Mystery-shop every rep,
every week.
Ai Shadow Caller dials your phone reps posing as a real customer, scores them against your rubric, and delivers a transcript + coaching note within minutes. Continuous quality assurance without the $200/shop bill or the 6-week turnaround.
For call centers, multi-location practices, and operations teams managing 20+ phone reps.
The problem
You only find out your reps are off-script
when a customer complains.
Most teams can't afford to run them more than 2-4 times a year.
By the time you read the findings, the rep has had six more bad calls.
Two shoppers grade the same call differently. Trend lines mean nothing.
How it works
Five steps to continuous phone-rep coaching.
Write your scenario
Tell the AI who it's pretending to be and why it's calling. "You're a new patient asking about pricing — friendly, a little nervous." 30 seconds, plain English.
Build a rubric
List 5-10 things the rep should do. "Greeted within 10 seconds." "Mentioned new-patient pricing." "Offered to book." Yes/no — that's it.
Add your reps
Name + phone number. Or upload a CSV of 50 reps. Tag them by team if you want.
Schedule a campaign
Run it once, weekly, biweekly, monthly. AI rotates through 10 different voices so reps can't spot the pattern.
Get the receipts
Every test produces a transcript, a rubric breakdown, a coaching note, and an overall score. Drill into any call to coach with the exact line they missed.
Built for
Any team that answers the phone for a living.
Test how front-desk staff handle insurance questions, new-patient inquiries, and appointment booking under pressure.
"I have Cigna, do you take that?" "Can I book my husband and me on the same day?"
Verify that hosts handle reservations, dietary requests, and large-party inquiries with the warmth your brand demands.
"I have a gluten allergy, what can you accommodate?" "Can we get a table for 12 next Saturday?"
Audit how reps explain coverage options, handle objections, and stay compliant with your scripts. Identify training gaps before they show up in regulators' complaints.
"I need to lower my premium." "Why did my rate go up?"
Catch the rep who quotes too low, forgets the warranty pitch, or fails to book a follow-up. Coach by example, with the transcript in hand.
"How much for an oil change?" "Do you guys do alignments?"
Why Shadow Caller
Better data, faster, at a fraction of the cost.
10× cheaper than human mystery shopping
Human shopper services charge $50-200 per shop. Ai Shadow Caller starts at under $1 per test. Run 500 tests this quarter for less than two human shops.
Continuous, not quarterly
Most teams "audit" twice a year and find out about service problems after they've already cost customers. With Shadow Caller you find out this week.
Standardized scoring
Human shoppers grade subjectively. The AI uses your rubric, the same way, every time. Trend lines on individual reps actually mean something.
Built on the same voice tech as our receptionist product
PressTwo already runs a multi-tenant voice agent network for inbound calls. Shadow Caller is the outbound mirror — same battle-tested Retell + Claude infrastructure, dedicated agent persona.
Pricing
Pay per test. Volume discounts kick in immediately.
Three flat plans, plus enterprise for teams running thousands of tests a month. Tests reset monthly. Annual saves 15%.
Starter
Single-location, 5-15 reps
or $1,499/yr · save 15%
Overage past 100 bills at $1.49 each.
Growth
1-3 locations, 15-50 reps
or $4,995/yr · save 15%
Overage past 500 bills at $0.99 each.
Scale
Call centers, 50-200+ reps
or $15,000/yr · save 15%
Overage past 2,000 bills at $0.75 each.
Enterprise
5,000+ tests / custom rubrics
Custom test volumes, custom rubrics, dedicated success manager, phone-system integration (Five9, NICE, RingCentral). For teams running 5,000+ tests / mo.
Compliance & ethics
We test reps. We don't deceive customers.
- ·The AI never claims a false identity (no "I'm Sarah from Cigna"). It plays a prospective customer scenario — same as a human mystery shopper.
- ·Employer consent for testing + recording is captured during onboarding. Your MSA covers the legal basis.
- ·Recording disclosure auto-toggles per state (two-party consent states like California, Florida, Illinois detected from rep area code).
- ·Results are owned by you. Used for coaching — never used for automated terminations or PIPs without your manager review.
Ready to see what your reps actually sound like?
We're running design-partner pilots with a few teams across healthcare, hospitality, and insurance. 30 days, free, real data. If your team handles inbound phone volume, we want to talk.