Zolve × Nbyula · method comparison

Same leads, two methods:
the team vs the AI agent

We took the exact 415 loan leads assigned to John Stanley and Swatika Reddy and asked a simple question — on these identical people, how does the human team's calling effort compare to Nbyula's loan AI voice agent, which dialled the same numbers?

Snapshot 6 Aug 2026 Leads 415 (the two advisors' full book) Source Nbyula CRM call history + AI voice-call log Same duration basis both sides measured on connected-call length

The finding: it's an effort gap, not a skill gap

When a human dials, they connect at 38% per attempt — the AI connects at only 15%. The people have the skill. But across the same 415 leads the team made 1,018 dials and the AI made 4,288 — the AI simply never stops, never skips a lead, and never decides a number "isn't worth another try." That tireless effort is why the AI reached 78% of the book to the team's 52%, and held 166 five-minute loan conversations to the team's 50 — on the very same leads. The gap isn't ability. It's follow-through.

2.5 → 10.3

dials per lead — team vs AI

52% → 78%

of leads actually reached

50 → 166

real 5-min+ conversations

14.6 → 41.3h

total talk hours delivered

103

leads the team never reached that the AI held a 5-min+ call with

Head to head, on the same book

Every metric below is the identical set of leads, scored the same way for both. Blue = the human team, green = the AI agent. The ×badge is how many times more the AI delivered.

Human advisorAI agent

All 415 leads

both advisors combined vs AI

Leads dialled at least once

Team

396 · 95%

408 · 98%

Total dials made4.2×

Team

1,018

4,288

Leads actually reached1.5×

Team

214 · 52%

323 · 78%

Leads with a 5-min+ conversation3.3×

Team

50 · 12%

166 · 40%

Total talk time (hours)2.8×

Team

14.6 h

41.3 h

John Stanley’s book

204 leads — the harder-working advisor

Leads dialled at least once

Team

194 · 95%

202 · 99%

Total dials made3.2×

Team

727

2,300

Leads actually reached1.3×

Team

122 · 60%

159 · 78%

Leads with a 5-min+ conversation2.0×

Team

40 · 20%

80 · 39%

Total talk time (hours)1.9×

Team

10.5 h

20.1 h

Swatika Reddy’s book

211 leads — where the gap is widest

Leads dialled at least once

Team

202 · 96%

206 · 98%

Total dials made6.8×

Team

291

1,988

Leads actually reached1.8×

Team

92 · 44%

164 · 78%

Leads with a 5-min+ conversation8.6×

Team

10 · 5%

86 · 41%

Total talk time (hours)5.1×

Team

4.1 h

21.1 h

103 qualified leads the team left on the table

These are leads the human advisor never reached (or caught for under a minute), where the AI went on to hold a genuine 5-minute-plus loan conversation — many of them tagged Doable or Borderline. This is revenue that would simply have been lost to no-follow-up.

Lead	Advisor	Human effort	AI effort	AI conversation length	Loan verdict
Priyanshu Choudhary	John	3 dials · 2 reached	2 dials · 2 reached	26:42 min	Borderline
Rajkumar	Swatika	1 dials · 0 reached	4 dials · 1 reached	21:40 min	Borderline
Meeneswararao konidena	John	1 dials · 0 reached	7 dials · 1 reached	20:22 min	Borderline
Mann R Mhamunkar	John	1 dials · 0 reached	3 dials · 2 reached	19:53 min	Borderline
Suzain Aijaz Mistri	Swatika	1 dials · 1 reached	2 dials · 2 reached	19:31 min	Borderline
Sheron Christian	Swatika	1 dials · 1 reached	2 dials · 2 reached	19:18 min	Borderline
Aditya Swain	John	1 dials · 1 reached	2 dials · 2 reached	18:49 min	Borderline
sarfaraz shaikhjiwala	John	8 dials · 0 reached	6 dials · 1 reached	18:06 min	Borderline
Rizvanhusen Shaikh	John	0 dials · 0 reached	7 dials · 1 reached	18:06 min	Doable / Fast-Track
Rohit Jamwal	John	4 dials · 1 reached	5 dials · 2 reached	17:35 min	Borderline
Abhishek	Swatika	1 dials · 0 reached	4 dials · 2 reached	17:15 min	Borderline
Shubham manij Sadafule	Swatika	1 dials · 0 reached	2 dials · 2 reached	17:03 min	Doable / Fast-Track
Alfiya L	Swatika	2 dials · 0 reached	2 dials · 1 reached	16:46 min	Borderline
Gourav Saini	Swatika	1 dials · 1 reached	2 dials · 2 reached	16:36 min	Borderline
Nallamandu bojanna	Swatika	1 dials · 0 reached	2 dials · 1 reached	16:31 min	Borderline
Chintabhathini Sasi kumar	John	5 dials · 0 reached	23 dials · 1 reached	16:18 min	Borderline
Khushi boriwal	John	2 dials · 0 reached	5 dials · 2 reached	15:51 min	Not Doable
Rohit Prakash Baisane	Swatika	1 dials · 0 reached	8 dials · 2 reached	15:48 min	Borderline
sujan namala	Swatika	0 dials · 0 reached	2 dials · 2 reached	15:21 min	Doable / Fast-Track
Raghav panwar	John	2 dials · 0 reached	2 dials · 2 reached	14:51 min	Doable / Fast-Track

The honest other side

Where the humans still win — and why this is a hybrid, not a replacement

Per dial, humans are far more efficient. A human connects on 38% of attempts; the AI on 15%. The AI only gets to 78% reach by brute-forcing ~10 dials a lead — a human's time is too valuable to spend that way, which is exactly the job the AI should take off their plate.

15 leads were reached deeply by a human and never by the AI — including cases where a human held a long call the AI couldn't land. Skilled humans still close relationships the AI opens.

The recommendation writes itself: let the AI do the tireless first-contact and qualification pass on every lead — the coverage the team cannot physically sustain — and route the reached, interested, and document-stage cases to the humans, whose per-dial skill and depth is where they earn their seat. The AI isn't beating the team at selling; it's doing the effort the team is skipping.

Lead by lead — team effort vs AI effort

Every one of the 415 leads, both methods side by side.

Lead	Advisor	Team dials	Reached	Best call	AI dials	Reached	Best call	Verdict	Edge