| Campaign | Status | Spend (USD) | % Budget | Clicks | Impressions | CTR | Avg CPC (USD) | Conversions | CPA (USD) |
|---|---|---|---|---|---|---|---|---|---|
| Generic – AI Agent Industry Review | Enabled | 7,495 | 30.5% | 1,984 | 25,885 | 7.66% | 3.78 | 475 | 16 |
| Generic – Conversational AI | Enabled | 5,858 | 23.8% | 1,762 | 32,710 | 5.39% | 3.32 | 291 | 20 |
| Generic – AI Agent Function | Enabled | 5,542 | 22.6% | 612 | 15,888 | 3.85% | 9.06 | 98 | 57 |
| Competitors | Paused | 3,060 | 12.5% | 240 | 33,627 | 0.71% | 12.75 | 37 | 83 |
| Brand | Paused | 2,601 | 10.6% | 585 | 5,258 | 11.13% | 4.45 | 207 | 13 |
The Brand campaign has the highest CTR (11.1%) and lowest CPA (USD 13) in the account — nearly 4.5× more efficient than the AI Agent Function campaign (CPA USD 57). Despite this, it is currently paused. Conversely, the Competitors campaign is also paused but still accumulated USD 3,060 in spend with a CTR of just 0.71%, meaning 99.3% of people who saw those ads did not click.
The short answer: yes, the keywords are related to AI and chatbots — but the majority are attracting the wrong audience. The campaigns are structured around broad, generic AI search terms that pull in a high volume of consumer-intent traffic (people looking for free ChatGPT alternatives, personal AI chat tools, or comparing AI chatbot brands) rather than business owners or decision-makers who would actually buy Astra.
| # | Keyword | Campaign | Intent Grade | Spend (USD) | Clicks | CTR | Conv | CPA (USD) |
|---|---|---|---|---|---|---|---|---|
| 1 | "best conversational ai" | Industry Review | Mixed | 5,693 | 1,506 | 7.92% | 370 | 15 |
| 2 | "conversational artificial intelligence chatbot" | Conversational AI | Mixed | 5,338 | 1,643 | 5.48% | 271 | 20 |
| 3 | "astra ai" | Brand | ✓ Brand | 2,511 | 561 | 10.77% | 196 | 13 |
| 4 | "chatbot manychat" | Competitors | ⚠ Competitor | 1,720 | 141 | 0.52% | 18 | 96 |
| 5 | "best ai chatbot" | Industry Review | Mixed | 1,533 | 428 | 6.92% | 100 | 15 |
| 6 | "chat bot for business" | AI Agent Function | B2B | 1,109 | 137 | 3.37% | 25 | 44 |
| 7 | "ai tool for sales" | AI Agent Function | B2B | 1,030 | 90 | 4.05% | 10 | 103 |
| 8 | "ai voice agent platform" | AI Agent Function | B2B | 962 | 100 | 4.16% | 20 | 48 |
| 9 | "ai chat bot for website" | AI Agent Function | B2B | 782 | 100 | 3.78% | 16 | 49 |
| 10 | "ai voice agent" | AI Agent Function | B2B | 607 | 78 | 4.57% | 13 | 47 |
* Keywords #1 and #2 alone account for USD 11,031 — 45% of total US Search budget. Both are phrase match keywords broad enough to attract significant non-business traffic.
Keywords are what we bid on. Search terms are the actual queries people typed before seeing our ad. The gap between them reveals the true audience we are reaching. Analysis of the top 100 search terms by spend shows a clear pattern: a large portion of paid traffic comes from people who are not potential Astra buyers.
Based on top 100 search terms. Remaining spend not in top 100 follows a similar distribution.
The vast majority of search traffic is driven by two broad phrase-match keywords — "best conversational ai" and "conversational artificial intelligence chatbot" — which are generic enough to match searches from consumers looking for personal AI chat tools, students doing research, and people comparing free ChatGPT alternatives. None of these searchers are Astra's intended buyer (a business owner or sales/ops manager seeking an AI agent for lead qualification or customer support).
The Competitors campaign (paused) was designed to intercept searches for tools like ManyChat, Tidio, and Zendesk. However, because the generic phrase-match keywords are so broad, they are already triggering on competitor-adjacent searches like "apps like chat gpt", "chat gpt alternative", and "similar to chatgpt" — but serving general ads not designed for that audience. This is the worst of both worlds: competitor-intent traffic being served the wrong creative.
Every ad group has Headline 1 pinned to "Astra by Wati AI Agent" — an unknown brand name to the US market. None of the running ads mention Astra's most compelling and differentiated features: Voice AI with voice cloning, WhatsApp deployment, or the no-code natural language builder. The copy reads generically ("Build no-code AI agents for sales, support, and customer engagement") — indistinguishable from dozens of competing products in the same search results.
All active US Search campaigns run Responsive Search Ads (RSAs). Below is the actual copy pulled from the account, graded across four dimensions, with suggested rewrites highlighting Astra's genuine differentiators.
Voice cloning for business AI agents is a genuinely rare capability. ElevenLabs does voice cloning — but it is a voice tool, not an AI agent platform. Bland.ai, Vapi, and Retell AI do voice agents — but none offer voice cloning. Astra is currently the only product that does both: clone your voice AND deploy it as a full business agent across calls, web, and WhatsApp. None of the current ads make this claim. Owning this positioning in ad copy would create immediate differentiation from every competitor in the US search results, and likely improve CTR and Quality Score significantly by attracting higher-intent searchers.