Last updated: October 27, 2026
An AI SDR works when the company has a low average contract value and a large pool of possible buyers, and it fails when the contract is large and the pool is small. That is the view of Jorge Macias, founder of The GTM Engineering Company, and this guide builds on it. You will see where an AI SDR fits, what the main vendors charge as of October 27, 2026, which sending limits apply, and what a human-in-the-loop alternative looks like.
Disclosure: The GTM Engineering Company builds with Clay, and its founder, Jorge Macias, is a member of Clay's Expert Program, where Clay lists him as a vetted expert.
Key Takeaways (TL;DR)
- Verdict: Jorge Macias of The GTM Engineering Company sorts every AI SDR decision on two axes, average contract value (ACV) and total addressable market (TAM). Low ACV with a big TAM suits an AI SDR. High ACV with a small TAM belongs to humans.
- Where the opening is: The mid-market on either axis. Jorge says that is where human-in-the-loop workflows, or deterministic AI outreach for chosen segments and CRM contacts, have the most room.
- What it costs: AiSDR lists $1.50 per person prospected. The Salesforge Agent Frank add-on is $499 a month billed quarterly. Amplemarket Startup is $600 a month on an annual term. 11x and Artisan show no dollar prices on their pages.
- What can go wrong: Google's email sender guidelines tell senders to keep the spam rate in Postmaster Tools below 0.10% and to avoid ever reaching 0.30%. A tool that sends high volume from one domain can cross that line.
- Evidence gap: The only public AI SDR figures with numbers are self-reported by the companies running them. Treat every response-rate claim as unaudited.
AI SDR: at a Glance
An AI SDR is software that finds prospects, writes outreach, sends it, handles replies and books meetings. The table below condenses the article, and each row points to the section that backs it up.
| Question | Short answer | Source of the answer |
|---|---|---|
| Do AI SDRs work? | Yes, in a defined slice of the market | Jorge Macias; SaaStr's self-reported results |
| Best fit | Low ACV, big TAM | Jorge Macias |
| Worst fit | High ACV, small TAM | Jorge Macias |
| Biggest opening | Mid-market ACV or mid-sized TAM, with human review | Jorge Macias |
| Cost, outbound | $1.50 per person prospected (AiSDR) | AiSDR pricing page, October 27, 2026 |
| Sending limit to respect | Spam rate under 0.10% | Google email sender guidelines |
Read the table as a screening step. A company in the worst-fit row can skip ahead to the build section.
What Is an AI SDR, and Do AI SDRs Work?
An AI SDR is an agent that does the early work of a sales development representative: prospecting, writing, sending, replying and booking. Vendors also call these tools AI sales agents. Outbound tools write to people who never asked to hear from you, while inbound tools such as Qualified's Piper talk to visitors already on your website. Qualified is now owned by Salesforce, which announced the deal on December 17, 2025 and completed it on April 1, 2026, according to Salesforce's newsroom (2025 and 2026).
The short answer
Do AI SDRs work? Yes, for a bounded set of accounts and with a person steering. Jorge Macias starts from the question of when they are useful, and the sections below follow that order. His rule of thumb comes before any tool comparison, since the same product can produce good results in one company and a burned market in another.
What the public numbers show
The most detailed public account comes from SaaStr's six-month write-up, by the publisher and event company that ran five AI agents. It carries no publication date. SaaStr reports a 6.7% overall response rate against an industry average it puts at about 3%, a 4% positive response rate, and 15 to 20 hours of human management each week across the five agents. These figures are self-reported, SaaStr promotes the tools it uses, and nobody audited them.
The same write-up says results rise in weeks when people spend more time on the agents. That matches Jorge's warning that a buyer who expects to set the tool up and walk away does not understand sales and marketing. No independent study with response-rate numbers turned up in research for this guide, so the evidence that AI SDRs work in general is thin.
The Two-Axis Test: ACV and TAM Decide Where an AI SDR Fits
Jorge Macias describes his thinking about AI SDRs as a two-axis table. The first axis is ACV, the average contract value of the company, product or service. The second axis is TAM, how small or large the total addressable market is. The matrix below shows his view, with the two corners and the middle band he described.
| ACV vs TAM | Big TAM | Mid-sized TAM | Small TAM |
|---|---|---|---|
| Low ACV (SMB) | AI SDR fits. Volume is the point and each account matters little. | Open space: human-in-the-loop or deterministic AI outreach for chosen segments | Not part of Jorge's view |
| Mid ACV (not SMB, not enterprise) | Open space: human-in-the-loop or deterministic AI outreach for chosen segments or CRM contacts | Open space: the middle of both axes | Open space: human-in-the-loop or deterministic AI outreach |
| High ACV (enterprise) | Not part of Jorge's view | Open space: human review before anything sends | Human workflow. An AI SDR burns the TAM and makes a weak first impression. |
The matrix is attributed to Jorge Macias, founder of The GTM Engineering Company. The cells marked "not part of Jorge's view" are the two corners he did not describe.
Low ACV and big TAM: where an AI SDR is strong
Jorge says an AI SDR is strong when the ACV is low and the TAM is large. One reading of that rule: a low contract value cannot pay for hours of human research per account, and in a big market a weak email to a few accounts costs little. SaaStr's write-up shows the same split by deal size, since its agent closes event tickets under $1,000 on its own (about 20% of SaaStr's event ticket revenue now comes from AI agents, per the write-up).
High ACV and small TAM: where an AI SDR is bad
When the ACV is high and the TAM is small, Jorge says an AI SDR is a poor choice for two reasons. It burns through the TAM quickly, because every account you email is one of a short list. It also makes a bad first impression, and Jorge points out that humans are naturally better at a first impression than AI, so this stays a human workflow. SaaStr reports the same handoff for sponsorships of $50,000 to $100,000 and above, where the agent qualifies and books the meeting and a person takes over.
The mid-market, on either axis
Jorge says no answer has been found yet in two middle bands. One is a mid-market where the TAM is neither big nor small. The other is a mid-market where the ACV is neither enterprise nor SMB. He expects a lot of the opportunity there, in human-in-the-loop workflows or deterministic AI outreach aimed at certain segments of the market, or at certain people in your CRM.
A PLG example
Jorge's example is a product-led growth motion. A small account that signs up can get an AI SDR, since the cost of a miss is low. A huge account that you badly want to land should get a person who gathers the intel and does manual outreach.
AI SDR vs Human SDR: Who Gets Which Accounts
The ACV and TAM test also answers the AI SDR vs human SDR question without a salary comparison. Give an AI SDR the accounts where volume matters and any single account carries little value. Give a human the accounts where one conversation can be a large share of the year's revenue.
SaaStr's reader-question guide on rolling out an AI SDR points the same way. It suggests starting with ignored leads and high-volume, low-ACV segments, and keeping human SDRs for complex enterprise deals. It also recommends reviewing the first 1,000 emails by hand and budgeting 30 to 45 minutes a day of tuning for the first 60 to 90 days.
Five Checks Before You Switch an AI SDR On
Jorge lists the things to understand before you deal with an AI SDR, and they work as a pre-launch checklist. How much volume is it sending, and from which domain? Will the personalization move the needle, and how short are the messages? How customized is the tool across segment, persona, use case, product and the combination of all of those?
Volume and domain
Volume and domain tie directly to Google's published limits. Google tells senders to keep the spam rate reported in Postmaster Tools below 0.10% and to avoid ever reaching 0.30%. Anyone sending more than 5,000 messages a day to Gmail accounts is a bulk sender and needs SPF, DKIM and DMARC records plus one-click unsubscribe. Warmup and a pre-launch review are the controls, and the Claude Code cold email automation guide walks through both.
Responsibility is a separate matter. The FTC's CAN-SPAM compliance guide says you cannot contract away your legal responsibility to comply, so hiring a vendor to send does not move the liability. The guide sets a penalty of up to $53,088 for each separate email that violates the Act, and requires honoring opt-outs within 10 business days. This is general information, not legal advice.
Personalization, length and segment fit
Personalization counts when it changes whether the reader replies. Short messages and tight segments leave the model less room to write something generic.
Your own time
Jorge adds the question of how much time you will put into it. If you iterate, improve the setup and think through the process, he expects most customers of AI SDR companies who do that are getting better results. SaaStr's 15 to 20 weekly hours is one data point for what that effort costs a team that runs five agents.
AI SDR Cost: Published Prices and the Time You Add
AI SDR cost has two parts: the price on the vendor's page and the hours a person spends reviewing and tuning. Published prices use different units, so a fair comparison needs the arithmetic shown below. Every price here was read on the vendor's page on October 27, 2026, with its billing basis.
What the pages say
AiSDR lists $1.50 per person prospected with no monthly fee, charged daily. Its FAQ gives two campaign minimums, $7.50 and "as little as" $3, and the page does not reconcile them. A warmed-up mailbox is a $33 add-on and a fully managed service is $2,500 a month.
Salesforge lists Pro at $40 a month billed monthly. Its Agent Frank AI SDR is an add-on at $499 a month billed quarterly, which the page ties to 2,000 or more contacts a month (about $0.25 per contact), and access starts with a demo. Amplemarket lists Startup at $600 a month on an annual term, with 2 users and 27,000 contacts.
Illustrative arithmetic
Prospecting 1,000 people on AiSDR's pay-as-you-go rate comes to $1,500, before mailboxes or review time. Agent Frank's $499 covers 2,000 or more contacts a month per its page, which is a different unit from "prospected", so the two cannot be ranked against each other. Add the Pro plan and the Salesforge bases differ too, monthly for Pro and quarterly for the add-on, so no single monthly total is shown here.
The human cost
No pricing page lists time. SaaStr's 15 to 20 hours a week across five agents is the one public number, and the roll-out guide's 30 to 45 minutes a day for the first 60 to 90 days is the other. This guide found no sourced figure for a human SDR's salary, so it does not compare against one.
Best AI SDR Tools and AI Sales Agents: What Each One Is
No independent group has tested these tools against each other, so this table lists what each one is and what it charges, not a ranking. Prices were read on October 27, 2026. Nothing here is a claim that any vendor's product works, and The GTM Engineering Company did not test them.
| Tool | What it is | Published price (October 27, 2026) |
|---|---|---|
| AiSDR | AI SDR for outbound | $1.50 per person prospected, pay as you go |
| Salesforge (Agent Frank) | Email tool with an AI SDR add-on | Pro $40 a month billed monthly; Agent Frank $499 a month billed quarterly |
| Amplemarket | Sales data, signals and outreach | Startup $600 a month, annual term; Growth and Elite custom |
| Qualified (Piper) | AI SDR for website visitors, owned by Salesforce | Not public on the pages checked |
| 11x | AI SDR | Pricing page returned an error on October 27, 2026 |
| Artisan | AI SDR | Credit-based; no dollar figures on the pricing page |
For the signal-driven tools that sit next to this category, the best signal-based outbound tools comparison lists nine options with pricing.
What to Build Instead: Human-in-the-Loop Outreach
A built system keeps the useful parts of an AI SDR and puts a person at the point where Jorge says humans are better. A signal such as a funding round, a hire or a site visit selects the account. A research step enriches it, a model drafts the message, a person approves, and a sender delivers it. Replies go to a person.
The GTM Engineering Company builds this kind of work inside a client's own CRM, and Clay sits at the center of most of its builds. Weigh that relationship to Clay accordingly.
Tools you can build with
Clay handles data and research. Its free plan lists 500 actions and 100 data credits a month, while paid plans show conflicting prices on the pricing page itself, $167 a month in one section and "starting at $185" in the FAQ, so check Clay pricing before budgeting. Claude Code is the logic option; the Team plan is $20 per seat per month billed annually for 2 to 150 seats, and it includes Claude Code. Claygent, Clay's research agent, covers the research step, as described in the Claygent use cases for B2B outbound.
The third piece is a sender. Whichever you pick, the approval step sits before the send, where one wrong message stops at one inbox instead of hundreds. The comparison of Clay vs Claude Code explains why client-facing work stays in visible steps with manual review between them.
Can you replace your GTM stack with Claude and a few APIs?
Claude can draft and reason over research. Contact data, mailbox warmup and the approval step still come from a data source, a sender and a person. The build wires a smaller stack together around those three pieces.
What segmenting a CRM looks like
The Fluint case study shows segmentation, though not an AI SDR. Over 12 weeks The GTM Engineering Company built a 2,062-company audience for ads and a subset of about 300 companies with an ICP fit of 85 or higher for outbound. The outbound subset is small by design. Read the Fluint case study for what was built and the numbers it reports.
Everything You Need to Know About AI SDR
| Topic | What to remember |
|---|---|
| Definition | Software that prospects, writes, sends, replies and books |
| Fit rule (Jorge Macias) | Low ACV and big TAM suit it; high ACV and small TAM do not |
| Open space | Mid-market on either axis, with human review or deterministic outreach |
| Pre-launch checks | Volume, sending domain, personalization, message length, segment fit |
| Sending limit | Google: spam rate below 0.10%, never 0.30% |
| Prices (October 27, 2026) | AiSDR $1.50 per person; Agent Frank $499 a month; Amplemarket $600 a month |
| Build option | Signal, Clay research, model draft, human approval, sender |
The table repeats the sections above. Check it against your own ACV and TAM before you book a vendor demo.
Request the 30-Day Audit
The GTM Engineering Company is a fractional go-to-market engineering partner for seed to Series B B2B SaaS companies. It builds an evergreen data layer and the revenue motions on top of it inside your own CRM, the tools stay in your name, and the playbook stays with you. That suits a team deciding which segments an AI SDR should touch, since the segment field lives in the CRM. Every engagement ends with a written runbook, the "how to keep it humming" guide, as the services page describes (as of October 27, 2026).
If your accounts sit in the mid-market band of Jorge's matrix, a segment-by-segment build usually beats one tool for everyone. Request the 30-day audit and bring your ACV and TAM. The services page describes a GTM Stack Audit and a 30-minute call that leaves you with a shortlist whether or not you hire the company. Starter is $5,000 a month and Growth is $7,000 a month, so you can see the plans before the call.
FAQs About AI SDR
What is an AI SDR?
An AI SDR is software that does the work of a sales development representative: it finds prospects, writes outreach, sends it, handles replies and books meetings. Outbound versions write to people who have not contacted you, and inbound versions such as Qualified's Piper talk to website visitors. Vendors also call them AI sales agents. Jorge Macias of The GTM Engineering Company treats the category as a tool whose fit depends on ACV and TAM.
How much does an AI SDR cost?
AI SDR cost depends on the vendor's pricing unit, and as of October 27, 2026 AiSDR lists $1.50 per person prospected, Salesforge's Agent Frank add-on is $499 a month billed quarterly, and Amplemarket Startup is $600 a month on an annual term. 11x and Artisan show no dollar prices on their pages. Prospecting 1,000 people on AiSDR's pay-as-you-go rate comes to $1,500. Mailboxes, data and review time come on top, and SaaStr reports 15 to 20 hours a week of management across five agents.
What is the best AI SDR tool?
The best AI SDR tool depends on your ACV, your TAM and how much review time you can give it, and no independent test ranks the vendors. Jorge Macias of The GTM Engineering Company says an AI SDR is strongest with a low ACV and a big TAM. AiSDR lists a pay-as-you-go price of $1.50 per person prospected, which suits a small trial, while Amplemarket Startup at $600 a month on an annual term includes 27,000 contacts. Compare the tools in the table above against your own volume.
How does an AI SDR vs human SDR comparison work?
An AI SDR vs human SDR comparison comes down to account value, and Jorge Macias says humans are naturally better at the first impression with high-ACV, small-TAM accounts. SaaStr's agents close event tickets under $1,000 on their own and hand $50,000 to $100,000 sponsorship deals to people. Use the AI for volume segments and a person for named accounts. The mid-market in between is where human-in-the-loop outreach has the most room.
Do AI SDRs actually work?
Do AI SDRs work? They do in a defined slice: SaaStr reports a 6.7% response rate against an industry average of about 3%, and a 4% positive response rate. Those numbers are self-reported and unaudited, and they came with 15 to 20 hours a week of human management. Jorge Macias says a buyer who expects the tool to run itself does not understand sales and marketing. Results track the time put into setup and iteration.
Will AI replace SDRs?
AI will replace SDR tasks before it replaces SDRs, and Jorge Macias of The GTM Engineering Company places the line at account value. Low-ACV, big-TAM work suits an AI SDR, while high-ACV, small-TAM work stays human. SaaStr's agents close sub-$1,000 tickets themselves and hand $50,000 to $100,000 deals to people. No independent study turned up that measures how many SDR roles AI has removed.
Can I build my own AI SDR instead of buying one?
You can build your own AI SDR from a signal source, a research tool, a model and a sender, with a person approving each message. Claude Code's Team plan is $20 per seat per month billed annually, and Clay's free plan lists 500 actions and 100 data credits a month, both as of October 27, 2026. You take on build time and upkeep, which a vendor otherwise carries. The Clay vs Claude Code guide covers the trade-off.
Is an AI SDR worth it if I sell to a small market?
An AI SDR is usually a poor fit for a small market with a high ACV, because Jorge Macias says it burns through the TAM quickly and makes a bad first impression. A person who researches each account and writes by hand protects a small list better. A mid-sized segment can still use human-in-the-loop outreach.
About the Author
Jorge Macias is the founder of The GTM Engineering Company. The company's about page says he spent more than three years at BrainHi as its second sales hire and is an industrial engineer by training.
