Claygent Navigator clicks, filters and paginates through websites for you.
We’ve been digging into what’s actually possible using this new agentic workflow engine inside Claygent.
So we took the 20 most useful Navigator workflows that were basically impossible before and turned them into a playbook.
01Pricing and Packaging Capture
Steps to build it:
- In Clay, add a Use AI column.
- Select Web Research → Claygent Navigator.
- Use and edit this prompt as per your use case:
Task: Visit the pricing page and extract pricing and packaging. Inputs: - Pricing URL: [PRICING_URL] - Preferred region: United States - Preferred currency: USD Instructions: 1) Load the pricing page. Accept cookie banner if needed. 2) If there is a calculator or toggle, interact with it to reveal prices. a) Capture both Monthly and Annual if shown. b) For seat selectors, set to the minimum allowed. Record that. c) For usage inputs, set to lowest valid values. Record that. 3) For each visible plan, capture: - Plan name - Monthly price amount and currency - Annual price amount and currency, note if billed annually with discount - Billing unit (per user, per seat, per month, etc.) - Minimum commitment or seats - Whether the plan is “contact sales” or “custom” - Whether a free trial exists and its length - Up to five key features listed 4) Only return visible values. Do not guess. 5) Use United States and USD if options exist. 6) Provide results as rows using the Output Fields schema. Include source URL. 7) Stop when all main plans are captured. - Define output fields:
source_urltierprice_type(listed, contact_sales, custom_calculated)monthly_priceannual_pricecurrencybilling_unitmin_seatsusage_inputs_usedannual_discount_notefree_trialfree_trial_length_dayskey_featuresnotes
- Run the agent.
- Open the replay to confirm exactly which clicks were used.
02Security and Compliance Check
Steps to build it:
- In Clay, add a Use AI column.
- Select Web Research → Claygent Navigator.
- Paste this prompt:
Task: Visit the security or trust portal for the company and extract compliance information. Inputs: - Security page URL: [SECURITY_URL] // if you only have a domain, search the site for “security” or “trust” page. Instructions: 1) Load the page. Accept any cookie banners if needed. 2) Scan for visible references to certifications or frameworks (SOC 2, ISO 27001, HIPAA, GDPR, etc.). 3) Capture any details about: - Compliance certifications (list them) - Security features or policies highlighted - Data residency notes (e.g., “all data stored in US or EU”) - Date of latest certification, if shown 4) Do not guess. Only return what is explicitly visible. 5) Provide results as rows using the Output Fields schema. Include the exact source URL. - Define output fields:
source_urlcertifications(comma separated)security_featuresdata_residencylast_certification_datenotes
- Run the agent.
- Open the replay to confirm the certifications and notes were captured from the page.
03Partner Directories
Steps to build it:
- In Clay, add a Use AI column.
- Select Web Research → Claygent Navigator.
- Use and edit this prompt as per your use case:
Task: Visit the partner or integration directory for the company and extract partner listings. Inputs: - Directory URL: [DIRECTORY_URL] Instructions: 1) Load the directory page. Accept any cookie banners if needed. 2) Apply category or filter options (for example: CRM, Marketing Automation, Analytics). 3) Paginate through all pages of results. 4) For each partner or integration, capture: - Partner or integration name - Category or filter applied - Short description, if listed - Link to the partner detail page 5) Do not guess. Only return what is explicitly visible. 6) Provide results as rows using the Output Fields schema. Include the exact source URL. - Define output fields:
source_urlpartner_namecategorydescriptionpartner_link
- Run the agent.
- Open the replay to confirm filters and pagination were applied correctly.
04Customer Proof Points
Steps to build it:
- In Clay, add a Use AI column.
- Select Web Research → Claygent Navigator.
- Use and edit this prompt as per your use case:
Task: Visit the case studies or customers section of the site and extract customer proof points. Inputs: - Case studies URL: [CASE_STUDIES_URL] Instructions: 1) Load the page. Accept any cookie banners if needed. 2) Paginate through all case study or customer listings. 3) For each case study, capture: - Customer name - Industry or sector, if listed - Headline outcome or metric (for example: “200% ROI in 6 months”) - Link to the case study detail page 4) Do not guess. Only return information explicitly shown. 5) Provide results as rows using the Output Fields schema. Include the exact source URL. - Define output fields:
source_urlcustomer_nameindustryheadline_outcomecase_study_link
- Run the agent.
- Open the replay to confirm all case studies were paginated and captured correctly.
05Event Sponsorships
Steps to build it:
- In Clay, add a Use AI column.
- Select Web Research → Claygent Navigator.
- Use and edit this prompt as per your use case:
Task: Visit the conference or event website and extract sponsorship information. Inputs: - Event sponsor or exhibitor page URL: [EVENT_URL] Instructions: 1) Load the sponsor or exhibitor directory. Accept any cookie banners if needed. 2) Apply available filters (for example: sponsor tier such as Gold, Silver, Bronze). 3) Go through all results until the full sponsor list is covered. 4) For each sponsor or exhibitor, capture: - Company name - Sponsorship tier, if labeled - Industry or category, if shown - Link to sponsor or exhibitor detail page 5) Do not guess. Only capture what is explicitly visible. 6) Provide results as rows using the Output Fields schema. Include the exact source URL. - Define output fields:
source_urlsponsor_namesponsorship_tierindustrysponsor_link
- Run the agent.
- Open the replay to confirm filters and pagination captured the full list.
06Vendor and RFP Portals
Steps to build it:
- In Clay, add a Use AI column.
- Select Web Research → Claygent Navigator.
- Use and edit this prompt as per your use case:
Task: Log into the vendor or RFP portal and extract registration or approval status. Inputs: - Portal URL: [PORTAL_URL] - Session cookie: [SESSION_COOKIE] // copy from your authenticated browser session Instructions: 1) Load the vendor or RFP portal using the provided session cookie. 2) Navigate to the registration or submission status section. 3) Apply any relevant filters (for example: by vendor name, submission ID). 4) For each record, capture: - Vendor or company name - Submission or registration ID - Status (for example: approved, pending, rejected) - Date of last update, if shown 5) Do not guess. Only capture what is explicitly visible. 6) Provide results as rows using the Output Fields schema. Include the exact source URL. - Define output fields:
source_urlvendor_namesubmission_idstatuslast_updated
- Run the agent.
- Open the replay to confirm the portal navigation and captured statuses.
07Review Filters
Steps to build it:
- In Clay, add a Use AI column.
- Select Web Research → Claygent Navigator.
- Use and edit this prompt as per your use case:
Task: Visit the software review site and extract filtered review data. Inputs: - Review site URL: [REVIEW_URL] Instructions: 1) Load the review directory page. Accept any cookie banners if needed. 2) Apply filters such as industry, company size, or segment. 3) Paginate through all results under the selected filters. 4) For each review, capture: - Product name - Industry or company size filter applied - Star rating - Review title or headline - Date of the review 5) Do not guess. Only capture information explicitly visible on the site. 6) Provide results as rows using the Output Fields schema. Include the exact source URL. - Define output fields:
source_urlproduct_nameindustry_segmentstar_ratingreview_titlereview_date
- Run the agent.
- Open the replay to confirm filters and pagination were applied correctly.
08Content Gap Sweeps
Steps to build it:
- In Clay, add a Use AI column.
- Select Web Research → Claygent Navigator.
- Use and edit this prompt as per your use case:
Task: Visit the blog or resource library and extract recent content topics. Inputs: - Blog or resources URL: [BLOG_URL] Instructions: 1) Load the blog or resource hub page. Accept any cookie banners if needed. 2) Paginate through at least the first 2–3 pages of posts. 3) For each post, capture: - Post title - Publication date - Category or tag, if shown - Link to the post 4) Do not guess. Only return what is explicitly visible. 5) Provide results as rows using the Output Fields schema. Include the exact source URL. - Define output fields:
source_urlpost_titlepublication_datecategorypost_link
- Run the agent.
- Open the replay to confirm pagination captured the latest posts correctly.
09Media Kits and Brand Assets
Steps to build it:
- In Clay, add a Use AI column.
- Select Web Research → Claygent Navigator.
- Use and edit this prompt as per your use case:
Task: Visit the media kit or brand assets page and extract available files and guidelines. Inputs: - Media kit or press page URL: [MEDIA_KIT_URL] Instructions: 1) Load the media kit or brand assets page. Accept any cookie banners if needed. 2) Scroll and capture each listed asset or guideline. 3) For each asset, capture: - Asset name (for example: logo, brand guidelines PDF) - File type (PNG, SVG, PDF, etc.) - Direct download link, if available - Any stated usage rules or restrictions 4) Do not guess. Only capture what is explicitly shown. 5) Provide results as rows using the Output Fields schema. Include the exact source URL. - Define output fields:
source_urlasset_namefile_typedownload_linkusage_rules
- Run the agent.
- Open the replay to confirm all listed assets and rules were captured.
10Webinars and Podcasts
Steps to build it:
- In Clay, add a Use AI column.
- Select Web Research → Claygent Navigator.
- Use and edit this prompt as per your use case:
Task: Visit the webinars or podcasts section of the site and extract event or episode details. Inputs: - Webinars or podcasts page URL: [EVENTS_URL] Instructions: 1) Load the webinars or podcasts page. Accept any cookie banners if needed. 2) Paginate or scroll through all available listings. 3) For each entry, capture: - Title of the webinar or podcast - Date (event date or publication date) - Speaker or host names, if listed - Link to the registration or episode page 4) Do not guess. Only return information that is explicitly visible. 5) Provide results as rows using the Output Fields schema. Include the exact source URL. - Define output fields:
source_urltitledatespeakersevent_link
- Run the agent.
- Open the replay to confirm pagination or scrolling captured all listings correctly.
11ATS Role Harvesting
Steps to build it:
- In Clay, add a Use AI column.
- Select Web Research → Claygent Navigator.
- Use and edit this prompt as per your use case:
Task: Visit the company’s ATS job board and extract open role details. Inputs: - ATS job board URL: [ATS_URL] Instructions: 1) Load the ATS job board. Accept any cookie banners if needed. 2) Paginate through all available job listings. 3) For each job posting, capture: - Job title - Department or team, if listed - Location or remote tag - Posting date, if visible - Link to the job description 4) Do not guess. Only capture what is explicitly shown on the page. 5) Provide results as rows using the Output Fields schema. Include the exact source URL. - Define output fields:
source_urljob_titledepartmentlocationposting_datejob_link
- Run the agent.
- Open the replay to confirm pagination captured all listings correctly.
12Developer Footprint Mapping
Steps to build it:
- In Clay, add a Use AI column.
- Select Web Research → Claygent Navigator.
- Use and edit this prompt as per your use case:
Task: Visit the product documentation, changelog, or linked repository pages and extract technical footprint signals. Inputs: - Docs, changelog, or repo URL: [TECH_URL] Instructions: 1) Load the documentation, changelog, or repository page. Accept any cookie banners if needed. 2) Scroll or paginate to capture recent entries. 3) For each entry, capture: - Feature or update title - Date of release or update - Technology, SDK, or language referenced (for example: Python SDK, iOS app, React component) - Link to the entry 4) Do not guess. Only capture what is explicitly shown. 5) Provide results as rows using the Output Fields schema. Include the exact source URL. - Define output fields:
source_urlfeature_or_updaterelease_datetechnology_referenceentry_link
- Run the agent.
- Open the replay to confirm it captured recent updates and technical references correctly.
13University Programme Lookups
Steps to build it:
- In Clay, add a Use AI column.
- Select Web Research → Claygent Navigator.
- Use and edit this prompt as per your use case:
Task: Visit the university career or program page and extract internship or co-op program listings. Inputs: - University portal or program URL: [UNIVERSITY_URL] Instructions: 1) Load the internship or program listings page. Accept any cookie banners if needed. 2) Apply filters if available (for example: by program type or department). 3) Paginate or scroll through all listings. 4) For each program, capture: - Program title - Application deadline, if shown - Eligibility criteria (for example: undergraduate, graduate) - Location or department - Link to program details 5) Do not guess. Only capture what is explicitly listed. 6) Provide results as rows using the Output Fields schema. Include the exact source URL. - Define output fields:
source_urlprogram_titleapplication_deadlineeligibilitylocationprogram_link
- Run the agent.
- Open the replay to confirm filters and pagination captured all programmes correctly.
14SEC EDGAR Queries
Steps to build it:
- In Clay, add a Use AI column.
- Select Web Research → Claygent Navigator.
- Use and edit this prompt as per your use case:
Task: Visit the SEC EDGAR search portal and extract company filing details. Inputs: - EDGAR search URL: [EDGAR_URL] - Company name or ticker: [COMPANY_NAME] Instructions: 1) Load the SEC EDGAR search page. Accept any cookie banners if needed. 2) Enter the company name or ticker in the search field. 3) Apply filters for filing type (for example: 10-K, 8-K). 4) Open the most recent filings returned. 5) For each filing, capture: - Filing type - Filing date - Link to the filing document - Headline section or first listed risk factor, if visible 6) Do not guess. Only capture what is explicitly listed on EDGAR. 7) Provide results as rows using the Output Fields schema. Include the exact source URL. - Define output fields:
source_urlfiling_typefiling_datefiling_linkrisk_factor_excerpt
- Run the agent.
- Open the replay to confirm the search, filters, and filing capture were applied correctly.
15FINRA BrokerCheck
Steps to build it:
- In Clay, add a Use AI column.
- Select Web Research → Claygent Navigator.
- Use and edit this prompt as per your use case:
Task: Visit the FINRA BrokerCheck portal and extract registration and disclosure details. Inputs: - BrokerCheck URL: [BROKERCHECK_URL] - Firm or individual name: [ENTITY_NAME] Instructions: 1) Load the FINRA BrokerCheck search page. Accept any cookie banners if needed. 2) Enter the firm or individual name in the search field. 3) Select the correct entity from the results. 4) Capture the following details: - Entity name - Registration or CRD number - Current status (active, inactive, etc.) - Licensing details - Disclosure flags (yes/no) if displayed 5) Do not guess. Only capture what is explicitly shown on the page. 6) Provide results as rows using the Output Fields schema. Include the exact source URL. - Define output fields:
source_urlentity_namecrd_numberstatuslicensing_detailsdisclosures_flag
- Run the agent.
- Open the replay to confirm the search steps and captured data are correct.
16Licence and Permit Verifications
Steps to build it:
- In Clay, add a Use AI column.
- Select Web Research → Claygent Navigator.
- Use and edit this prompt as per your use case:
Task: Visit the state or federal license registry and extract license or permit status. Inputs: - Registry URL: [REGISTRY_URL] - Entity or license holder name: [ENTITY_NAME] Instructions: 1) Load the license or permit registry page. Accept any cookie banners if needed. 2) Enter the entity or license holder name in the search field. 3) Apply filters if available (for example: license type, state). 4) For each record returned, capture: - Entity or license holder name - License or permit number - Status (active, expired, suspended, pending, etc.) - Expiration or renewal date, if shown - License type or category 5) Do not guess. Only capture what is explicitly visible in the registry. 6) Provide results as rows using the Output Fields schema. Include the exact source URL. - Define output fields:
source_urlentity_namelicense_numberstatusexpiration_datelicense_type
- Run the agent.
- Open the replay to confirm search inputs and captured results are correct.
17Sanctions and Watchlist Screens
Steps to build it:
- In Clay, add a Use AI column.
- Select Web Research → Claygent Navigator.
- Use and edit this prompt as per your use case:
Task: Visit the sanctions or watchlist search portal and extract screening results. Inputs: - Watchlist portal URL: [WATCHLIST_URL] - Entity or individual name: [ENTITY_NAME] Instructions: 1) Load the sanctions or watchlist search page. Accept any cookie banners if needed. 2) Enter the entity or individual name in the search field. 3) Submit the search form. 4) For each result, capture: - Entity or individual name returned - List type (for example: OFAC, UN, EU, state-level) - Status or match flag - Additional notes if shown (for example: date of listing, ID number) 5) Do not guess. Only capture what is explicitly listed on the portal. 6) Provide results as rows using the Output Fields schema. Include the exact source URL. - Define output fields:
source_urlentity_namelist_typestatusnotes
- Run the agent.
- Open the replay to confirm the search form inputs and results were captured correctly.
18Changelog and Status Monitors
Steps to build it:
- In Clay, add a Use AI column.
- Select Web Research → Claygent Navigator.
- Use and edit this prompt as per your use case:
Task: Visit the product changelog or status page and extract recent updates. Inputs: - Changelog or status URL: [CHANGELOG_OR_STATUS_URL] Instructions: 1) Load the page. Accept any cookie banners if needed. 2) Scroll or paginate to capture recent entries. 3) For each entry, capture: - Update or incident title - Date and time, if shown - Category or component, if shown - Brief summary text that is visible - Link to the entry detail page 4) Do not guess. Only capture what is explicitly visible on the page. 5) Provide results as rows using the Output Fields schema. Include the exact source URL. - Define output fields:
source_urlentry_titledatetimecomponentsummaryentry_link
- Run the agent.
- Open the replay to confirm the scroll or pagination captured the latest entries correctly.
19Usage Dashboard Snapshots
Steps to build it:
- In Clay, add a Use AI column.
- Select Web Research → Claygent Navigator.
- Use and edit this prompt as per your use case:
Task: Log in to the web dashboard and capture visible metric values for a specified account or date range. Inputs: - Dashboard URL: [DASHBOARD_URL] - Session cookie: [SESSION_COOKIE] // from an authenticated browser session - Account identifier or filter value: [ACCOUNT_FILTER] - Date range, if required: [DATE_RANGE] Instructions: 1) Load the dashboard using the provided session cookie. 2) Apply the account filter and date range as specified. 3) Wait for charts or tables to finish loading. 4) Capture the visible metric values and labels from tables or numeric widgets (for example: active users, sessions, tickets, NPS). 5) Do not guess. Only capture values that are explicitly shown in the UI. 6) Provide results as rows using the Output Fields schema. Include the exact source URL. - Define output fields:
source_urlaccountdate_rangemetric_namemetric_valuenotes
- Run the agent.
- Open the replay to confirm filters were applied and values were captured after the page finished loading.
20Support Policy Comparisons
Steps to build it:
- In Clay, add a Use AI column.
- Select Web Research → Claygent Navigator.
- Use and edit this prompt as per your use case:
Task: Visit the support or plans page and extract support entitlements by plan. Inputs: - Support or plans URL: [SUPPORT_URL] Instructions: 1) Load the page. Accept any cookie banners if needed. 2) Identify plan-specific support details. 3) For each plan, capture: - Plan name - Support channels offered (for example: email, chat, phone) - Stated response time or SLA, if shown - Support hours (for example: business hours, 24/7) - Any noted limitations or premium support notes 4) Do not guess. Only capture what is explicitly listed on the page. 5) Provide results as rows using the Output Fields schema. Include the exact source URL. - Define output fields:
source_urlplan_namesupport_channelsresponse_time_slasupport_hourslimitations_notes
- Run the agent.
- Open the replay to confirm the plan sections and support details were captured accurately.
Questions people ask
What does Claygent do?
Claygent is the AI research agent inside Clay. With Navigator, it can load a page, accept cookie banners, use toggles and dropdowns, apply filters and page through results. It then returns the values you asked for as columns in your table, with the source URL.
