The Account Executive You Want Now Directs the AI and Reads the Room

What Does an Account Executive Do When AI Drafts Every Email?

Monday, your rep's sequence goes out at six without her touching it, and two of the eleven replies say the same thing: good note, did you write it? The account executive still closes. The work around closing, account research, meeting prep, CRM updates, forecasting and follow-up, has moved into software ³.

What remains is tacit and hard to fake. Reading which of five people on a buying committee is quietly against you. Knowing when a discount buys nothing. Holding a point of view about the buyer's business that survives being argued with by someone who runs that business.

The shift is not confined to sales. BCG's 2026 survey of 11,749 workers found 47% now spend more time managing and directing AI than doing the work itself, and 72% say AI has considerably changed the skills their role expects ². PwC's barometer points the same direction on pay: roles where AI magnifies expert judgment show roughly twice the job growth and 42% faster salary increases than roles where AI flattens the skill premium ¹.

The tells that separate a real AE from a performed one are specific. A real one narrates a lost deal by naming the moment the economic buyer disengaged, and can say what they would test differently. A performed one narrates the same loss as a procurement problem. Ask what the last buyer taught them about their own product; the answer is either a concrete correction to their pitch, or a compliment.

How Do You Screen an AE's AI Practice Without a Mock Cold Call?

Give them a real account and forty minutes with an assistant, in the open. Hand over a public company you sell into, your one-page product summary, and the ask: a point of view on why this account would buy, three risks to the deal, and a first-call plan. Then read how the work happened, not just what came back.

The session shows you things a panel cannot. Did they frame the account before they generated anything, or start typing at the model. When the assistant asserted a funding round or a reorg, did they go check it, and did they say where they checked. When it produced a confident and generic paragraph about digital transformation, did they cut it. Sellers who work this way narrate the cut out loud: this part is wrong about how procurement works here, so it goes. Your Monday rep would do this in ten minutes, which is worth knowing, because the exercise is only useful if it is calibrated against someone you already trust.

Two failure modes show up repeatedly and they look nothing alike. One rep accepts everything the assistant wrote and polishes the tone, which is the habit that puts a fabricated detail in front of a buyer. Another refuses the tool outright and does it all by hand in forty minutes, producing less and calling it rigor. Neither is the person managing the work rather than doing it that the BCG numbers describe ².

Run the same exercise for every candidate and write down what you saw at each decision point. A structured record of moments beats a panel's impressions, and it also survives the scrutiny an AI transformation consultant or your own counsel will apply to a hiring process later.

Recruiters and Domain Switchers Multithread Better Than Career AEs

The reliable feeders are still enterprise sales, solutions consulting, and customer success with an expansion quota. The unexpected ones are worth more right now: contingency recruiters, agency account directors, ex-founders who sold their own first fifty deals, and specialists who left a domain, a nurse, a claims adjuster, a plant supervisor, to sell into it.

Each of those backgrounds trains the same muscle from a different angle. A contingency recruiter has spent years multithreading between a hiring manager and a candidate who both have veto power, which is a buying committee under another name. An agency account director has defended a strategy to a client who paid for it. A domain switcher can ask a question that makes a buyer sit forward, and no model has that question because it never sat in that room.

What none of those backgrounds guarantees is the AI practice, and the practice is a separate thing you have to look for. The sellers who got good at it built habits, not prompt collections. They keep the drafting and the deciding apart: the model writes the recap, the rep decides what the recap admits. They force a source out of the assistant on any claim that will appear in front of a customer, because a hallucinated funding round in a first email ends the account. They keep a running file on each account and feed it in rather than starting cold every week. And they can tell you what they stopped delegating after it burned them once.

That practice looks different from the technical side of the same shift. If the person you actually need builds the pipeline systems instead of working them, read the GTM engineer hire before you write the AE job description, because those two roles get conflated in job postings constantly.

Read the Deal Teardown Before You Read the Resume

Not on job boards, mostly. The sellers worth hiring are employed and hitting quota, so they surface in the places where sellers argue about the craft: Pavilion, RevGenius, Bravado, the Sales Enablement Society, and the practitioner tracks at SaaStr Annual. Referrals from your own best rep still convert better than any of it.

Sourcing signal beats sourcing volume here. A rep who posts a teardown of their own lost deal is telling you more than a resume can. So is one answering technical questions in a Gong or Outreach user community, or writing the internal enablement doc their team actually uses. Those artifacts are cheap for you to read and expensive for a candidate to fake, which is the only ratio that matters in sourcing.

Adjacent roles are the underpriced pool. Solutions consultants who ran discovery and want a quota. Customer success managers who have been quietly closing expansions with no commission on them. Support engineers at a technical vendor who know the product better than the sales team does. Partner managers who have multithreaded through two organizations at once for years.

Feeder companies matter less than feeder motions. A rep from an org that ran real MEDDIC-style qualification and lost deals to committees will transfer; a rep who carried an inbound-only book at a category leader often will not. Ask what percentage of their pipeline they sourced themselves, then ask what they did in the quarter that number dropped.

What Should You Pay, and Who Actually Hit the Number Last Year?

As of mid-2026, one self-reported aggregator, levels.fyi, puts account executive median total compensation at $220,000, with the 25th percentile at $148,000 and the 75th at $315,000 ⁴. Read that as a tech-weighted band rather than a national wage series: it skews toward software companies and toward the reps who file their own numbers.

No published series tracks the AI-augmented version of the title separately yet, so treat any premium as a bet rather than a benchmark. The directional evidence is PwC's, which finds faster salary growth in roles where AI magnifies expert judgment than in roles where it levels the field ¹. Enterprise AE pay stays roughly a 50/50 base-to-variable split, and the honest question a strong candidate asks is not the on-target number but whether anyone hit it last year.

The work is hybrid by gravity. Committee-heavy enterprise deals pull people onto planes and into rooms, and most teams settle on remote-based with heavy travel, or an office anchor of two days for pipeline reviews and deal strategy. A fully remote AE role closes fine when the territory is remote too; it strains when the buyer expects on-site and the rep does not want to travel.

Remember that your Monday rep is already inside somebody's band, and three companies are calling her this quarter. What closes them: a territory with named accounts rather than a vague region, a comp plan that pays on multi-year deals in the year they land, a product they can hold a real opinion about, and a manager who runs deal strategy instead of forecast interrogation. What kills the offer: a quota nobody has ever hit, a variable plan that changed twice last year, and a slow process. The best sellers have three conversations running. If you need six weeks to decide, you are hiring from whoever is left, which is the same failure the revenue AI systems architect search runs into.

See the benchmarks >