An independent, funding-weighted map of the AI SDR and agentic-outbound category as of Q4 2026. This edition replaces our Q2 2026 map, which is preserved as published. It grades that map’s calls, re-scores all four positions, and adds a fifth that the Q2 framework could not accommodate.

The headline is a capital inversion. In Q2 we wrote that money was concentrated in the replace-the-SDR position “despite it being the position with the worst retention signal,” and predicted a re-rating toward augmentation and vertical depth within twelve months. That re-rating has happened, and it is more complete than we expected: every dollar of new capital raised in this category during 2026 went to Positions 2 through 5. Position 1 — the two vendors whose names defined the category — has not raised since November 2024.

The capital ledger, Q4 2026

Position	Named vendors	Disclosed total	Last raise	New capital in 2026
1 — Replace the SDR	11x, Artisan	~$113M combined	Nov 2024	$0
2 — Augment the SDR	Actively AI	$68M	Apr 2026	$45M
3 — Vertical specialist	Monaco	$85M+	May 2026	$85M
4 — Voice for outbound	Bland	~$106M	Jun 2026	$50M
5 — Product-grounded agents (new)	Alta, Sable	$70M combined	Jul 2026	$70M

Figures are disclosed rounds as recorded on our funding tracker. Totals for 11x and Artisan are the ones we can source; both companies’ lifetime totals are reported inconsistently across directories, which is itself a signal about disclosure discipline in this position.

Position 1 — Replace the SDR (11x, Artisan): downgraded

11x raised a $24M Series A led by Benchmark at a reported ~$90M valuation, then roughly $50M in a Series B led by a16z at a reported ~$350M — a round TechCrunch surfaced on September 30, 2024 and the company announced that November. Total funding clusters around $76M across sources. That was twenty-three months ago, and nothing has been announced since.

Artisan raised $12M as a standout of Y Combinator’s W24 batch and a $25M Series A led by Glade Brook in April 2025, for roughly $37M total. Also no round since.

Two things changed for this position in 2026, and both cut the same way. First, Artisan stopped publishing a price list. Its pricing page now routes to a sales conversation, which removes the one concrete advantage our Q2 map credited to it over 11x. Both Position 1 vendors are now quote-only, and third-party trackers put a single AI SDR seat in the $30,000–$60,000 a year range — unverified, because neither company publishes.

Second, and more interesting: Artisan’s own product moved toward Position 2. Ava 2.0, launched in May 2026 per secondary coverage, runs outbound end to end with approval gates. An approval gate is a human pressing send. That is the augmentation thesis, shipped by a company whose brand was built on “Stop hiring humans.” When the flagship product of the replacement position adds a human checkpoint, the position is conceding the argument rather than winning it.

Our Q2 call — “one Position-1 leader exits to a CRM incumbent by Q4 2026” — has not happened. Both CRMs did buy in 2026: HubSpot took Warmly in the signal layer (June 30) and Salesforce agreed to buy Fin in customer support for $3.6B (June 15). Neither bought an AI SDR. The quarter runs to December 31, so this is not formally closed, but we were wrong about what the acquirers wanted.

Position 2 — Augment the SDR (Actively AI): upgraded

Our Q2 map said this position “becomes a feature, not a category,” and that Actively AI would “either win decisively or get absorbed into a Clay-class platform.” Neither happened. Actively AI raised a $45M Series B at a $250M valuation on April 28, 2026 (TCV with First Harmonic), taking it to $68M total. A position we described as structurally fragile attracted the category’s second-largest round of the year.

The fragility argument was that the augmentation surface is also what a Clay-plus-Smartlead stack does competently. That remains true, and it is why this position will never produce a $10B company. But it mis-read the buyer: augmentation sells against a budget line that already exists (rep productivity), while replacement sells against a headcount decision that requires someone to be fired to validate the purchase. In a year when AI budgets came under scrutiny, the easier sale won. We should have weighted procurement friction more heavily than product ceiling.

Position 3 — Vertical specialists (Monaco): upgraded sharply, and our estimate was badly wrong

This was the Q2 map’s best structural call and its worst number. We wrote that vertical depth “is the one moat the foundation models can’t trivially erode,” which has held. We also put Monaco at “~$8M seed (disclosed)” and projected the whole-market outcome for a vertical player at “$30–80M ARR — meaningful but not a category winner.”

Monaco has since raised a $35M Series A (February 15, 2026) and a $50M Series B (May 12, 2026), for $85M+ total. A single vertical specialist is now funded at more than ten times the figure we used to characterise the position. Our contrarian read — that the 2027 category winner “has not yet raised a Series B, ships in a single vertical, and is invisible to the average tech-Twitter feed today” — is now half-wrong in the most useful way: the vertical bet is being validated, but the company doing it raised its Series B eight months after we called it invisible.

Position 4 — Voice for outbound (Bland): upgraded, and the regulatory trigger never came

Bland raised a $50M Series C on June 16, 2026, led by Dell Technologies Capital with HubSpot Ventures, Archerman and Tribeca joining existing backers Emergence, Upfront, Scale and Y Combinator — taking total funding past $100M. The company reports 250+ enterprise customers and 3.5 million calls a week (company-stated), and it builds its own voice models rather than wrapping a frontier model, which is the “voice-infrastructure company, not LLM wrapper” shape our Q2 map said would win here.

A disclosure: we found this round while re-scoring this map. It was not on our tracker, which means our Q2 2026 quarterly figures understated the category. Both are now corrected — Q2 2026 reads seven rounds and $1,355M, and the Q3 report carries a correction note.

Our Q2 threat call here was that “one TCPA enforcement action against an AI-dialed outbound campaign reprices the entire sub-category overnight,” and that outbound voice would become “a compliance product, not a growth product.” No enforcement action materialised in the window. What did happen is that Bland raised at a premium and explicitly directed the capital at regulated industries — healthcare and financial services. Read carefully, that is the compliance-product thesis being funded rather than punished. The trigger we named is still loaded; the direction we inferred from it arrived without it.

Position 5 — Product-grounded agents (Alta, Sable): new

Two rounds in July 2026 describe a shape the Q2 framework had nowhere to put. Alta raised a $25M Series A (July 8) for a “Company Brain” that coordinates agents from prospecting through AI calling on top of the CRM you already run. Sable raised $45M (July 16, co-led by Sequoia and 8VC, no stage label in the primary source) for an AI sales engineer that runs live product demos by seeing and clicking the product in real time.

What unites them is the grounding. Positions 1 and 2 are bounded by the prospect list — the agent’s quality depends on research about someone it has never met. Alta is bounded by the customer’s own CRM and Sable by the customer’s own product. Both replace open-ended inference with a closed system the agent can actually read. That is the single most reliable pattern we have observed across this category: the agents that work are the ones whose job has edges. If the 2027 winner comes from outside Position 3, we expect it from here.

Threats, re-scored

1. Email deliverability collapse — still live, still unrealised. We called this “the single highest-probability sub-category extinction event in the next 18 months.” No Gmail or Microsoft policy change in the window produced the reply-rate cliff we described. We are not withdrawing it; we are noting that six months of the eighteen have passed without it.

2. Voice regulation — unrealised, see Position 4.

3. AI-SDR fatigue — unscoreable as written. Our Q2 map asserted reply rates falling from 4–6% in 2024 to sub-2% in Q1 2026. Those were our own qualitative estimates from operator conversations, not measured data, and there is no mechanism by which a reader could verify or falsify them. We are flagging it rather than repeating it, and future editions will not publish a rate we cannot source.

Buyer decision tree, Q4 2026

Unchanged in structure, with one substitution and one addition. Removing headcount? Position 1 — and note that neither vendor has raised in two years, so price the renewal risk. More from existing headcount? Position 2. Regulated domain? Position 3, which is now genuinely capitalised. ICP unreachable by email? Position 4. Want the agent bounded by a system rather than a list? Position 5 — the newest and least proven, but the most structurally sound. None of the above, with a competent RevOps team? Still don’t buy an AI SDR; build the equivalent on Clay plus a sender plus a frontier model.

Predictions for 2027, with resolution criteria

1. Neither 11x nor Artisan announces a priced round by June 30, 2027. Resolves: a dated funding announcement from either. Two years of silence is a pattern, not a gap.

2. A Position 5 vendor (Alta or Sable) raises again at a higher mark before either Position 1 vendor does. Resolves: dated announcements, compared.

3. At least one more vertical AI-SDR company raises $25M+ by Q2 2027. Resolves: our tracker. Monaco’s $85M makes the vertical thesis fundable; capital follows proof.

4. No AI SDR vendor publishes a per-seat or per-outcome price list by Q2 2027. Resolves: a public pricing page. Artisan withdrawing its list was a move away from transparency, not toward it.

Methodology: funding figures are disclosed rounds as recorded on the GTMLens funding tracker, which requires an exact date plus either a primary source or two independent sources. Where a lifetime total is reported inconsistently across directories we say so rather than pick one. Position assignments reflect each company’s primary GTM message and product surface, not internal roadmaps. Customer counts, call volumes and revenue figures attributed to a company are company-stated and not independently verified. Retention characterisations in our Q2 edition were qualitative analyst priors and are not repeated here as data. GTMLens accepts no vendor money, paid placements or affiliate commissions — see our editorial policy.