Save time
Collectors spend their day on negotiations
Voice agents make the reminder calls from a contact list, retry the ones nobody answered and transfer the debtor to a collector when a negotiation needs a person.
Done with Campaigns Collections
Call every account on the list, and evidence whether each call followed your rules.
Drellia voice agents take your pre-due and early-stage reminder calls inside your calling hours and record how each one ended: promise to pay, callback, dispute. Every recorded call, human or AI, is then scored against your own collections script, so quality review stops being a sample.
- Reminder and early-stage campaigns
- Promises to pay recorded per call
- Script checks on every call
What changes
Time, performance, cost and customers: what Drellia changes for your team.
Improve performance
Every collector scored on all of their recorded calls
Each call is scored against your script, question by question. Results roll up per collector and team, so coaching starts from the steps a collector actually skips.
Done with Dashboards Save money
Collector hours go where a person changes the result
A voice agent uses 1 credit per minute, so a 2-minute reminder call uses 2 credits and no collector time, and it can send the payment link during the call. The worked example below shows the math for a month.
Done with Campaigns Customers and compliance
Respectful calls you can evidence
Check on every call that the collector reached the right person, gave the disclosures your rules require and handled disputes your way. A miss you flag raises a high, medium or low alert with the quote from the call.
Done with Alerts Use cases
Collections: the calls Drellia takes on.
Each use case starts from how the work runs today, shows the part of Drellia that handles it and what your team gets out of it.
Voice agents
Pre-due and early-stage reminders
- Today:
- Reminder lists grow faster than the team can dial, and a small balance takes as much agent time as a large one.
- With Drellia:
- Upload the list as a CSV, pick a voice agent and set calling hours, retries and how many calls run at once. The agent follows the script in its prompt, including how it confirms it is speaking to the right person before mentioning the debt, and its recorded calls can be scored against the same questionnaire as your collectors.
- What you get:
- Every account on the list is queued and dialed inside your hours, with retries, and you follow attempts and outcomes while the campaign runs.
Voice agents
Promises to pay you can follow up
- Today:
- Promises and callback dates end up in free-text notes that are hard to report on and easy to lose.
- With Drellia:
- Each call ends with an outcome from your own catalog, such as promise to pay, callback requested, dispute or wrong person. For a promise, the agent records the amount, the date and the payment method, and can send a payment link over WhatsApp or SMS while the customer is still on the line.
- What you get:
- Filter calls by outcome, export the promises with their amounts and dates, and load them into a reminder campaign ahead of the promised date.
Voice agents
Disputes and hardship reach a person
- Today:
- An automated call that keeps pushing on a disputed debt, or on a customer in hardship, turns into a complaint.
- With Drellia:
- Give the campaign a transfer number. The agent hands the call to a collector when the conversation needs one, following the instructions in its prompt.
- What you get:
- Automation covers the routine reminders; the agent is instructed to transfer disputes and hardship, and Audit can flag recorded calls where it didn't.
Audit
Right-party contact and disclosures on every call
- Today:
- QA can listen to a few calls per collector each month. A skipped identity check or a missing disclosure on the other calls goes unseen.
- With Drellia:
- Turn your collections script into a questionnaire: right-party verification, required disclosures, balance and due date stated, no pressure. Drellia scores every call, human or AI, against it.
- What you get:
- Coverage goes from the sample you had time for to every recorded call from your phone system, with a result for each question.
Audit
Alerts on the misses that matter
- Today:
- The call that broke a rule is found weeks later, if it is found at all.
- With Drellia:
- Link a high alert to the questions that must never be missed, such as the disclosure, and add conversation alerts for threats or abusive language.
- What you get:
- Supervisors start the day from the flagged calls and the exact quote, instead of listening at random.
Chat
Portfolio questions without a report request
- Today:
- Questions like which campaign produced the most promises wait for someone to build a report.
- With Drellia:
- Ask Drellia Chat in plain words, for example “Which campaign got the most promises to pay last week?”
- What you get:
- A short summary and a chart from your own data, in the meeting where the question came up.
A day with Drellia
The same work, before and after.
Reminder calls
Today Collectors dial the list between negotiations; unanswered numbers wait until someone has time.
With Drellia A voice agent campaign calls the list inside your hours, retries by your rules and transfers the debtor to a collector when needed.
Recording outcomes
Today Free-text notes typed after the call.
With Drellia An outcome from your catalog on every AI call, with the promised amount and date, the transcript and the recording.
Quality review
Today A supervisor listens to a hand-picked sample, for example 2 in every 100 calls.
With Drellia Every recorded call is scored against the same questionnaire, human or AI.
Compliance evidence
Today Rebuilt from recordings when a complaint or an audit arrives.
With Drellia Per-question results, alerts with the quote and CSV exports, ready when someone asks.
Management questions
Today Wait for the next report.
With Drellia Asked in Chat and answered with a summary and a chart.
Illustrative example
What a month could cost, worked out step by step.
An example with its assumptions in plain sight, so you can redo it with your own volumes. It is not a quote, and it does not claim any saving.
Illustrative example
Assumptions
- 1,000 answered reminder calls a month, 2 minutes each on average
- 1,000 collector calls a month, recorded outside Drellia, 5 minutes each, scored against a 10-question questionnaire (billed per block of 10 questions, so the 7-question sample below costs the same)
- Not included: evaluating the voice agent calls themselves, Chat answers, transferred minutes and phone-line costs. Rates and prices as published on the pricing page.
If people did this work
- About 33 hours of talk time on the voice agent calls (1,000 × 2 min), before dialing, unanswered attempts and wrap-up.
- About 83 hours to listen to every audited call once (1,000 × 5 min). A hand-picked sample of 2 in every 100 would cover 20 calls.
Credits used
- Voice agent: 1,000 calls × 2 min × 1 credit per minute
- 2,000
- Audit: 1,000 calls × 0.3 credits (0.05 for the evaluation + 5 min × 0.05 credit per minute for transcription)
- 300
- Total per month
- 2,300 credits
Fits the Standard plan: 2,500 credits a month for €390 per month, billed yearly. That is about €0.16 per credit.
Real usage depends on your call volumes and lengths. See plans and credit rates
Sample questionnaire
What Audit would check on your calls.
A starting point to edit with your own rules. Each question has an answer type, the answer that counts as compliant and the alert a miss raises. Drellia scores every recorded call against it.
Collections call check
7 questions-
Did the agent confirm they were speaking to the debtor before discussing the account?
- Answer type
- Yes / No
- Compliant answer:
- Yes
- Alert if missed:
- High
-
Did the agent give the disclosures your policy requires, such as saying the call is about collecting a debt?
- Answer type
- Yes / No
- Compliant answer:
- Yes
- Alert if missed:
- High
-
Did the agent state the balance and the due date correctly?
- Answer type
- Yes / No
- Compliant answer:
- Yes
- Alert if missed:
- Medium
-
If the customer disputed the debt, how was it handled?
- Answer type
- Multiple choice
- Compliant answer:
- No dispute · Dispute logged and escalated
- Alert if missed:
- High
-
Did the agent avoid threats, pressure and misleading statements?
- Answer type
- Yes / No
- Compliant answer:
- Yes
- Alert if missed:
- High
-
What payment did the customer agree to (amount and date)?
- Answer type
- Open answer
- Compliant answer:
- Captured as text, not scored
- Alert if missed:
- No alert
-
Did the agent close with a clear summary of the next step?
- Answer type
- Yes / No
- Compliant answer:
- Yes
- Alert if missed:
- Low
FAQ
Collections: common questions
Can an AI voice agent make debt collection calls?
Yes, for reminder and early-stage calls. You write the agent's instructions in plain words, upload a contact list and set calling hours, retries and how many calls run at once. The agent records an outcome for each call and can transfer the call to a collector when the conversation needs a person.
Does Drellia make our collections calls compliant?
No tool can do that for you. Drellia helps you check and evidence your own rules: you write them as questions, every call is scored against them, and misses you flag raise alerts with the quote from the call. Your compliance team decides what the rules are.
Can we audit the calls our human collectors make?
Yes. Connect your phone system over SIP, upload recordings from your current system or send them through the API. Recorded voice agent calls flow into Audit as well, and both are scored against the same questionnaire.
Can Drellia tell us whether promises to pay were kept?
Not by itself. Drellia records each promise to pay with its amount and date. Whether the payment arrived lives in your own collection system, so export the promises, or read campaign call sessions through the API, and match them there.
How are voice agent calls charged?
In credits: one minute of voice agent uses 1 credit. Plans include monthly credits and start from €120 per month, billed yearly.
Which languages do the agents and the audit handle?
A voice agent speaks the one language you choose for it, from any language its voice supports, including English and Spanish. Audit evaluates conversations in English and Spanish.
Other industries
Next step
See Drellia on your own calls.
Tell us how your operation works, and we will walk you through the use cases that fit it in a 20-minute demo.
Book a demo