How AI-Native Companies Are Scaling on a Direct Enterprise Sales Motion

Delivered to you by Primer – Primer unlocks B2C advert channels for B2B entrepreneurs with precision focusing on and next-gen measurement

There are two methods you’re losing advert finances: (a) you pay to achieve individuals who won’t ever purchase, and (b) the consumers who do present up stroll away unidentified. Primer fixes each. This consists of excluding the incorrect folks throughout Meta and Google earlier than you pay for the impression. No extra burning finances on college students, opponents, and tire-kickers.

You get limitless web site reveals: each account researching you, recognized – no caps, no per-reveal pricing. See who’s really in-market and feed it straight again into focusing on.

GTMnow Community unique provide: Prolonged 45 day trial and 15% off with code GTM15 – simply apply the code if you go to improve from the free trial.

See who’s actually in your funnel: sayprimer.com

How AI-Native Firms Are Scaling on a Direct Enterprise Gross sales Movement

The PLG story has owned the AI development narrative for 2 years, and it deserved to. However the firms quietly posting the identical numbers (or extra) are doing the other.

Legora went from $1M to $100M in ARR in about 18 months, promoting to probably the most procurement-heavy consumers on earth, international regulation corporations like White & Case and Linklaters.

Sierra hit a $100M ARR run charge in 21 months and now works with 40% of the Fortune 50, with year-one contracts working into the a whole lot of hundreds.

Decagon tripled income 12 months over 12 months to roughly $35M by October 2025, including greater than 100 enterprise logos throughout airways, banks, and telecom.

None of those is a self-serve product you’ll be able to join at your leisure. Each firm has a gross sales rep in each deal, a six-figure ACV, and a purchaser who didn’t get full product entry earlier than a contract existed. They’re rising at PLG pace on a sales-led movement.

Right here is how they’re doing it and how one can too.

The quick risers aren’t all product-led

The default rationalization for AI’s development information is bottom-up adoption. Cursor earned that repute, reaching $100M ARR in twelve months and $2B+ by February 2026 with zero outbound gross sales till late 2025, just because builders couldn’t cease pulling it into their firms. Enterprise income climbed from roughly 25% of the combo to just about 60% as the primary reps arrived previous $200M ARR to formalize demand that already existed.

That movement works when one individual can undertake the product alone and get worth in minutes. It breaks the second the client is a common counsel, a head of CX, or a CIO who won’t ever personally attempt the instrument, and whose firm wants the factor wired right into a CRM, a billing system, and a compliance evaluation earlier than it does something helpful.

So a unique set of firms skipped the PLG flywheel fully.

They promote on to the financial purchaser, they usually develop simply as quick. The pace is coming from a gross sales movement constructed particularly for AI.

The 2 AI-native sales-led firms reached $100M ARR as quick because the PLG document holder.

The movement: gross sales lands it, an engineer makes it actual

The standard enterprise playbook is linear: construct the product, hand it to gross sales, gross sales closes, buyer success onboards, and the client figures it out. That works for a CRM with a recognized job to be carried out. It collapses for AI as a result of the hole between a fantastic demo and a working deployment is the place most AI pilots crumble. MIT’s 2025 analysis put the failure charge of enterprise AI pilots at 95%, and a lot of the failure was implementation, not the mannequin.

The businesses profitable closed that hole by shifting an engineer into the room. The function has a reputation now, the forward-deployed engineer, and it was Palantir’s invention from the early 2010s, when its intelligence company prospects couldn’t even describe what they wanted. Palantir’s reply was to embed engineers contained in the buyer to construct in actual time. That mannequin was mocked for a decade as costly consulting carrying a software program costume, proper up till the inventory returned roughly 640% over 5 years.

Nearly each enterprise AI pilot dies on the integration wall. The FDE is who will get it previous.

What a Ahead Deployed Engineer (FDE) is

An FDE is a senior engineer who embeds with the client for the size of a deployment, writes manufacturing code within the buyer’s atmosphere, runs technical discovery, and ships the mixing that makes the agent work. It’s not a renamed options engineer. A options engineer sells the product and palms it off pre-contract. An FDE, then again, ships the product into manufacturing post-contract. One closes the demo, the opposite closes the hole between the demo and dwell.

That is now the usual org sample, not a fringe experiment. Decagon assigns devoted Agent Product Managers and forward-deployed engineers who embed with every enterprise buyer, taking deployments from discovery to dwell in roughly six weeks. The frontier labs run the identical play: Anthropic’s $1.5B Deloitte three way partnership and OpenAI’s $10B Bain alliance are, beneath the framing, FDE staffing pacts. When the mannequin makers themselves resolve deployment wants embedded engineers, it’s an excellent signal for application-layer startups constructing on high of them ought to take the trace.

The embedded engineer is the load-bearing rent, and the market is paying for it.

PLG lets demand construct, then formalizes it. The FDE movement manufactures demand contained in the account.

Why the FDE movement compounds as an alternative of simply costing extra

The apparent objection is the margin. Embedding engineers in each account appears like a companies enterprise with a software program emblem, and for fifteen years, that was the consensus learn on Palantir. The rationale it isn’t a consulting entice is a suggestions loop, Palantir named it the gravel highway to paved freeway.

An FDE builds a quick, customized workflow for one buyer, a gravel highway. The core engineering staff then research the gravel roads throughout many shoppers, finds the repeated sample, and turns it right into a product – a paved freeway each future buyer will get out of the field. Every customized construct makes the following deployment cheaper, so the movement will get extra environment friendly at scale reasonably than much less. So the companies’ work isn’t the price of the mannequin, it’s the R&D enter that feeds it.

Decagon productized this immediately. Its Agent Working Procedures let CX groups outline workflows in pure language that compile into code, which is precisely the gravel-road work an FDE used to do by hand, paved right into a characteristic prospects run themselves. Sierra did the identical with Ghostwriter, an agent that builds brokers from a buyer’s SOPs and recordings, explicitly constructed to scale back the on-site engineering every new deployment wants. The customized work of the early offers turns into the product floor of the later ones.

The economics again it up. A forward-deployed engineer runs roughly $250K to $450K all-in at Sequence A, which is lower than the totally loaded value of an enterprise AE plus a options engineer, and the FDE produces compounding development. The compounding curve and the sales-only curve diverge round month six.

The loop that turns embedded fieldwork right into a product that closes the following deal quicker.

Pricing is doing the promoting

A direct AI sale has one objection above all others: will this really work on our information, or are we paying six figures for a demo that doesn’t survive contact with actuality? The quickest firms reply it with the contract itself.

Sierra costs on outcomes. As founder Bret Taylor, the previous Salesforce co-CEO now working Sierra, describes it: the client pays a pre-negotiated charge when the AI agent resolves a difficulty autonomously and pays nothing when it escalates to a human. He frames it because the pure subsequent step in software program’s evolution, from perpetual licenses to subscriptions to paying for a job properly carried out.

Decagon gives per-conversation and per-resolution pricing on the identical logic. The impact is similar in each circumstances: the seller solely makes cash when the product works, so the client’s danger drops and the deal strikes.

Pricing on outcomes shifts adoption danger off the client, and it has constructed a $15.8B firm.

PLG made software program promote itself, and that story isn’t incorrect. However it influenced a notion {that a} direct gross sales movement was the gradual, previous solution to develop. Legora, Sierra, and Decagon are the counterexamples. Constructed proper, direct enterprise gross sales transfer as quick as any flywheel.

Tag @GTMnow so we will see your takeaways and assist amplify them.

SpaceX formally agreed to amass Cursor for $60 billion in inventory, days after its document Nasdaq IPO.

Snap unveiled SPECS, their new augmented actuality glasses. Absolutely standalone, 132 grams, 51-degree discipline of view and obtainable for pre-order at $2,195 this fall. Then Meta mentioned it’s launching a brand new line of good glasses, dubbed Meta Glasses, beginning at $299.

VC: “We Don’t Fund Good Firms” : A $1.5B VC Explains Why | Ben Lerer, Lerer Hippeau

GTM: The right way to Use Gross sales Comp as a GTM Lever | Siva Rajamani (Everstage CEO)

GTM: The right way to Win at Paid Promoting within the AI Period | Keith Putnam-Delaney, Primer CEO

VC: How Benchmark Invests & What to Know on GTM | Chetan Puttagunta (GP)

Pay attention by means of the hyperlinks within the web page above or by looking wherever you get your podcasts “The GTMnow Podcast.”

Prometheus – raised $12B at a $41B valuation. That is Jeff Bezos’s bodily AI startup constructing software program to automate the design and manufacturing of advanced bodily methods, from jet engines to drug compounds.

Newton – launched mainnet beta – an onchain authorization layer that stops transactions earlier than they settle in the event that they break coverage. Crypto solved settlement and skipped authorization. Newton is constructing that lacking layer.

Mutiny – launched an AI agent that takes accounts from chilly to closed – one platform changing ABM, gross sales enablement, and AI gross sales instruments. Connects to your CRM, calls, and electronic mail to automate the client experiences that truly win offers. 3,000 firms together with Rippling, Snowflake, and Uber are already on it.

Consideration – raised $30M Sequence B led by RTP International. They’re the AI platform that acts on gross sales calls, not simply information them. Most instruments inform you what occurred. Consideration drafts the follow-up, updates the CRM, and runs the following play.

Redo – raised $81M at a $1.25B valuation led by Smash Capital, Pelion Enterprise Companions, and Cervin Ventures. To have a good time, they’re giving customers $100 to spend at Redo manufacturers as an alternative of working a conventional announcement.

Runlayer – raised $30M from Felicis and Khosla Ventures to assist firms say sure to brokers with out dropping management. Id-aware permissions, full observability, runtime safety – one platform. Clients embody Instacart, Gusto, Decagon, and dbt Labs.

Coval – raised $28M Sequence A led by Norwest VP to construct the simulation and analysis platform for voice and chat brokers. The identical method that made autonomous autos reliable can do the identical for conversational AI.

Scaled Cognition – raised $100M Sequence A led by Khosla Ventures – AI fashions constructed to inform the reality, not simply sound assured. Enterprise hallucination charges are 5x what firms suppose. They rebuilt the structure from first ideas so reliability is engineered in, not bolted on.

Common Instinct – raised $320M at a $2.3B valuation led by Khosla Ventures to construct AI brokers that generalize from gameplay to the true world. Billions of action-labeled gameplay clips from 17M customers – not simply video, however the precise buttons pressed and when. The identical mannequin taking part in Fortnite is powering their quadruped robotic.

Airwallex – raised $320M at an $11B valuation led by Addition to construct the clever layer on international monetary rails. Two new merchandise: T:0, an AI-native finance platform, and Airi, an agentic pockets for when software program brokers transact in your behalf.

Account Govt, SMB at Closinglock (Austin, TX)
Enterprise Account Govt, APAC (Sydney, Australia)
Account Govt, SLED Authorities at Vanta (Distant – US)
Dealer Companion Supervisor (Dealer Improvement Consultant) at TrustLayer (Hybrid – Tampa / St. Pete, FL)
Options Engineer (Pre-Gross sales) at Noibu (Hybrid – Ottawa, Ontario)