LLM OptimizationWrite it so a model can quote it.
We restructure your pages so a language model can lift a correct and complete answer without filling in gaps.
- Passage level answer blocks
- Entity and fact clarity
- Source and proof on the page
- Re check across engines
Addition is a marketing agency for ecommerce and B2B brands
We turn your searches into paying customers.
A model repeats what your page made easy to repeat.
LLM Optimization problems we solve.
Do you say even one of these?
-
Our pages are written for people and the models still get us wrong.
-
The AI assistant quotes an old price from somewhere and we cannot find where.
-
Our best information is locked inside a PDF nobody can read.
Three promises. Hold us to them.
- We restructure your key pages so a model can lift a complete answer from them.
- We put the source and the proof on the page, so nothing has to be assumed.
- We re check across engines, because they do not all read the same way.
A roadmap no prompt can write.
Ideation, strategy and the roadmap all happen in these six steps. This is the part no AI can replicate, and it is where the value of our work comes from.
-
Step 01
Onboarding
We take access, agree the prompts that decide a purchase, and set the single metric we will be judged on.
-
Step 02
The audit
Each dated run shows which brands every engine names for your buying prompts.
-
Step 03
Brand DNA
We correct the facts about your product and add the proof behind them, so an AI assistant does not have to guess.
-
Step 04
Market & Competitors
We read the sources the answers are built from and name what the recommended brands have that you do not.
-
Step 05
Roadmap & Strategy
We start with the missing information that has the greatest effect on the answer.
-
Step 06
Execution
We make the changes and re-measure against the same prompts, so the comparison shows what changed.
Where do you need support?
Questions people ask about LLM Optimization
What is LLM optimisation?
Restructuring your key pages so a language model can lift a correct and complete answer from them without filling the gaps on its own. Clear headings, one answer per section, and the facts in the page rather than in an image or a script.
What is LLM optimisation versus SEO?
SEO gets the page found and ranked. LLM optimisation gives a model as little as possible to fill in when it reads the page, so the answer it extracts is more likely to be the one you meant. A page can rank well and still be summarised wrongly.
Do we need an llms.txt file?
It is quick to add, but easy to forget to update. There is no published evidence yet that the major assistants use it to choose sources, so it is not worth prioritizing. Keep the pages themselves current, with every claim tied to a source.
What do you change on a page?
Headings that name the question, an answer block that stands alone, specifications in text rather than images, and structured data that matches what the page says. Every change must also make the page easier for a person to read.
Find out what AI says about you.
Then be the brand it names.
We run the prompts that decide a purchase in your category, record who gets named instead of you, and show you the content and proof changes that would put you in the answer.