible to you.

Examples from our work

What we found.

Case A · Swiss online shop

Well known, but rarely mentioned in shop comparisons.

Starting point: the shop is well known. Among 47 brands counted, it ranked 13th. We did not give the models any competitors.

Measurement: for questions about the manufacturer, a product feature or delivery time, the shop appeared in 18% of answers. For questions about the best shop, a shop comparison or returns, in 3%. Of the answers that recommended exactly one shop, none chose it (0%). We formed the groups after the measurement.

The more the question focused on returns, the less often the shop appeared: 14%, 5%, 0%. Three questions show a direction, not an established difference.

Finding on its own pages: three statements on returns appeared there at the same time. A person usually reads one of them; an AI system can read all of them. Without a comparison group, the measurement does not show whether the contradictions explain the difference.

Recommended: clean up the purchase conditions, that is, make returns, shipping costs, delivery time and payment methods unambiguous, consistent and machine-readable. Then measure again with the same questions.

Case B · Swiss manufacturer selling to businesses and consumers

For crawlers without JavaScript, the website was empty.

Starting point: anyone fetching the pages without JavaScript saw an empty page.

Changed: the way the pages are delivered. Afterwards, the content was visible to 6 of 10 crawler user agents checked. The assistants' live fetches still receive an empty page.

Observed after the change: on the same 6 access paths, the Mention Rate rose from 4.0% to 6.6%. The 95% intervals of 2.9 to 5.5% and 5.1 to 8.5% only just overlap. By our rule, that is not yet an established difference.

Across all access paths in the re-test, the figure is 8.7%. However, this includes systems that were not measured in the baseline. Only the 6 access paths are comparable.

Next lever: the assistants' live fetches.

Mention Rate on the same 6 access paths, with 95% intervalCase B, anonymised

Baseline
August 2026

4.0% · 2.9 to 5.5%

Re-Test
30 Sep to 5 Oct 2026

6.6% · 5.1 to 8.5%

4.0%6.6%on the same 6 access paths, observed after the change. The intervals only just overlap.

8.7% across all access paths in the re-test. Not directly comparable: the re-test measures additional systems that were missing from the baseline.

Real measurement, anonymised

Price

Price on quotation. Clearly broken down.

We quote for questions, baseline, diagnosis and re-test according to scope. The re-test uses the same questions, the same systems and the same duration as the first measurement.

Implementation is billed directly by the company that does the work, or by your agency.

Let's talk about your category.

In a first conversation, we clarify which questions decide for you and where a Partner project would start.

Get in touch

Free readiness check

No account, no subscription, no cost.