# How to create original research for AI-search authority. - working worksheet

Reviewed alongside the Eli guide: https://hireeli.io/resources/original-research-for-ai-search-authority

## Decision to make

Publish a transparent dataset that people and answer engines can use

Owner: ____________________
Review date: ____________________
Canonical page or destination: ____________________

## Buyer questions

- [ ] How to create original research for AI-search authority?
- [ ] What should a B2B startup do first to publish a transparent dataset that people and answer engines can use?
- [ ] Which evidence shows that publish a transparent dataset that people and answer engines can use is working?
- [ ] What is the most common mistake when teams try to publish a transparent dataset that people and answer engines can use?

## Evidence-led work plan

### 1. Why this decision appears: a survey headline without sample

A survey headline without sample, dates, questions or limitations gives publishers and buyers little reason to trust or cite it.

Evidence or artifact: ____________________
Status: [ ] not started  [ ] working  [ ] verified  [ ] blocked

### 2. Start with a decision-relevant question, publish the method and usable result, and make every claim traceable to the data

Start with a decision-relevant question, publish the method and usable result, and make every claim traceable to the data.

Evidence or artifact: ____________________
Status: [ ] not started  [ ] working  [ ] verified  [ ] blocked

### 3. Define the population and collection window, then publish the questions or coding method

Define the population and collection window.

Evidence or artifact: ____________________
Status: [ ] not started  [ ] working  [ ] verified  [ ] blocked

### 4. Measure track qualified citations

Track qualified citations, dataset use, methodological corrections and relevant referral visits.

Evidence or artifact: ____________________
Status: [ ] not started  [ ] working  [ ] verified  [ ] blocked

### 5. Do not cross this boundary: do not label generated examples

Do not label generated examples, scraped private data or an undisclosed convenience sample as market research.

Evidence or artifact: ____________________
Status: [ ] not started  [ ] working  [ ] verified  [ ] blocked

## Source ledger

- [ ] Google Search Central: optimizing for generative AI features: https://developers.google.com/search/docs/fundamentals/ai-optimization-guide
- [ ] Google Search Central: people-first content guidance: https://developers.google.com/search/docs/fundamentals/creating-helpful-content
- [ ] Reddit Help: spam policy: https://support.reddithelp.com/hc/en-us/articles/360043504051-Spam
- [ ] OpenAI: publishers and developers FAQ: https://help.openai.com/en/articles/12627856-publishers-and-developers-faq

## Measurement boundary

Baseline date and exact context: ____________________
7-day observation: ____________________
14-day observation: ____________________
30-day observation: ____________________
Known limitations: ____________________

Keep technical eligibility, citations, recommendations, visits, leads and revenue as separate evidence. This worksheet does not guarantee an external ranking or commercial result.