Definition:
PASO, an acronym for Personal Assistant Search Optimization, is a label proposed to group practices intended to increase the likelihood that an organisation’s content will be found, interpreted, or used by digital assistants such as Siri or Alexa. The term appeared in 2017 in an article by John E. Lincoln about optimisation for personal assistants.
PASO is not a discipline officially recognised by the main assistant providers or a separate ranking system. It describes an area of work combining SEO, structured data, local information, integrations, and content design, but each assistant uses different sources and mechanisms.
A voice query is not always equivalent to a web search. The assistant may open results, answer with information from an app, perform an action, or generate a synthesis from several sources. Therefore, there is no single technique that guarantees a page will be selected as the answer.
Table of contents
Origin and evolution of PASO
The expression Personal Assistant Search Optimization was popularised by John E. Lincoln in the article Why the future is all about PASO. At the time, smart speakers, mobile assistants, and voice search were presented as a new environment for SEO.
Assistants have since expanded their functions. In addition to retrieving search-engine results, they may consult knowledge bases, local services, apps, catalogues, and generative systems. PASO remains useful as a descriptive label, but its current scope is broader than optimising a page for a spoken question.
The way answers are presented has also changed. An assistant may show several links, cite sources, request clarification, combine information, or return no web answer. The former automatic association between voice, zero position, and the first organic result does not represent all these cases.
Where personal assistants obtain information
The source depends on the device, query, location, permissions, and agreements of each platform. Five possible sources and signals include:
- Web indexes: pages crawled and indexed by search engines or a platform’s own robots.
- Knowledge graphs and databases: entities, attributes, and relationships used to answer factual questions.
- Connected apps and services: information and actions provided through integrations, APIs, or device-specific functions.
- Local and commercial sources: business profiles, maps, opening hours, inventory, and other location-related data.
- Generative systems: models that may synthesise an answer from several sources and may or may not display supporting links.
Apple explains that Applebot obtains data used in experiences such as Siri, Spotlight, and Safari. Amazon distinguishes in its documentation about Amazonbot between robots that power its search experiences and requests that retrieve current information to answer Alexa queries. These examples show that there is no common index or single selection rule for all assistants.
Technical access is only an initial condition. A platform’s ability to crawl content does not guarantee that it will index, cite, or convert it into an answer.
What can be optimised for assistants and voice search
PASO shares most of its foundations with SEO and does not require artificially conversational phrases. These six areas can improve content accessibility and comprehension:
- Solve a specific need: identify the questions, actions, and contexts for which the page provides a useful answer.
- Support crawling and indexing: use accessible HTML, internal links, canonicals, and directives compatible with the relevant robots.
- Structure the information: use headings, paragraphs, lists, and tables that clearly express the relationship between the question and answer.
- Add valid structured data: apply Schema Markup when it matches the content, without assuming it will generate a voice feature or rich result.
- Maintain consistent local data: review name, address, phone number, category, opening hours, and location when the query has local intent, as part of a local SEO strategy.
- Check the actual result: test relevant queries and devices, review logs, and analyse available traffic without attributing every change to voice.
Spoken queries may be longer or phrased as questions, but this tendency does not require repeating every variant literally. Content should answer using the necessary vocabulary while remaining natural for people.
Measurement has important limits. Search tools do not fully identify every query made by voice or every answer spoken by an assistant. Impressions, clicks, and visits describe only the observable part of the journey.
Differences between PASO, SEO, and GEO
PASO overlaps with other approaches but focuses on mediation by an assistant. The differences can be summarised in four points:
- SEO: improves a site’s accessibility, comprehension, and presence in search-engine results.
- PASO: applies discovery and structuring practices to queries or actions managed through personal assistants, with or without voice.
- GEO: works to make an entity and its content understandable, selectable, and citable in generative answers.
- Local and app optimisation: manages sources, profiles, and integrations that may supply the assistant without relying on a ranked page.
One case may belong to several areas. A local question addressed to an assistant may use a web index, a business profile, and a generative system in a single interaction. Separating these layers helps determine which source must be corrected and which metric can be evaluated.
PASO has no universal positions, ranking factors, or shared console. Nor does it guarantee that an assistant will mention a brand or send a visit. Its value lies in recognising that discovery can occur through interfaces other than a traditional results list, provided the sources, limitations, and observed outcome are documented.