RESEARCH · IMPLEMENTATION · MEASUREMENT
GEO services for US businesses
Explore the scope, delivery process and commercial terms for working with Auden AI and Prime Team Agency.
FROM RESEARCH TO ACTION
Know what’s missing.
Make it useful.
Generative engine optimization, or GEO, helps you work on how your business appears in AI answers. Auden helps US business owners and marketing teams understand what AI says and what their website needs.
Start with real questions.
Find the questions your customers ask. Save the AI answers and check their source links. See what is clear and what is missing.
Make your pages more useful.
Explain what you do, answer common questions and back up your claims. Fix problems that stop the right services from reading your pages.
Check what changed.
Ask the same questions again. Record the date and AI service. Track mentions, website visits and customer inquiries separately.
Agree the work, who does what, timing and fees in writing before starting. Plans start at USD $3,900 per month with a six-month minimum. Rankings, citations, traffic and revenue are not guaranteed.
WHO WE HELP
Your business knowledge.
A clear plan for the work.
Auden AI is a division of Prime Team Agency. We provide GEO research and implementation for US businesses and marketing teams that can supply accurate product or service information, approve public evidence and assign an approver for website changes.
Good fit: you want to understand real buyer questions, improve the pages that answer them and review dated evidence of the work. Outside this scope: guaranteed recommendations, paid placement inside AI answers, fabricated reviews, unapproved client claims or a promise to rank first.
Onboarding and delivery
Phase 1: establish the baseline
Agree the domains, CMS, audiences, markets, competitors and priority pages. Define the KPIs and baseline before changes. Record approved business facts and claim restrictions, then name the approver, developer and publisher.
Review the measurement and publishing access the scope needs: GA4, Search Console, Tag Manager, Clarity, CMS and server logs. Document the dates, timezone, sampling and gaps in each source. Crawler logs require appropriate plugin or server access and do not show what an AI model recommends.
Approve, stage and verify
Turn the audit into a prioritized queue with an owner, affected page and acceptance check for each change. Stage changes before publishing, retain a rollback and check the released pages. Weekly refreshes, an evidence board and the live dashboard support review of completed work and the next priorities.
Phase 2: deepen the evidence
Starting in month three, prioritize canonical product and service pages. Collect who, what, when, where, why and how facts, customer questions and resolutions, approved redacted feedback, verified launch and use-case history, and dated evidence for claims. Map buyer prompts to the pages that answer them. Keep the evidence owner and publication permission attached to each claim.
Strategy and prompt research, visibility and opportunity reviews, content audits, competitor intelligence and optimization of existing pages form the shared program. The agreed tier sets prompt counts, competitors, engines, new pages, technical scope, reporting cadence and dashboard access.
Before work starts, the engagement brief identifies delivery owners, milestone dates, review rounds and measurement access. Availability, onboarding lead time and the start date are confirmed in the proposal.
Review pricing and contract terms · Compare PTA and Profound · Read the measurement guide
Start with a scope discussion.
Bring your website, target market, priority buyer questions, an approver and evidence you can publish. Discuss your project with PTA. Use the quote checklist to confirm the full scope, extras, ownership and handoff.
Retest and handoff
The analyst repeats the same question set and records engine, model or displayed mode, date, location settings, citations and sample count. The handoff separates completed changes, unresolved items and observed answer differences. Mentions, visits, inquiries and revenue remain separate measures.
Reproducible validation: retain the original questions and before/after records; repeat under comparable conditions; keep negative results; check that cited pages support the answer. A successful page check does not prove indexing or a recommendation. Our measurement guide explains the limits, and the public checker source makes the narrower page-evidence check inspectable.
How we repeat an AI-answer check
Before collecting a baseline, the brief fixes the exact questions, selected engines, run count per question and engine, and baseline and retest windows. Candidate engines include ChatGPT, Claude, Gemini, Perplexity, Grok and DeepSeek; the agreed scope identifies which are included.
Use a fresh conversation for each run. Keep the question unchanged, and record the displayed model or mode, search setting, account and personalization conditions, language, location settings and UTC time. Test unaided discovery separately from tests supplying your website URL.
Repeat the same number of runs after the documented changes, using comparable conditions. Preserve every answer, cited URL, unsuccessful run and negative result. For each engine separately, report completed and attempted runs, brand mentions, explicit recommendations and citations to the business’s domain. Open cited pages to check support. Record unavailable settings and changed conditions; do not silently combine unlike runs.
These observations describe the collected sample, not all users’ answers or proof that the website changes caused a difference. Download the blank observation and change-log template (CSV). It is a template, not a completed test or client result. Record non-identifying session conditions; do not put credentials, personal information or confidential customer material in the log.
For a worked example of checking sources, see our walkthrough of one ChatGPT, Claude and Grok answer from 22 October 2026. It is a single small sample, not this repeat protocol, a client result or evidence of ranking change.
For an agreed multi-turn buyer scenario, preserve the ordered questions and the answer, cited sources and evaluation at every turn. Keep those conversation tests separate from fresh, single-question checks.
PRICING & TERMS
From USD $3,900 per month.
Minimum contract: six months. Base service fees therefore start at USD $23,400 over the minimum term, before approved extras or applicable payment fees. Pricing confirmed 19 October 2026. Your proposal confirms the tier, scope, total fees and start date.
What the entry scope covers
The entry package covers 100 tracked prompts, three competitors, Google AI Overviews and AI Mode, ChatGPT and Gemini, basic technical work, two new answer pages per month, a monthly scorecard and one dashboard login. It includes the shared research, weekly refresh, evidence board and existing-page work described in our delivery process. Higher-volume programs are scoped separately.
How the engagement is agreed
- Six-month renewal periods. Written termination notice is required at least 30 days before the end of the applicable term; it does not shorten the minimum commitment.
- The first invoice is due on receipt; subsequent invoices are Net 15. Card payments carry a 3% fee.
- Approved third-party media costs are additional. The standard agreement provides a 5% annual direct-fee increase from the first anniversary.
- The signed agreement sets the final deliverables, access, data and intellectual-property terms. Confirm any exceptions in writing before work starts.
Commercial process follows the standard agreement dated 9 October 2026, with the starting price updated above. We do not guarantee rankings, citations, traffic or revenue. Request a scope-specific proposal.
PTA VS PROFOUND
Compare the workflow.
PTA configures and runs the agreed research and implementation program. Profound provides a configurable AI marketing platform. Compare the exact workflow, service responsibilities and proposal before choosing. Reviewed 19 October 2026.
| Capability / proposed distinction | Profound | PTA program |
|---|---|---|
| Visibility, competitor and page tracking | ✓ Published | ✓ Included |
| Multi-turn buyer tests with saved checks at each turn | Confirm exact workflow | ✓ Scoped tests |
| Page-type schema coverage with a prioritized fix queue | Confirm exact workflow | ✓ Scoped technical work |
| Answer-change audit showing brands and sources | Confirm exact workflow | ✓ Evidence review |
Reading the table: a tick identifies a published platform capability or an agreed PTA service capability. “Confirm exact workflow” means the reviewed public material does not establish that precise combination. It does not mean the capability is unavailable. PTA test counts, engines, implementation responsibilities and reporting are set in the proposal.
Profound’s published features include visibility, competitor and citation analysis, technical insights and content generation. Its current product description also describes creating and optimizing pages. The distinctions above concern the specified workflow and who operates it. This table does not establish comparative outcomes or a like-for-like total cost.
PTA service pricing starts at USD $3,900 per month with a six-month minimum. Obtain a current Profound quote and compare the same scope, implementation ownership, tools, total cost and limitations.
Comparing GEO agencies?
Use our practical GEO agency scorecard to compare evidence, responsibilities, costs and reporting. Apply the same checks to our proposal.