Blog
A clearer view of the age of AI search
Explore how AI answers shape discovery. Kept brings together practical guides, product notes and thoughtful analysis to help your team understand its visibility and decide what to do next.
Designing Shopping Visibility Checks Across Regions
Build a practical comparison of shopping answers across US and European markets by controlling prompts, locale, product availability, and the evidence captured during each observation.
Two Illustrative Team Setups for Testing AI Search Ideas
Explore two illustrative team setups for testing AI search improvements, choosing a manageable hypothesis, assigning ownership, and deciding what evidence would justify the next investment.
Build a Visibility Dashboard Your Team Can Use
Plan shared dashboard views around real decisions, selecting relevant prompt groups, comparison periods, and annotations so different stakeholders can interpret the same visibility evidence consistently.
Publishing Statistics That Deserve to Be Cited
Learn to publish useful statistics with traceable methods, clear limitations, and accessible source material, then assess citations without confusing repeated claims with independently verified evidence.
Why Prompt Tracking Needs a Stable Baseline
Understand how prompt selection, sampling, and measurement settings can change visibility scores, and establish a consistent baseline before interpreting apparent gains or losses in performance.
Reading the Story AI Tells About Your Brand
Create a repeatable process for reviewing brand descriptions in AI answers, separating sentiment from factual accuracy, and turning recurring themes into questions for deeper investigation.
Following an AI Referral From Visit to Outcome
Map the journey from an AI assistant referral to a meaningful website outcome, documenting attribution limits and choosing metrics that connect visits with business questions.
How to Investigate an AI Search Index
Learn to distinguish crawling, indexing, retrieval, and answer generation when investigating AI search, using observable evidence and explicit uncertainty instead of assuming a single search provider.
Testing How a Company Name Shapes AI Answers
Design a controlled test of company naming effects in AI answers, keeping background information consistent and checking whether wording changes alter descriptions, associations, or recommendations.
Making Content Easier to Evaluate During Reranking
Explore the role reranking can play between retrieval and answer generation, then review passage clarity, context, and supporting evidence as practical content evaluation criteria.
Where GEO Extends Your Existing SEO Work
Compare the questions SEO and GEO measurement can answer, identify shared content foundations, and decide which additional observations help explain your presence in AI responses.
A Working Method for Monitoring Ads in AI Answers
Set up an observation log for ads encountered alongside AI answers, recording prompt context, placement, and disclosures while keeping paid exposure separate from organic visibility.
What Server Logs Can Tell Your AI Search Team
Use server logs to investigate crawler access, request patterns, and technical obstacles, while recognizing that a recorded visit alone cannot establish indexing, citations, or downstream visibility.
Auditing Site Access for Deep Research Workflows
Plan a site access audit for research-oriented assistants by comparing server requests, rendered content, and accessible documents, then documenting what the evidence can actually support.
Investigating Translated Community Sources in AI Answers
Examine how translated community pages appear in AI citations, distinguishing original discussions from translated versions and checking language, provenance, and relevance before drawing visibility conclusions.
Preparing Product Information for AI Shopping Discovery
Review the product information shoppers need when asking AI assistants for recommendations, and build a maintenance checklist for accurate specifications, availability, pricing context, and trustworthy supporting pages.
Building a Useful AI Shopping Measurement Plan
Choose a focused set of shopping prompts and product comparisons, then organize observations around recommendation presence, cited information, and changes that your ecommerce team can investigate.
Planning MCP Workflows for Search Analysis
Outline how connected tools could support search analysis, with explicit data permissions, reliable inputs, and review checkpoints before an assistant turns observations into recommendations or actions.
Giving AI Overviews a Place in Your Tracking Plan
Build a sampling plan for observing AI Overviews alongside other search experiences, recording appearance conditions and response differences without treating one prompt sample as universal coverage.
An Illustrative Fintech Agency Builds a GEO Service
Follow an illustrative agency team as it defines a fintech visibility service, organizes client questions, and separates repeatable reporting from experiments that still need validation.
Studying List Position Without Confusing Cause and Correlation
Design an analysis of list position and brand mentions in AI answers, accounting for source selection, repeated observations, and alternative explanations before interpreting any apparent relationship.
Reading Query Fanouts as Clues to Search Intent
Learn to group observed query expansions by intent, compare them across repeated prompts, and use those patterns to frame content questions without assuming complete access to internal reasoning.
Making Visibility Data Useful in Connected Workflows
Plan how structured visibility data could move into connected assistant workflows, defining useful fields, freshness expectations, permissions, and review steps before relying on generated analysis.
Evaluating a Platform Referral Program Before You Join
Assess platform referral programs through eligibility rules, reward terms, disclosure expectations, and audience fit, so your recommendation remains useful even when a commercial incentive exists.
Mapping Cited Domains for Your Own Search Questions
Create a domain citation map from a defined prompt set, distinguishing citation frequency from relevance and identifying which source relationships deserve closer editorial or technical review.
An Ecommerce Checklist for Investigating AI Product Rankings
Work through a practical product visibility review, checking catalog consistency, supporting content, comparison prompts, and measurement limits before choosing which improvements your ecommerce team should test.
Connecting AI Visibility Metrics to Business Questions
Choose metrics that distinguish answer presence, website visits, and business outcomes, then document attribution assumptions so teams can discuss progress without treating every movement as revenue impact.
Evaluating the Evidence Behind Promotional Comparison Lists
Review promotional comparison lists for transparent criteria, conflicts of interest, and verifiable claims, then investigate how those pages appear in AI citations using a defined sample.
Comparing AI Visibility Plans Beyond the Monthly Price
Compare AI visibility platform plans by measurement scope, usage limits, collaboration needs, and total operating effort, using a checklist that makes pricing tradeoffs easier to discuss.
An Illustrative Agency Organizes Visibility Across Clients
Explore an illustrative agency workflow for managing visibility across a startup portfolio, balancing shared reporting standards with client-specific prompts, priorities, and evidence review before recommending changes.
Building a Citation Benchmark You Can Interpret
Define what counts as a citation, select a relevant comparison group, and record sampling choices before using citation rates to evaluate content or set team expectations.
Reviewing AI Content for Risks That Matter
Build an editorial review process for AI-assisted content that checks factual support, originality, reader usefulness, and accountability before publication, with clear ownership for corrections and updates.
Testing Search Language Across Multilingual AI Prompts
Design a multilingual prompt comparison that records answer language, observable search queries, and cited sources, helping your team investigate language mismatches without assuming a universal preference.
Tracking Query Expansion Changes Across Countries and Time
Plan a repeatable study of query expansion length and structure across countries, keeping prompt cohorts and collection methods consistent so changes can be interpreted with appropriate caution.
Turning Content Gaps Into Reviewable Next Steps
Translate a content gap into a specific proposed change, linking each action to evidence, an owner, and a follow-up check before adding it to your work queue.
Finding Brand Mention Gaps Worth Investigating
Compare brand mentions across a focused prompt set, identify relevant omissions, and investigate context before deciding whether a gap suggests a content opportunity or a measurement issue.
Tracing Source Gaps Behind AI Search Answers
Compare cited sources across related AI answers to find missing or underrepresented material, then prioritize source gaps according to relevance, accessibility, and the evidence your team can improve.
A Repeatable Method for Reviewing Brand Sentiment
Develop a consistent rubric for reviewing brand sentiment in AI answers, separating tone, factual claims, and prompt context before comparing patterns across models or reporting periods.
Choosing Tracking Prompts That Match Real Decisions
Build a tracking prompt set around audience needs and buying questions, balancing topic coverage with manageable scope and documenting changes that could affect later visibility comparisons.
Questions Ecommerce Teams Should Ask About Commerce Protocols
Use a structured checklist to evaluate Universal Commerce Protocol proposals, examining interoperability goals, merchant responsibilities, and implementation evidence before deciding whether further technical investigation is worthwhile.
Setting AI Search Priorities for a Marketing Team
Create a practical AI search roadmap by connecting audience questions, content quality, technical access, and measurement capacity, then choosing a small set of priorities your team can sustain.
An Illustrative B2B Team Connects AI Visits to Demo Interest
Follow an illustrative B2B team as it maps AI referrals to demo interest, checks attribution gaps, and chooses experiments without presenting hypothetical outcomes as customer results.
Looking Beyond Funding Headlines When Choosing an AI Platform
Evaluate an AI platform company through product fit, service continuity, data portability, and transparent commitments, treating funding announcements as one input rather than proof of long-term suitability.
Defining the Denominator in AI Search Market Share
Examine how visits, searches, users, and assistant interactions produce different market-share estimates, then choose a definition that matches your planning question and clearly states its limitations.
An Illustrative Data Platform Team Reviews Its AI Visibility
Explore an illustrative technical marketing team reviewing visibility for a data platform, aligning specialist prompts, documentation gaps, and source evidence before prioritizing a focused content experiment.
A Practical Starting Point for Generative Engine Optimization
Build a foundation for GEO by mapping audience questions, reviewing answer visibility, improving useful source material, and establishing a measurement routine that supports learning rather than guaranteed rankings.
Testing the Usefulness of llms.txt and Markdown Pages
Evaluate proposed llms.txt files and Markdown alternatives against a clear hypothesis, checking accessibility, maintenance cost, and observable use before treating either format as an AI visibility solution.
Planning for AI Search While Keeping Strong SEO Foundations
Consider how changing search experiences affect content planning, technical priorities, and measurement, then identify durable practices your team can maintain while testing new approaches with explicit evidence.
Turning a Source List Into an Editorial Work Plan
Organize cited sources by relevance, ownership, and content role, then turn your review into a focused work plan with evidence checks and realistic options for improving source material.
See your brand through the lens of AI
Bring your questions. Explore the answers with Kept.