Can swap in recipient- and company-specific details, but the generated email body still keeps the same overall structure, messaging, and value proposition across different prospects.
What was measured
Personalization quality
How well the generated email reflects recipient-specific and context-specific details.
decisive for this rankingtransformation
How well the email reflects the recipient and context is central to judging AI outreach quality. (3 of 3 judges)
What was given, what came back
Test input: Aditya Nichite prospect lookup · text · group: prospect-search
Input — what we sent
The exact prompt
Aditya Nichite
Direct people/prospect search for Aditya Nichite to test known-prospect retrieval and compare how well tools return validated contact records.
Output — unretouched
Campaign email draft with merge fields and prospect list.
Personalized outreach email draft for Tytus.
Also checked on this input — same tool, 5 other criteria
People Search✓ WorkedCan locate a specific prospect and return validating context such as company affiliation, job title, location, and email, which is sufficient to confirm a match before outreach.People Search✗ FailedCan also fail completely on a fully specified person query, returning 0 matches and no usable prospect record for a known individual.Relevance✗ FailedCan fail to match a known person at all, returning 0 results for a named-person search instead of a usable prospect record.Supporting prospect information✓ WorkedCan return a single prospect record with enough validation context — name, role, location, and company — to sanity-check the match before outreach.Template dependency⚠ StruggledRelies on a largely fixed outreach template, because two generated emails shared the same structure and core messaging with only limited prospect-specific edits.
Provenance
- Observation
- e4058e25-58d6-473e-a052-af67fe4f9d53
- Evidence run
- 9e5960d0-1a72-4f29-9e1e-c7fc96120e61
- Study
- AI Outreach
- Research task
- 86ba3udqq
- Tested at
- Jun 15, 2026
- Source
- first-party
- Evidence state
- verified
- Proof shown
- input + output shown
- Cost / latency
- not captured
- Repeat run
- not captured
- Tester
- not captured
The last three rows are honest blanks, not placeholders — our capture has no field for them yet.
Query this
get_evidence({
tool: "snov-io",
scenario: "prospect-search"
})MCP · mcp.aidemos.com/api/mcp
Free with attribution.
Same input, same check — 0 other tools
measured on Personalization quality
No other tool was measured on this criterion for this input.
This evidence is published in
Real inputs and real outputs, no retouching · every cell queryable via API & MCP · aidemos.com