How long does an AI SDR take to ramp versus a human SDR?

A human SDR takes an average of 3.0 months to ramp after their start date, according to the Bridge Group’s 2025 survey of 351 B2B companies, and you have to recruit them first. An AI SDR skips the learning ramp but not the market’s clocks. In our worked model, the AI SDR spends 9.7–11.1 of its first 12 months at its own steady state. The human manages 7.5–9.5.

How long does a human SDR take to ramp?

A human SDR’s average ramp time is 3.0 months, the lowest the Bridge Group has recorded since 2010 and down from a peak of 3.8 months in 2014. Its 2025 SDR Models, Motions & Metrics report surveyed 351 B2B companies, 83% of them B2B SaaS. Ramp is counted from the start date, so it excludes the time it takes to hire the person.

Three further figures from the same report shape the real timeline:

• Attrition: 40% median annual turnover in 2024 (25th to 75th percentile: 21% to 57%), so a two-seat team should expect to re-run hiring and ramp most years.
• Tenure: 1.9 years on average. On those averages, ramp takes up roughly an eighth of a typical SDR’s time in the seat (3.0 months of about 23).
• Quota attainment: only 60% of SDRs were at quota, the lowest in the study’s history. A completed ramp doesn’t guarantee a productive rep.

We found no credible public benchmark for how long it takes to recruit an SDR specifically, so the worked model below uses one, two and three months as labelled assumptions.

How long does an AI SDR take to ramp?

An AI SDR has no learning curve in the human sense, but it still has a calibration period. The script, the qualification rules and the knowledge base get corrected against real conversations until the booking rate settles. No independent survey measures how long that takes yet. The Bridge Group’s 2025 report is the first edition in which AI SDRs appear as a distinct category at all, at 1% of respondents.

So the honest answer is a set of stages, not a benchmark: setup (data, integration, script, sending infrastructure), then calibration at partial output, then steady state. The week-by-week shape of that for a full outbound programme, including which metrics become readable when, is on our page about how long AI outbound takes to ramp. This page is about the head-to-head with a human hire.

Shared clocks and private clocks: what actually sets each ramp

The shared-clocks rule says that any delay set by the market, the carriers or the inbox providers applies equally to an AI SDR and a human SDR. Only recruiting and learning are private to the human. Most “AI ramps in days” claims compare the AI’s private clock with the human’s total.

The longest single stage for a human SDR is recruiting plus ramp, at four to six months in our bands. For an AI SDR the longest stage is usually a shared clock, most often sending reputation on a new email domain, which no vendor can shortcut.

What makes each one slip: for the human, a slow hire, a mid-ramp resignation or a manager with no time to coach. For the AI, a dirty list, an SMS campaign rejection that restarts registration, or nobody on your side reviewing transcripts during calibration.

Productive months in year one: the worked calculation

This model counts how many of the first 12 months, measured from the day you decide, each option spends at its own steady state. Ramp and calibration months count at half output, which is an assumption standing in for a linear ramp. Every band input other than the Bridge Group’s 3.0-month ramp and 40% attrition is a labelled assumption. Replace them with your own.

To recompute the mid band: the human loses 2 months to recruiting and half of 3.0 months to ramp, so 12 − 2 − 1.5 = 8.5. The AI loses 4 weeks of setup plus half of 6 weeks of calibration, which is 7 weeks or 1.6 months (at 4.33 weeks a month), so 12 − 1.6 = 10.4. The gap is about two months in year one. After that it widens by the human’s re-ramp each year.

A steady-state month from an AI SDR and from a human SDR aren’t the same output. This table tells you when each one is working. What each produces per month, and what it costs per qualified meeting, is a separate calculation. Ours is in AI SDR vs human SDR cost, and the human seat is priced line by line in what a US SDR costs in 2026.

When should ramp time decide between an AI SDR and a human SDR?

Ramp time should decide the choice only when the deadline is shorter than the human’s recruiting-plus-ramp clock. In every other case, cost per meeting and meeting quality matter more.

The sales-cycle row matters more than it looks. Whichever option books the meeting, revenue arrives one sales cycle later. Our B2B SaaS sales cycle benchmarks show how long that is for your deal size.

What shortening the ramp yourself involves

You can compress both ramps without outside help, at a cost in time and skill:

• Human: recruit before the seat is open, give the new rep a pre-built list and tested scripts on day one, and have a manager review calls daily for the first month. The manager’s time is the real cost.
• AI: register the US SMS brand and campaign and start warming sending domains before you choose the vendor, because those clocks are shared and can run in parallel. Then assign someone to read transcripts every day during calibration. Calibration stalls without that.
• Both: clean the target list first. A bad list slows every clock at once.

The setup and calibration stages are the part an AI appointment-setting service carries on its own infrastructure, with its sending reputation and scripts already running. LeadsNow has booked 50,769+ AI-booked sales appointments since 2017. Your sales cycle stays yours either way.

Frequently asked questions

What is the average ramp time for a human SDR?

3.0 months, according to the Bridge Group’s SDR Models, Motions & Metrics report (10th edition, February 2025, 351 B2B companies). That is the lowest since 2010, down from a peak of 3.8 months in 2014. It is measured from the start date, so recruiting time comes on top.

Is there a benchmark for AI SDR ramp time?

Not a survey-grade one. The Bridge Group’s 2025 report is the first edition where AI SDRs appear as a distinct category, at 1% of respondents. Vendor ramp claims are vendor claims. Build your estimate from the stages instead: setup, calibration at partial output, then steady state.

How long does US SMS registration add to an AI SDR’s ramp?

Microsoft’s Azure Communication Services documentation says A2P 10DLC brand registration with The Campaign Registry typically takes two to three business days, and campaign registration typically three to five. Either can take longer if information is missing. Start it before you choose a vendor, because the clock is shared.

Does an AI SDR need email domain warm-up like a human SDR does?

Yes. Inbox-provider rules apply to whoever sends the email. Google’s email sender guidelines say to increase sending volume slowly to avoid delivery problems, and require spam rates below 0.3%. Google publishes no fixed number of days, so treat any exact warm-up period as an estimate.

Does a faster ramp mean revenue arrives sooner?

Only by the size of the ramp gap. Revenue from a booked meeting arrives one sales cycle later whoever booked it. In our worked model the gap is about two months in year one, which matters on a 30-day sales cycle and barely registers on a nine-month one.