Answer Shape answer_shape
The shape the model plans for its answer before writing it: a single pick, a ranked shortlist, a comparison table or a how-to-choose guide. For "best roofer in Tulsa" it usually settles on a short ranked list.
answer.presentation (was query.context in the provisional inventory: Plans answer format: single pick, shortlist, table) · also called response_format, output_structure
Used in 2 of 100 census runs this week: “Provide a brief intro in Spanish, list the top 3 services with key strengths, include contact details and why they rank highest, and add a d”
Explicit Intent explicit_intent
What the query literally asks for, read at the interpretation stage: the trade, the place and any stated qualifier. In "emergency plumber in Mesa open now" the explicit intent is a plumber, in Mesa, available immediately.
What you can do: Name the exact service and city in page titles and headings, in the words customers type, so the stated request matches your page without interpretation.
query.intent · also called clear_intent, stated_goal
Term: search-intent
Used in 1 of 100 census runs this week: “provide_recommendation_for_trusted_hvac_contractor_in_austin_tx”
Hidden Intent Possible hidden_intent_possible
A flag that the query may carry an unstated goal, such as "tree removal near me" really meaning storm damage that needs fast, insured work. The model then considers businesses that fit the likely underlying need.
What you can do: Cover the common reasons behind your service on the site, e.g. a storm-damage page with response time and insurance details, not just a generic services list.
query.intent · also called implicit_intent, subtle_goal
Term: intent-coverage
Risk Assumption risk_assumption
The model's working assumption about how much can go wrong for the person asking, set while it interprets the query. Gas line work or roofing is assumed higher risk than buying a pet bed, which raises the bar for recommended businesses.
What you can do: State licences, insurance and safety practices plainly on the site, especially for work where mistakes are costly or dangerous.
evidence.assessment (was query.context in the provisional inventory: Judges how much can go wrong; risk assessment) · also called risk_factor, safety_assessment
Excluded Or Lower Priority excluded_or_lower_priority
Things the model decides to leave out or push down when scoping the answer, such as national chains or lead-generation directories when someone asks for a local, independent landscaper.
What you can do: Make clear on your site that you are the actual provider, not a referral service, and that you operate locally, so you are not filed with the deprioritised group.
query.scope · also called deprioritized_items, excluded_elements
Included Services included_services
The specific services the model treats as part of the request when scoping it, for example gutter cleaning and gutter guard installation both counting for a "gutter company" query.
What you can do: List every service you offer on its own page or clearly on the services page, using the plain names customers use.
query.scope · also called relevant_services, service_inclusions
Used in 2 of 100 census runs this week: “custom signs, vehicle wraps, exterior signage”
Context context
The circumstances the model reads around the query before judging candidates: who is asking, why and under what conditions, such as a homeowner after a hailstorm looking for a roofer.
What you can do: Describe the situations you handle (storm damage, move-out cleaning, new puppy) on your pages so the model can connect them to your business.
query.context · also called background_info, situational_context
Property Context property_context
Details about the property or thing being serviced that the model notes while scoping, such as a two-storey house, a flat commercial roof or a large dog breed, which narrow which businesses fit.
What you can do: State which property types, sizes and materials you work with, e.g. "flat and pitched roofs, homes and small commercial buildings".
query.context · also called subject_attributes, property_background
Excluded excluded
Items the model marks as outside the request while scoping it, such as excluding fence repair when the query asks for new fence installation.
query.scope · also called not_included, omitted_items · same label written as excluded_items
Used in 1 of 100 census runs this week: “companies with recent complaints, very low review count (<5), outdated websites”
Geographic Context geographic_context
The place the model ties the query to, and how far around it a business can reasonably be, when deciding which candidates count. "Plumber in Aurora" usually rules out shops based across the metro with no Aurora service.
What you can do: List the towns and neighbourhoods you serve on the site and keep them consistent with your Business Profile service areas.
query.geography · also called location_context, geographical_scope · same label written as context_geography
Checked by The places your site names against your Business Profile · Term: service-area
Included included
Items the model marks as inside the request while scoping it, such as counting both emergency and scheduled repairs for "HVAC repair in Boise".
query.scope · also called included_elements, accepted_items · same label written as included_items
Service service
The kind of service the query is about, as the model names it while scoping, such as "residential roof replacement" or "dog grooming". Candidates that do not offer it are dropped.
What you can do: Use the common name of each service in headings and in your Business Profile categories, so your offering is recognisable as that service type.
service.fit · also called service_type, offering · same label written as service_class, service_type
Used in 32 of 100 census runs this week: “moving company”
Included Context included_context
Background facts the model decides to carry into its evaluation, such as the season, the property type or a stated budget, which then shape how businesses are screened.
query.context (was query.scope in the provisional inventory: Background facts carried into evaluation) · also called relevant_context, included_background
Excluded Context excluded_context
Background details the model decides to ignore when scoping the request, such as an unrelated aside in a long question about choosing a pest control company.
query.context (was query.scope in the provisional inventory: Background details deliberately ignored) · also called irrelevant_context, omitted_background
Additional Scope additional_scope
Adjacent needs the model thinks may be relevant beyond the core request, such as stump grinding alongside tree removal, which can widen which businesses it considers.
What you can do: Mention related services you also handle, and link them from the main service page, so the model sees you cover the wider job.
query.scope (was query.context in the provisional inventory: Widens what services are in scope) · also called extended_scope, further_areas
Used in 9 of 100 census runs this week: “Links to review sites.”
Audience audience
Who the model thinks the answer is for, such as a first-time homeowner, a landlord or a small business, which shapes which providers and details it highlights.
What you can do: Say who you serve on the site (homeowners, property managers, businesses) and show examples for each group.
query.context · also called target_audience, intended_users · same label written as audience_type
Audience Context audience_context
What the model assumes about the asker's situation and knowledge, such as a renter who cannot authorise major work, which changes which businesses and caveats suit them.
What you can do: Address common customer situations directly, e.g. a page for landlords or renters explaining how approval and billing work.
query.context · also called audience_background, user_context
Context And Time Horizon context_and_time_horizon
The model's read of when the need arises and how long it matters, such as a burst pipe tonight versus a kitchen remodel next spring, which shifts the weight on availability.
What you can do: Publish your hours, emergency availability and typical lead times for booked work, and keep them the same across listings.
query.context · also called timeframe_context, temporal_scope
Checked by Your hours, day by day, everywhere they are printed
Domain domain
The broad field the model places the query in, such as home services, pet care or legal, which sets the sources and quality bars it applies to candidates.
What you can do: Use Business Profile categories and schema types that match your actual field, so the model files you in the right domain.
query.context · also called field_of_interest, knowledge_area
Domain Context domain_context
Background knowledge about the field that the model brings to the query, such as typical licensing for electricians or seasonal demand for pool services, used to judge candidates.
What you can do: Show the field's usual credentials and norms on your site, e.g. licence numbers for an electrician, so you meet the expectations the model applies.
query.context · also called domain_background, field_context
Excluded Scope excluded_scope
The boundary the model draws around what it will not cover in its answer, such as DIY instructions or businesses outside the named city.
query.scope · also called out_of_scope, irrelevant_scope · same label written as out_of_scope
In Scope in_scope
What the model decides falls inside the request's boundaries, such as which services, places and business types it will consider for "best fence company in Denver".
query.scope · also called within_scope, relevant_items · same label written as included_scope
Included Region included_region
The region the model accepts as in scope for candidates, such as a city plus nearby suburbs, used to keep or drop businesses by where they operate.
What you can do: Name the cities and suburbs you serve on the site and in your Business Profile, and keep those lists matching.
query.geography (was query.scope in the provisional inventory: Defines the accepted service region) · also called relevant_region, geographical_inclusion
Checked by The places your site names against your Business Profile · Term: service-area
Independent Local Mode independent_local_mode
A mode where the model favours independent, locally owned businesses over chains or franchises, often triggered by phrasing like "local" or "family-owned".
What you can do: If you are independent and locally owned, say so on the About page with owner names, founding year and local address.
query.intent (was query.context in the provisional inventory: Implied preference for independent local businesses) · also called local_mode, independent_operation
Planned Policy Mode planned_policy_mode
A mode where the model expects the asker is following a planned process, such as scheduled maintenance or an insurance claim, and favours businesses that work within that process.
What you can do: Explain how you handle planned jobs and insurer or HOA requirements, e.g. a page on working with insurance claims.
query.context · also called policy adherence, strategy alignment
Regional Context regional_context
Regional and cultural factors the model notes, such as local language, building styles or customs, that affect which businesses suit the asker.
What you can do: Mention regional specifics you serve, such as languages spoken or local styles and building codes you work with.
query.geography · also called geographical factors, cultural context
Risk Avoidance Mode risk_avoidance_mode
A mode where the model leans toward the lowest-risk choice, favouring licensed, insured, well-reviewed businesses over cheaper or less proven ones, common for electrical or tree work.
What you can do: Show licence, insurance, warranties and verifiable reviews prominently, and keep them consistent across directories.
query.intent (was query.context in the provisional inventory: Implied preference for lowest-risk choice) · also called caution mode, safety-first approach
Value Mode value_mode
A mode where the model weighs value for money, favouring businesses that offer solid quality at a fair, clearly stated price over the cheapest or the premium option.
What you can do: Publish typical prices or price ranges and what is included, so value can be judged from your own site.
offer.fit · also called importance assessment, worth evaluation
Checked by The claims your site makes: years, prices, refund window