About Elementera: answer engine optimization (AEO) company in Vancouver
We are AI & ML experts in AI search visibility.
Elementera is an AI search visibility company in Vancouver, Canada. You know your business better than anyone, but you don’t have time to write about it. Elementera takes content off your plate from A to Z.
Four paths.
Decades of experience.
One emerging problem.
The advent of Generative AI is changing what comes next. Increasingly, people are no longer searching the internet themselves. They are asking AI systems to search, interpret, compare, evaluate, and recommend on their behalf. That creates a fundamentally different challenge for businesses. It is no longer enough to be visible online. Brands need to speak the language AI models understand to be trusted and ultimately recommended by them. Elementera was created to solve that problem. It began with four people who had spent decades, independently, working on the technologies and problems that would eventually make this challenge possible.This is our story.
Your Team
Ready to Deliver
An expert team of scientists, engineers, and product strategists delivering bespoke solutions with the scientific rigor of top research institutions & the technical mastery of leading tech firms.
From Search Optimization to AI Communication
For years, the four founders were pursuing different problems. But they kept crossing paths. They worked together. They mentored together. They built things together. They debated technology together. And they watched AI change, together. The four founders came together after decades of building intelligent systems, conducting AI research, and creating software and products on a global scale. They saw a fundamental shift taking place: as AI becomes the new gateway to discovery, brands must do more than simply be visible. They must be understood, trusted, and recommended by AI. That is why we built Elementera: because your next customer may not explicitly search for your service. Instead, they may describe what they need in a long, natural-language prompt and let AI determine which businesses best meet their needs.
Elementera’s mission is straightforward: Make brands understood. Make them trusted. Make them recommended.
Our Vision
By 2030, Elementera will be the benchmark for how original content is created and quantified.
Our Mission
We turn what founders, small-business owners and creators know into original content for AI search, and we measure the results engine by engine, with the rigor of the research we come from.
Read more about Elementera
What does Elementera do?
You know your business better than anyone, but you don’t have time to write about it. When you use AI to fill the gap, it sounds like everyone else, and ChatGPT and Google pass you over.
Elementera takes content off your plate from A to Z. First, Elementera Lenz learns your business objectives and vision, and shows you which topics you need to win to reach them. Then Originum interviews you and draws out what only you know: your insights, your take on the industry, the customer stories that prove it.
We turn that knowledge into your digital footprint: blog posts, service pages, FAQs and social posts, each built to win one of those topics. Originum gives each draft an originality score: how much of it comes from your own material rather than from what is already published.
The result: when a potential customer asks an AI model about a problem you solve, it understands your business, trusts it and recommends you.
Our work has three parts:
Who is Elementera for?
Elementera is for startup founders, small-business owners and content creators in Canada and the United States who know their field and have little time to write, and for marketing leads and heads of growth at SaaS companies, startups, fintech and other technology companies. The situations we see most often are a rebrand or new product that AI engines have not caught up with, a move upmarket where enterprise buyers do their research through AI, comparison and "best X for Y" questions where competitors are named first, and investor or buyer due diligence done partly through an AI assistant.
Security, data handling, fit and how results are judged: Security, data and fit
What research is Elementera built on?
Elementera is built on machine learning research and patents by its co-founders, most of it peer-reviewed. The work covers how learning systems make decisions, what they pick up from their training data, and how to measure them, which is the groundwork for measuring AI engines. Each title links to the paper.
Elementera works on AI search visibility because of the founders' research background: an AI answer engine is a learning system that decides which sources to read and which companies to name, and how such systems decide, what they learn from their data and how to measure them carefully are the questions the founders had been publishing on for years.
Dr. Maziar Gomrokchi did his PhD at McGill University and Mila, supervised by Doina Precup, according to his own homepage. His full record is on Google Scholar.
- Differentially Private Policy Evaluation. Balle, Gomrokchi, Precup. ICML 2016.
- Locally Persistent Exploration in Continuous Control Tasks with Sparse Rewards. Amin, Gomrokchi, Aboutalebi, Satija, Precup. ICML 2021.
- Membership Inference Attacks Against Temporally Correlated Data in Deep Reinforcement Learning. Gomrokchi, Amin, Aboutalebi, Wong, Precup. IEEE Access, 2023.
- AdCraft: An Advanced Reinforcement Learning Benchmark Environment for Search Engine Marketing Optimization. Gomrokchi, Levin, Roach, White. arXiv preprint, 2023.
AdCraft is the closest of these to our work today. It models how automated systems decide which advertiser to show for a search, which is a close cousin of the decision an AI answer engine makes when it chooses which company to name.
Soroush Mehri did his research at Mila and Université de Montréal and then at Microsoft Research, according to his GitHub profile; his team page places the Microsoft work at Microsoft Research Montreal. His papers are listed on arXiv.
- On Binary Classification with Single-Layer Convolutional Neural Networks. Mehri. arXiv preprint, 2015.
- SampleRNN: An Unconditional End-to-End Neural Audio Generation Model. Mehri, Kumar, Gulrajani, Kumar, Jain, Sotelo, Courville, Bengio. ICLR 2017. First author.
- Rapid Adaptation with Conditionally Shifted Neurons. Munkhdalai, Yuan, Mehri, Trischler. ICML 2018.
- Increasing Robustness to Spurious Correlations using Forgettable Examples. Yaghoobzadeh, Mehri, Tachet, Hazen, Sordoni. EACL 2021.
- US patent 11,833,422 B2, Providing automated user input to an application. Assigned to Microsoft, granted December 5, 2023. This is the US patent his team page refers to.
The spurious-correlations paper studies when a language model answers from a shortcut it learned in training instead of from the evidence in front of it. That is one of the reasons an AI answer can describe a company wrongly.
Dr. Mehrdad Mehrbod is the first-listed inventor on US patent 10,508,305 B2, DNA sequencing and processing, granted December 17, 2019.
When did Elementera start, and what came before it?
Elementera was founded in 2024, and the founders' published research goes back to 2015. The timeline below lists only dates we can point to a public record for.
| Year | What happened | Record |
|---|---|---|
| 2015 | Single-author arXiv paper by Soroush Mehri, on convolutional neural networks | arXiv 1509.03891 |
| 2016 | Differentially Private Policy Evaluation, ICML (Dr. Maziar Gomrokchi) | arXiv 1603.02010 |
| 2017 | SampleRNN, ICLR, first author Soroush Mehri | arXiv 1612.07837 |
| 2018 | Rapid Adaptation with Conditionally Shifted Neurons, ICML (Soroush Mehri) | arXiv 1712.09926 |
| 2019 | US patent 10,508,305 granted, first-listed inventor Dr. Mehrdad Mehrbod | Google Patents |
| 2021 | Locally Persistent Exploration, ICML (Dr. Maziar Gomrokchi); spurious correlations paper, EACL (Soroush Mehri) | arXiv 2012.13658, arXiv 1911.03861 |
| 2023 | AdCraft, a reinforcement learning benchmark for search engine marketing (Dr. Maziar Gomrokchi); US patent 11,833,422 granted, assigned to Microsoft (the patent on Soroush Mehri's team page) | arXiv 2306.11971, Google Patents |
| 2024 | Elementera founded | |
| Since 2024 | Managed AEO services, then Lenz | this page |
How we work
We measure before we claim. These are the rules we hold ourselves to.
- We ask each question more than once. In our research measurements every question is asked at least three times, and repeated on another day before we believe a change.
- We never pool engines. ChatGPT, Google AI Overviews, Perplexity, Claude and Gemini read different parts of the web. A result measured on one engine is reported for that engine only and never moved to another.
- We publish a finding only with its scope. Every finding we quote names the engine, the kind of question (for example, searches that do not name a company) and the sample it was measured on, and it is checked separately before it is published on this site.
- We record which model answered. Every measurement call records the model we asked for and the model the provider says actually answered, so a quiet model change does not pass as a real result.
Who runs Elementera?
Elementera is run by its four co-founders. Each name links to a page with that person's background and published work.
- Dr. Maziar Gomrokchi, Co-Founder and CEO. Reinforcement learning, privacy-preserving machine learning and multi-agent decision making; PhD from McGill University and Mila.
- Soroush Mehri, Co-Founder and Chief AI Officer. Neural audio generation, meta-learning, robustness of language models and reinforcement learning; previously at Mila and Microsoft Research.
- Dr. Mehrdad Mehrbod (Dr. Med), Co-Founder and Chief Product Officer. Dr. Med leads product, including Lenz, and is the first-listed inventor on a US patent.
- Mojtaba Ebadi, Co-Founder and Chief Operating Officer. Mojtaba runs operations and engineering delivery.
Frequently asked questions
Is Elementera an SEO agency?
No. We work on how AI answer engines read, trust and recommend companies. Search rankings are one input to that.
What is Lenz?
Lenz is our measurement product. It writes buyer questions from your website, asks AI engines those questions without naming you, and reports how often each engine names you, engine by engine.
Where is the research published?
The founders' papers are on arXiv and in the proceedings of ICML, ICLR and EACL, linked above.
Why does a company founded in 2024 talk about research from 2015?
Because the founding date and the expertise are different facts. The company was founded in 2024; the founders' work on how machine learning systems learn, decide and can be measured goes back a decade, and every paper is linked above.
Sources
- Elementera company page on LinkedIn (founding year): linkedin.com/company/elementera
- Dr. Maziar Gomrokchi, homepage: maziarg.github.io; Google Scholar: profile
- Soroush Mehri, arXiv author page: arxiv.org/a/mehri_s_1; GitHub: github.com/soroushmehr
- Balle, Gomrokchi, Precup. Differentially Private Policy Evaluation. ICML 2016. arXiv 1603.02010
- Amin, Gomrokchi, Aboutalebi, Satija, Precup. Locally Persistent Exploration in Continuous Control Tasks with Sparse Rewards. ICML 2021. arXiv 2012.13658
- Gomrokchi, Amin, Aboutalebi, Wong, Precup. Membership Inference Attacks Against Temporally Correlated Data in Deep Reinforcement Learning. IEEE Access, 2023. arXiv 2109.03975
- Gomrokchi, Levin, Roach, White. AdCraft. Preprint, 2023. arXiv 2306.11971
- Mehri. On Binary Classification with Single-Layer Convolutional Neural Networks. Preprint, 2015. arXiv 1509.03891
- Mehri et al. SampleRNN. ICLR 2017. arXiv 1612.07837
- Munkhdalai, Yuan, Mehri, Trischler. Rapid Adaptation with Conditionally Shifted Neurons. ICML 2018. arXiv 1712.09926
- Yaghoobzadeh, Mehri, Tachet, Hazen, Sordoni. Increasing Robustness to Spurious Correlations using Forgettable Examples. EACL 2021. arXiv 1911.03861
- US patent 11,833,422 B2. Google Patents
- US patent 10,508,305 B2. Google Patents