Help your customers find relevant products and content.
Combine semantic search and exact matching to search your catalog and content. Anteater, our own technology, lets you configure ranking rules and analyze zero-result searches, clicks and abandonment.
Finding the right results takes query understanding, ranking and measurement.
- 01
Different ways to search
People may search for the same product or content by name, a synonym or a description. Search needs to connect these different queries to the information available.
- 02
Relevance and business priorities
Inventory, seasonality, margin and private labels can influence result order. Those priorities need to be evaluated alongside what the person is looking for.
- 03
Visibility into the search experience
Zero-result queries, clicks and abandonment help identify search problems. Reviewing them alongside result quality helps the team decide what to improve.
Meaning, exact matching and business rules in one search.
We connect the sources behind your catalog and content. The engine searches by meaning and exact terms, combines the results and applies the agreed ranking rules. We evaluate that combination using your real queries and usage metrics.
- 1
Connect catalog and content
Integrates products, attributes, inventory, prices and content from the sources defined for the project.
- 2
Interpret the query
Processes terms, synonyms and natural-language descriptions to relate them to indexed information. Performance is evaluated using representative queries, including spelling errors.
- 3
Combine meaning and exact matching
Runs semantic and exact search and merges the results into one list.
- 4
Apply ranking rules
Configures business priorities such as inventory, seasonality, margin or private labels, and evaluates their effect on result relevance.
- 5
Measure to improve
Records searches, clicks and abandonment. The dashboard helps the team review zero-result queries and identify changes to the catalog, content and rules.
Our technology · Anteater
Anteater is Amarello's search and recommendation technology. It combines semantic search, exact matching, result fusion and usage metrics. Intelligent recommendations can be added as a separate solution on the same platform.
Integration, ranking control and measurement using your queries.
- Integration with your sources
- Connectors to catalogs, PIM, CRM and e-commerce platforms, with a search API for your site or interface.
- Relevance and measurement
- Testing with real queries and review of results, zero-result searches, clicks and abandonment against the agreed criteria.
- Ranking rules
- Business rules your team can configure to set priorities and evaluate their effect on results.
- Operations and support
- Operations under agreed service levels, with defined owners for the catalog, content and rules.
- Continuity and handover
- Exportable indexes, rules and dashboards, with documentation and handover terms defined in the contract.
Anteater in live operations.
combines meaning and exact matching in one engine
databases, PDFs, websites and CRM as information sources for the index
Amarello implementations across industries since 2008
Banking · Content search
Two of the largest banks in Mexico
Anteater in production on the digital channels of two of the country's largest banks, with semantic and exact search connected to their content and business rules.
Retail · Search and recommendations
Supermarket chain in Mexico
Search across the online store's full catalog, with recommendations from the same engine and business rules configurable by the chain. Anteater also ran on the marketplace of a retail group.
Companies Amarello has worked with
Banking · Retail · Consumer goods
We move phase by phase, with verified results.
We start with a catalog or content collection and representative queries. We agree on how to evaluate relevance, zero-result searches and engagement with results before expanding traffic. The timeline depends on the scope, sources and required integrations.
- 1
Sources and search criteria
- What we do
- We define the catalog or content, channel, rules and baseline metrics. We agree on what makes a result relevant to a query and how to evaluate the experience.
- Your team's role
- Provide access to sources, available query history, current analytics and catalog or content owners.
- Criteria to proceed
- Sources, scope, metrics and acceptance criteria agreed.
- 2
Indexing and testing
- What we do
- We index the sources and test real queries, including those with no results and those using synonyms or descriptions. We review relevance and the effect of ranking rules.
- Your team's role
- Provide representative queries and business rules, and validate the expected results.
- Criteria to proceed
- Results evaluated against the agreed criteria and required adjustments identified before the pilot.
- 3
Pilot with real traffic
- What we do
- We enable a share of traffic and measure searches, clicks, abandonment and response times. We tune rules with your team and evaluate conversion from search when measurement is available.
- Your team's role
- Involve business and technical owners to review results, rules and integration performance.
- Criteria to proceed
- Result quality, integration and performance validated with pilot traffic.
- 4
Production and operations
- What we do
- We expand traffic according to the agreed plan, document the configuration and activate operations under defined service levels. We establish metric reviews and updates to sources and rules.
- Your team's role
- Assign owners for the catalog, content and ranking priorities.
- Criteria to proceed
- Operations active, sources updated under the agreed approach, owners assigned and documentation delivered.
Search engine, integrations, data and deployment.
Integration
Connectors to catalogs, PIM, CRM and e-commerce platforms, with a search API for the site's interface. During the technical assessment, we define the sources, connectors and required adaptations. The connection to recommendations is configured if that solution is included in scope.
Engine
A hybrid index that combines vectors for semantic search and text for exact matching. Results are merged and ranked using the configured business rules. We evaluate the combination with representative queries for the catalog and content.
Data
Databases, PDFs, websites and CRM can provide information for the index, with incremental updates. The project defines which sources are integrated, which fields are used and how available content, prices and inventory are updated.
Measurement
Search, click and abandonment events are recorded, with a zero-result query dashboard for the responsible team. The project's measurement plan defines events, relevance criteria and, where applicable, conversion tracking from search.
Deployment
Google Cloud Platform (GCP) is our primary runtime environment. Availability on Amazon Web Services (AWS) or private infrastructure is confirmed during the technical assessment, based on the environments validated for the solution.
Data export and handover
The contract defines exports of indexes, rules and dashboards, the documentation to be delivered and the terms for handover to the client's team.
Questions before getting started.
Does it work with our e-commerce platform?
Integration uses APIs and connectors to the catalog and the site's interface. During the technical assessment, we review the available interfaces and the adaptations needed to work with your current platform.
How are products and content in the search engine updated?
The index supports incremental updates from integrated sources. Together, we define which information is used and how changes are received. Update frequency and scope are agreed based on the sources and integration; real-time synchronization is not assumed.
How fast does it respond?
Response times are evaluated using your catalog or content, configured rules and pilot traffic. We agree on performance criteria and verify them before expanding usage.
Can we control which products appear first?
Your team can configure business rules related to inventory, seasonality, margin or private labels. We test how those priorities affect result order and relevance to the user's queries.
How do you evaluate search quality?
We review results using representative queries and measure zero-result searches, clicks and abandonment. When analytics allow, we evaluate conversion from search. The goal is to show relevant results; fewer zero-result queries alone do not demonstrate a better experience.
Does it include recommendations?
Intelligent recommendations are an additional solution on the same platform. They can share the catalog, signals and dashboard when integrated into the project. You can start with search, recommendations or both, with a defined scope for each solution.
What is Anteater?
Anteater is Amarello's own search and recommendation technology. It combines semantic search, exact matching, result fusion, ranking rules and metrics. It also supports vector-based recommendations when that solution is added to the project.
How do you support continuity without dependence on Amarello?
The contract defines exports of indexes, rules and dashboards, documentation, rights to components and handover to the client's team.
Related solutions
- Intelligent recommendationsRecommendations based on popularity, trends and affinity, connected to your catalog, events and business rules. An additional solution powered by Anteater.Explore the solution
- AI creative generationGenerate campaign variants for different channels and formats using your brand guidelines, with a review of each asset.Explore the solution
Let's review what your customers find when they search.
We start with your catalog, content and real queries. We agree on how to evaluate relevance, zero-result searches and engagement with results, then test with a limited share of traffic before expanding the scope.