The short version
- AI builds answers from many places. In one of our studies, only about a third of recommendations leaned on the firm's own website.
- So we work in ten places, not one. Your site, articles, LinkedIn, Reddit, video, directories, reviews, the press and more.
- The study decides the order. We close the biggest evidence gap first.
- Done for you or with your team. Nothing is published without your approval.
- No shortcuts. No fake reviews, no hidden accounts, no guarantees.
Why a page rewrite isn't enough
When an AI tool recommends a firm, it leans on whatever evidence it can find. Often that isn't the firm's own website.
- Our own data. In one study we traced 365 recommendations to their sources. 120 leaned on the firm's own site, 123 leaned on other sites, and 122 showed no source at all.
- The tools differ. In the same study, Perplexity showed the firm's own site in 58 of 66 recommendations. Gemini did in 12 of 112.
- Reddit. Profound analysed 680 million AI citations. Reddit was the most-cited domain in Google AI Overviews and in Perplexity. Wikipedia led in ChatGPT.
- LinkedIn. For professional questions, Profound later found LinkedIn was the most-cited domain across six AI platforms.
- Video. Ahrefs studied 75,000 brands and found that mentions on YouTube had the strongest link with being visible in AI answers. That's a correlation, not a proven cause.
Even the biggest source is a small slice. In Profound's data, Reddit made up 6.6% of Perplexity's citations and 2.2% of Google AI Overviews'. No single source decides the answer, so the work has to follow the evidence wherever it sits.
The ten ways we work
| Where | What we do |
|---|---|
| 1. Your website | Decision pages for each buyer situation, pricing and terms pages, proof beside every service, team and role pages, and rewrites of pages that describe the firm you used to be. |
| 2. Technical access | Indexing settings, robots rules, canonical addresses, sitemaps, and structured data where it helps. |
| 3. Articles and research | Answer-first articles for the questions buyers ask, plus original data and reports published under your name. |
| 4. LinkedIn | Your company page, leaders' posts and long articles, and study findings turned into posts people share. |
| 5. Reddit and forums | Finding the threads your buyers read and answering them usefully, under real names, with your firm disclosed. |
| 6. Video and podcasts | YouTube videos with transcripts, short clips cut from longer material, and podcast guest spots. |
| 7. Directories and profiles | Industry directories, Google Business Profile, Bing Places, your Businesses in ChatGPT profile and association listings, all stating the same current facts. |
| 8. Reviews and client proof | Review profiles your buyers trust, a simple way to ask clients for reviews, and client stories published with permission. |
| 9. Press and guest articles | Bylined pieces in trade publications, entries in credible lists, and credits from partners, suppliers and associations. |
| 10. Re-testing | The same buyer questions asked again after the changes, so you can see what moved. |
Your website comes first
Outside sources matter, but they point back to you. If your own pages can't answer a buyer's question, nothing else holds.
- In one study, a firm with dozens of well-written blog articles was cited in 0 of 160 answers. The pages AI did use answered what a buyer asks before a call: cost, time, and what you leave with.
- In another, one setting was keeping nine of nine sampled pages out of Google. That's case file 003.
- One AI tool recommended a firm and then said it couldn't verify the firm's prices. Its competitors had published theirs.
For a service firm, LinkedIn is where the people behind it can be checked. We plan and write posts and long articles for your leaders and your company page, and turn each study finding into something worth sharing. It's also where we publish our own findings first.
We have done this for years. For one specialist in medical education we ran LinkedIn, Instagram, short-form video and outreach together for nearly twelve months: read that case study.
Reddit and forums
Buyers ask each other for recommendations in communities, and AI tools read those threads. We find the ones that matter in your field and help your people answer real questions there.
We do it openly: real names, your firm disclosed, and answers that are useful even to someone who never hires you. Communities spot promotion dressed up as a customer's opinion, and it backfires.
Video and podcasts
A video with a transcript is evidence a machine can read and a buyer can trust. We script and edit videos, cut short clips from longer material, and arrange podcast guest spots. Video and social content is where Riseklix started. See how we turned long videos into short ones for a family entertainment brand and a music creator, or browse all our earlier work.
Directories, reviews and the press
These are the sources you don't control, which is why they are believed.
- In one of our tests, three of four AI tools described a firm as the business it used to be. They were reading old directory profiles.
- In the contractor study, one competitor's recommendations were repeatedly backed by a third-party directory that listed its past projects.
So we correct what's out of date, fill the profiles that are missing, help you collect reviews from real clients, and place bylined articles where your buyers read.
How we decide what to do first
The study ranks the gaps. We start where a small change removes the biggest doubt, and not every firm needs all ten. A typical first three months looks like this.
- Month 1. Make the pages buyers land on ready.
- Month 2. Answer the highest-priority buyer questions.
- Month 3. Build outside evidence, then check progress.
Done for you, or with your team
We can write, produce and publish for you, or brief your team and review their work. Either way, nothing goes out without your approval.
What we won't do
- Fake or paid-for reviews.
- Posts from accounts that hide who they are.
- Paid placements dressed up as editorial.
- Hundreds of near-identical pages.
- Schema or AI files sold as a guaranteed shortcut.
- Promise a ranking or a placement in any AI tool.
What it costs
Implementation is quoted per scope, after a study, and usually runs month by month. See pricing, or read how a study works first.
Sources. Profound, AI Platform Citation Patterns, 680 million citations, August 2024 to June 2025. Profound, LinkedIn is the most-cited domain for professional queries in AI search, March 2026. Ahrefs, Top Brand Visibility Factors in ChatGPT, AI Mode, and AI Overviews, December 2025. All other figures are from Riseklix studies.
Questions
Do I need a study first?
In most cases, yes. The study shows which evidence is missing and for which buyer situations, so the work goes where it can change answers. Without it we would be guessing.
Do you write and post the content yourselves?
We can. We write, produce and publish for you, or brief your team and review their work. Nothing is published without your approval.
Do you post on Reddit as our firm?
Only openly. We help your people answer real questions under their own names, with the firm disclosed. We don't use hidden or fake accounts.
How long until AI answers change?
It varies by tool, and nobody can promise a date. That's why we ask the same buyer questions again after the work, and track them monthly if you want.
Can you guarantee we'll be recommended?
No. Nobody can. We can make sure the evidence a tool needs is published, accurate and easy to find, and then measure what changes.