We tested 54 home care keywords across ChatGPT, Perplexity, Google AI Overview, and Claude. Directories appear in AI answers 4x more often than agency websites. Most agency content is structured in ways AI models cannot summarize confidently.
Key Findings
- Directories appear in AI-generated answers 4x more often than individual agency websites across all four platforms tested
- Google AI Overview appears for 67% of home care informational queries — primarily how-to, what-is, and cost-related searches
- Perplexity cites sources explicitly: agencies whose content was cited received referral traffic; agencies whose content was not structured for citation received none
- ChatGPT recommends specific agencies only when prompted for local results — and only in markets where it has sufficient local data
- "How much does home care cost?" is the most AI-visible home care query — appearing in AI Overview and Perplexity answers in nearly every test
- Agency FAQ pages are the most AI-extractable content format — short, specific answers to common questions are cited more frequently than long-form service page copy
The AI Search Shift
AI-generated answers are changing how families research home care. A family member typing "what does non-medical home care include?" into ChatGPT or asking Google and receiving an AI Overview is getting an answer without clicking a result. The implications for home care marketing are direct: if the answer Google or ChatGPT surfaces is authoritative and complete, the family's next step may be a local search for an agency — not further research.
For agencies, this means two things. First, appearing in or being cited by AI answers for informational queries builds brand awareness earlier in the research process. Second, the agencies cited tend to be directories, not individual agencies — which means directories are capturing the first impression with families who are still forming their understanding of what home care is.
What AI Models Are Drawing On
AI models answer home care questions by synthesizing content they have indexed from the web. The sources that appear most frequently in our tests:
- Medicare.gov and related government resources
- AARP and major nonprofit caregiving publications
- Care.com, A Place for Mom, Caring.com — content sections, not listing pages
- Major health system websites (Mayo Clinic, Cleveland Clinic) where home care content exists
- Individual agency content in specific cases where it is well-structured and comprehensive
Individual agency websites appear when they have:
- FAQ pages with specific, direct answers to common questions
- Blog or resource content that answers the question being asked
- Sufficient domain authority for AI models to treat them as credible sources
The content format matters. Long service page copy written primarily for conversion is not well-suited to AI extraction. A page that says "Our compassionate caregivers provide personalized in-home care services tailored to your loved one's unique needs" is not answering a question. A page that says "Non-medical home care includes personal care (bathing, dressing, grooming), companion care, meal preparation, light housekeeping, and transportation" is.
Google AI Overview
AI Overview appears for 67% of home care informational queries in our test set. It does not appear for transactional queries ("home care agency near me," "home care services [city]") — those continue to show the standard local pack and organic results.
The content Google surfaces in AI Overview is drawn primarily from pages that rank well organically. An agency that ranks on page one for an informational query has a reasonable probability of contributing to the AI Overview for that query. An agency on page three does not.
The feedback loop: good organic ranking leads to AI Overview visibility. AI Overview visibility increases brand awareness among families in the research phase. Families who encounter a brand during research are more likely to search for it specifically when they are ready to act — which in turn improves that agency's local pack performance.
Perplexity and Source Citation
Perplexity is the AI search platform most likely to generate referral traffic for cited sources, because it shows source links explicitly in its answers. In our testing, cited sources received measurable referral traffic from Perplexity queries on home care topics.
The agencies and directories cited by Perplexity share a characteristic: their content directly answers the question being asked, in a format Perplexity can extract a specific answer from. Pages that rank for the query but answer it obliquely — or bury the answer in five paragraphs of context — are less likely to be cited.
Structuring content for Perplexity citation is not different from structuring it for featured snippets: lead with the direct answer, be specific, be complete. The same page that wins a Google featured snippet tends to be cited by Perplexity.
Chatbot Responses to Local Queries
ChatGPT and Claude recommend specific agencies when asked for local recommendations, but only in markets where sufficient local data exists and only when the user specifies their location. For most markets, the AI response is a set of criteria for evaluating agencies, not a list of agency names.
This pattern will likely change as AI models ingest more local data. The agencies that will be recommended when these models do name specific agencies are the ones with the strongest signal of authority in their market: high review counts, complete GBP profiles, cited content, mentions across local directories and publications.
Building that signal now positions an agency for AI recommendation when the capability becomes common. Waiting until AI starts recommending local businesses is too late to build the underlying reputation.
What Agency Content Needs to Change
The gap between how most agencies write their web content and how AI models can use it is significant. Most agency websites are written for conversion — emotional language, benefit-focused copy, calls to action. That content is appropriate for visitors who have already decided they want home care and are evaluating providers.
It is not appropriate for families in the research phase who are asking "what is home care?" and "how does home care work?" Those families are being answered by directories and government sites. Agencies that add a resource center, FAQ page, or blog that directly answers informational questions become eligible for the AI-driven research phase — not just the decision phase.
The investment is relatively small: ten to twenty pages of specific, direct answers to common home care questions. The return is visibility at the top of the research funnel, where families are forming their understanding of what home care is and who provides it in their area.
Methodology
We tested 54 home care keywords across ChatGPT (GPT-4o), Perplexity (online mode), Google with AI Overview enabled, and Claude (claude-3-5-sonnet) in 2026. Queries were run from US IP addresses. Source citation data was collected for Perplexity where available. AI Overview presence was assessed against current Google SERP results for each keyword.
If your agency website has no informational content and no structured data, you're invisible to the AI systems that are beginning to answer home care questions for families at scale. Our AI marketing service for home care agencies addresses both — the content that gets cited and the structure that makes it usable by AI.