How Erlin works
How Erlin suggests prompts, collects AI answers and calculates its metrics, why results change between runs, and what the numbers can and cannot tell you.
Erlin helps you choose questions to track, collect AI answers and investigate what those answers say about your business. Prompt Discovery estimates demand. Tracked runs provide answers to inspect. Recommendations connect findings with possible actions.
Understanding those separate steps helps you use the results without treating an estimate as an observed customer visit.
How Erlin recommends prompts
Prompt Discovery combines partner data with Erlin's proprietary algorithm to identify relevant questions, estimate volume and assess buying intent.
Estimated volume describes demand for a prompt. It is not an exact count of every conversation across AI platforms.
Buying intent helps distinguish a question about learning a topic from a question about choosing a provider or product. Use it alongside relevance: a question matters only when it relates to customers your business can serve.
See Choose the prompts to track.
What a tracked run collects
A tracked run collects AI answers to the prompts you selected. Erlin uses a combination of collection methods. The result is an answer from a particular check, not a record of a private buyer conversation.
Choose the available platforms for the questions you want to investigate. Checks run from your brand's office locations. When the brand has two or more locations, Run prompts from lets you choose which ones. In Prompt Discovery, Language for suggested prompts sets the language of the prompts Erlin suggests.
A check's location gives it its intended context. It does not establish that every person in that market receives the same answer.
How to interpret the metrics
The metrics summarize different aspects of the collected answers. Use the linked definitions when you need the exact calculation.
| Metric | The question it helps answer |
|---|---|
| AI visibility | How often do the included answers mention the brand? |
| Share of voice | How does the brand's presence compare with other brands in the included answers? |
| Average position | Where does the brand appear when an answer lists brands? |
| Citation rate | How often do the included answers cite the website? |
A mention is not a visit, and a visit is not a sale. Read an answer and its cited sources before interpreting a summary as evidence of a business outcome.
Brand mention rules let you define accepted names and spellings. Changes to recognition can affect how a result is counted without changing what the AI platform said.
Why results change
Results can change when you select different prompts, platforms, run locations or periods. Answers can also vary between runs.
Keep those settings consistent when you want to compare like with like. Review individual answers to understand a change instead of assuming that one movement reflects every customer's experience.
When recommendations appear
Erlin needs each prompt to have run at least 4 times on different days in the last 90 days, and you need to have picked competitors. One run is not enough, because AI answers change from run to run. Until then, Find new tasks is off and its hint reads, for example, "Run your prompts at least 4 times first (3 of 12 ready)."
Once the gate is met, Erlin also finds tasks on its own after a scheduled run and after you run a single prompt. Running all your prompts by hand does not start the search. Finding tasks costs no credits.
A recommendation is a proposed next action based on the evidence Erlin collected. Review the finding and relevant pages before accepting the work. Further runs help you reassess the answer after a change.
See Turn recommendations into actions.
How scheduling and access work
Use the run settings to choose the supported platforms and a schedule: Weekly, Every 2 weeks, Monthly or Off. New brands start on Monthly. Each run costs 2 credits per prompt per platform that answered.
Usage limits, permissions and credit allowances can differ. Consult Plans and billing and Credits and usage before committing to a run schedule.
How this connects to business outcomes
Erlin connects lead journeys across days and channels and links a lead to a won deal and its revenue when that information is available. Cross-device tracking is not supported.
Those records help you investigate the path to an outcome. They do not establish that a single prompt result or content edit caused the revenue.
For the measurement details, continue to lead journeys and revenue.