Dialer software is a system that places outbound phone calls automatically from a contact list, so sales and service teams spend their time talking instead of dialing. The main types are preview, power (progressive), predictive and auto dialers that play recorded messages. A newer option, the AI calling agent, dials and also holds the conversation itself.
This buyer's guide compares dialer types, explains how predictive dialer technology paces calls and why abandoned calls happen, and summarizes the rules in the US, UK and India, including India's September 2026 rules for AI-voiced calls. It then covers where AI calling agents fit, what each option costs, and a pre-launch checklist.
What is dialer software?
Dialer software sits between your contact list and the phone network. It takes numbers from a CRM or spreadsheet, places the calls, detects the outcome (busy, no answer, voicemail or a live person) and passes connected calls to an agent with the customer's record. Teams use it for prospecting, follow-up, collections, renewals and reminders.
A rep who dials by hand spends much of each hour listening to rings and voicemail greetings; a dialer removes that idle time. How aggressively it does so separates the main types. It is also what regulators care about, because dialing ahead of available agents is what produces silent and abandoned calls.
What are the types of dialers? Preview, power, predictive and auto dialer software
Vendors use the names inconsistently, especially “power dialer”, so check how a product actually paces calls rather than trusting the label.
- Preview dialer: the rep sees the record first, then clicks to call. Slowest, but best for complex or high-value calls.
- Power or progressive dialer: dials automatically as soon as a rep is free, at a fixed number of lines per rep, often one. Abandoned calls are rare.
- Predictive dialer: an algorithm dials more numbers than there are free reps, forecasting who will answer and when reps will finish. Most talk time per hour, but some answered calls find nobody available.
- Auto dialer or voice broadcasting: dials a list and plays a recorded or synthetic message, sometimes with a keypress menu. In most markets this is the most tightly regulated pattern.
| Option | How it paces calls | Abandoned-call risk | Best for |
|---|---|---|---|
| Preview dialer | Rep reviews the record, then dials one number | Very low | Account-based sales, complex B2B, collections that need notes |
| Power / progressive dialer | Dials when a rep is free, at a fixed number of lines per rep | Low | Small and mid-size sales teams, follow-ups, warm lists |
| Predictive dialer | Algorithm dials ahead of rep availability and adjusts the ratio | Highest; must be capped | Large consumer lists with low answer rates and many agents |
| Auto dialer / voice broadcast | Dials a list and plays a recorded or synthetic message | No agent involved, but strict consent rules | Opted-in reminders, alerts and notifications |
| AI calling agent | Dials and holds the conversation itself, one agent per call | No agent shortage; consent rules still apply | Qualification, reminders, confirmations, callbacks |
A short definition of each dialing mode, and what changes when the agent is software.
How does predictive dialer technology pace calls?
A predictive dialer runs a feedback loop. It watches the answer rate, call lengths and when each agent is likely to become free, then sets how many lines to dial per agent. Cisco's public documentation for its Outbound Option dialer shows the moving parts: in predictive mode, the configured lines per agent is only a starting value, a maximum caps how far the dialer can go, and it adjusts after a set number of live connections by an amount set by a gain parameter. The abandoned-call limit is a configurable percentage with a default of 3.
Abandoned calls happen when the prediction is wrong. If more people answer than expected, or agents stay on calls longer, someone picks up and nobody is there. Answering machine detection adds another source: when it mistakes a live person for a voicemail greeting and hangs up, that person gets a silent call. Ofcom's 2008 policy statement concluded that companies using answering machine detection must include a reasoned estimate of those false positives in their abandoned-call rate. In Cisco's API, the default treatment for a detected answering machine is to abandon the call, so check that setting, and the maximum lines per agent, in any product before going live.
Dialer compliance rules in the US, UK and India
Every major market regulates silent calls, abandoned calls and automated voices. Here is a summary for three large markets, checked against primary sources on 28 October 2026.
United States: the TCPA, FCC rules and the FTC's Telemarketing Sales Rule
Under the Telephone Consumer Protection Act and the FCC's rules in 47 CFR 64.1200, a telemarketer may not abandon more than 3 percent of calls answered live, measured over a 30-day period per campaign. A call is abandoned if it is not connected to a live sales representative within two seconds of the person's completed greeting; if no rep is free, a prerecorded message must start within those two seconds, naming the seller, stating the call was for telemarketing and offering a do-not-call number and an automated opt-out. Unanswered calls must ring for at least 15 seconds or four rings. The FTC's Telemarketing Sales Rule sets a parallel 3 percent safe harbor.
Telephone solicitations are limited to 8 a.m. to 9 p.m. in the called party's local time and may not go to numbers on the National Do Not Call Registry. Artificial or prerecorded voice messages must identify the business at the start and give a callback number.
AI voices are covered. In a declaratory ruling adopted on 2 February 2024, the FCC confirmed that the TCPA's restrictions on an “artificial or prerecorded voice” include AI technologies that generate human voices, so these calls need prior express consent, and prior express written consent for telemarketing. The FCC added that the TCPA allows no carve-out for technologies that claim to be the equivalent of a live agent. Consumers can sue for their actual loss or $500 per violation, whichever is greater, and courts may triple that for willful or knowing violations. Separately, in Facebook v. Duguid (1 April 2021) the Supreme Court held that equipment is an autodialer only if it can store or produce numbers using a random or sequential number generator, but the artificial-voice rules apply however a number was dialed.
Consent, calling hours, Do Not Call and how the 2024 AI-voice ruling applies.
United Kingdom: Ofcom's persistent misuse policy and PECR
Ofcom deals with silent and abandoned calls under its persistent misuse powers in the Communications Act 2003 and can impose penalties of up to £2 million. Its current policy statement, published on 20 December 2016 and applied from 1 March 2017, says calling parties should not make silent or abandoned calls at all. Ofcom calculates an abandoned call rate per calendar day, per campaign or call center, and uses it to prioritize enforcement rather than as a safe harbor; the earlier 2008 statement had referred to a 3 percent ceiling per campaign over 24 hours.
Factors Ofcom says can aggravate a case include an information message that starts later than two seconds after pick-up, or that fails to identify the caller, omits a basic-rate number for opting out or contains marketing; calling an abandoned number again within 72 hours without a guaranteed live agent; ringing for less than 15 seconds; and withheld caller ID. It expects compliance records to be kept for at least six months. Separately, the ICO's PECR guidance says automated marketing calls with a recorded message need specific consent to automated calls, live marketing calls must be screened against the TPS and Corporate TPS, and your number must be displayed.
India: TRAI's 140 series and the 2026 rules for AI-voiced calls
India's rules come from TRAI's Telecom Commercial Communications Customer Preference Regulations (TCCCPR). In its February 2025 amendment, TRAI said promotional auto-dialer or robo-calls should be permitted through 140-series numbers only, with the 1600 series allotted for service and transactional calls, and required senders to tell their operator in advance about auto-dialer and robo-call use.
On 18 September 2026, TRAI notified the Third Amendment. It defines an Application-to-Person (A2P) call as a voice call initiated by an application, software system or automated platform without direct human dialing, including autodialing, robo-calls and pre-recorded or artificial voice technologies. Senders must declare A2P use to their originating operator in advance, with the caller ID (CLI) ranges they will use, recorded on the DLT platform; undeclared A2P calls are treated as unsolicited commercial communication. Terminating operators may also charge originating operators up to ₹0.05 per minute for A2P calls, except calls from the 140xx, 1600xx and 1601xx series. TRAI's reasoning: in April 2026, 15,736 of 20,591 disconnections for violations, about 75 percent, were SIP/PRI connections used for A2P calls.
Most of the amendment takes effect 30 days after publication in the Official Gazette, but the A2P declaration and the termination charge take effect after 60 days, and some complaint and appeal provisions after 90. TRAI also declined to exclude calls where a live agent participates, noting that networks may not be able to tell whether a human or an artificial voice is speaking.
Registration, headers, the 140 and 1600 series, and what the 2026 amendment changes.
AI dialer vs AI calling agent: what is the difference?
An AI dialer, sometimes sold as an autocaller AI, is still a dialer: it uses machine learning to choose when and whom to call, detect voicemail or score leads, then connects a human rep. An AI calling agent places the call and holds the conversation itself, asking questions, answering from approved information, booking a slot and logging the outcome, and it hands over to a person when needed.
A dialer connects humans faster, so its ceiling is the number of reps on shift, and over-dialing to keep them busy creates abandoned calls. An AI agent does not wait for a free rep, so there is no reason to dial ahead of capacity. What it does trigger is the artificial-voice rules: consent under the FCC's 2024 ruling in the US, PECR's specific-consent rule for automated marketing calls in the UK, and A2P declaration in India.
| Predictive or power dialer | AI calling agent | |
|---|---|---|
| Who holds the conversation | Your human reps | The AI agent, with a warm transfer to a rep when needed |
| What limits volume | Reps on shift | Concurrent phone lines and the size of your list |
| Abandoned-call risk | Built into predictive pacing; must be capped | No rep shortage, so no reason to dial ahead of capacity |
| Main compliance triggers | Abandonment, calling hours, Do Not Call lists | The dialer rules, plus artificial-voice consent and, in India, A2P declaration |
| Typical pricing | Per seat per month, plus calling minutes | Per minute of conversation, plus telephony |
| Best at | High-value or complex conversations that need a person | High-volume, repeatable calls: qualification, reminders, confirmations, callbacks |
When should you use a dialer, an AI agent or both?
- Preview or power dialer: when every conversation is complex or valuable, such as enterprise sales, negotiated renewals or collections that need judgment.
- Predictive dialer: only with a large team, large lists with low answer rates, and a reliable way to keep abandonment within your market's limits.
- AI calling agent: for repeatable calls at volume, such as calling new leads back within minutes, appointment reminders, order confirmations, renewal nudges and after-hours callbacks.
- Both: the AI agent calls every lead, asks your qualification questions and warm-transfers interested prospects to a rep with a summary, while everyone else gets a booked follow-up or a polite close.
The hybrid model keeps people on the calls that need them. Vaami's AI voice agents place outbound calls, handle interruptions, update your CRM and calendar, and transfer to a person with context when the prospect is ready. Replies come with sub-300ms end-to-end response latency in production, in 30+ languages including Hindi, Hinglish, Tamil and Spanish, switching mid-call when the caller does. For India, outbound calling is TRAI/DLT-aware.
How an AI agent qualifies leads by phone and books qualified prospects into your reps' calendars.
Campaigns, retries, warm transfers and CRM updates from one outbound AI agent.
How much does dialer software cost?
Dialer software is usually priced per seat, per user per month, often with an annual commitment, with calling minutes bundled for some countries or billed on top. Two examples from vendors' own pricing pages, checked on 28 October 2026: Aircall lists its Professional plan, which includes a Power Dialer, at $50 per license per month billed annually, with a minimum of three licenses. CloudTalk sells a Power Dialer add-on at €15 per user per month billed annually on its lower plans, and a Parallel Dialer add-on that dials up to 10 numbers at once at €39 per user per month billed annually.
The bigger cost sits outside the software: every rep hour, plus numbers, minutes, list data and Do Not Call scrubbing. AI calling agents are usually priced per minute instead. Vaami charges $0.025 per minute for AI usage (₹1.99 per minute in India), billed per second, with telephony billed separately or brought by you, and no setup fee, seat fee or monthly minimum on the standard rate. To compare fairly, work out the cost per qualified conversation, including rep time spent on calls that go nowhere.
How sales teams use AI agents for speed to lead, qualification and follow-up.
Compliance checklist for outbound call center solutions
- 01Classify every campaign: promotional or telemarketing, service or transactional, and whether it uses an artificial or prerecorded voice.
- 02Record consent for each number, with wording, timestamp and source, and act on revocations quickly.
- 03Scrub lists before every campaign: the National Do Not Call Registry in the US, TPS and CTPS in the UK, customer preferences under the TCCCPR in India, and your own do-not-call list.
- 04Respect calling hours: 8 a.m. to 9 p.m. local time for US telephone solicitations, and no unsociable hours anywhere.
- 05Cap abandonment: set the abandon limit and maximum lines per agent, count answering machine detection false positives, and play a compliant identification message within two seconds.
- 06Ring unanswered calls for at least 15 seconds, and in the UK do not call an abandoned number again within 72 hours unless a live agent is guaranteed.
- 07Present a valid caller ID. In India, use the 140 series for promotional calls and declare A2P calling and its CLI ranges to your operator.
- 08Say who is calling and on whose behalf, tell people when they are speaking with an AI agent, and offer a person on request.
- 09Keep call logs, consent evidence and dialer settings; Ofcom expects at least six months of records.
Sources
- US eCFR, 47 CFR 64.1200 Delivery restrictions, version current as of 1 October 2026: ecfr.gov/current/title-47/chapter-I/subchapter-B/part-64/subpart-L/section-64.1200
- US eCFR, 16 CFR 310.4 Abusive telemarketing acts or practices (Telemarketing Sales Rule), version current as of 1 October 2026: ecfr.gov/current/title-16/chapter-I/subchapter-C/part-310/section-310.4
- Federal Communications Commission, Declaratory Ruling FCC 24-17, Implications of Artificial Intelligence Technologies on Protecting Consumers from Unwanted Robocalls and Robotexts, adopted 2 February 2024, released 8 February 2024: docs.fcc.gov/public/attachments/FCC-24-17A1.pdf
- Cornell Law School LII, 47 U.S. Code § 227, Restrictions on use of telephone equipment: law.cornell.edu/uscode/text/47/227
- Supreme Court of the United States, Facebook, Inc. v. Duguid, No. 19-511, decided 1 April 2021: supremecourt.gov/opinions/20pdf/19-511_p86b.pdf
- Ofcom, Persistent Misuse: a statement of Ofcom's general policy on the exercise of its enforcement powers, 20 December 2016: ofcom.org.uk/__data/assets/pdf_file/0027/96309/Persistent-Misuse-Policy-Statement-edit.pdf
- Ofcom, Revised statement of policy on the persistent misuse of an electronic communications network or service, 10 September 2008: ofcom.org.uk/phones-and-broadband/unwanted-calls-and-messages/persistent_misuse
- ICO, Telephone marketing (Guide to PECR), accessed 28 October 2026: ico.org.uk/for-organisations/direct-marketing-and-privacy-and-electronic-communications/guide-to-pecr/electronic-and-telephone-marketing/telephone-marketing
- TRAI, Telecom Commercial Communications Customer Preference (Second Amendment) Regulations, 2025, 12 February 2025: trai.gov.in/sites/default/files/2025-02/Regulation_12022025.pdf
- TRAI, Telecom Commercial Communications Customer Preference (Third Amendment) Regulations, 2026, 18 September 2026: trai.gov.in/sites/default/files/2026-09/Regulation_18092026.pdf
- TRAI, Press release on the TCCCPR Third Amendment, 18 September 2026: trai.gov.in/sites/default/files/2026-09/PR_No119of2026.pdf
- Cisco DevNet, Outbound Campaign API, Packaged Contact Center Enterprise, accessed 28 October 2026: developer.cisco.com/docs/packaged-contact-center/outbound-campaign-api
- Aircall, Pricing, accessed 28 October 2026: aircall.io/pricing
- CloudTalk, Pricing, accessed 28 October 2026: cloudtalk.io/pricing
- Vaami, Pricing, accessed 28 October 2026: vaami.ai/pricing
See how Vaami is deployed in this industry — real use cases, sample transcripts, and ROI data.
Frequently asked questions
What is dialer software?
What is the difference between a predictive dialer and a power dialer?
What is the maximum abandoned call rate for a predictive dialer?
Are AI voice calls legal under the TCPA?
Do AI calls need to be declared to telecom operators in India?
Is an AI dialer the same as an AI calling agent?
How much does dialer software cost?
Can an AI calling agent transfer a call to a human sales rep?

Sandeep is a founding member of EmpowerX Labs and a core technical member of the team. A technology leader with an engineering-manager background, he brings deep technical experience, has built multiple startups and products, and leads AI development at Vaami.