Get started

Taking an assessment rather than buying one? This page is written for employers. Here is the page for candidates.

High-velocity and transactional sales · Mid level

How to assess a SMB Account Executive

The mock demo is the step that looks most like the job and predicts it least, for a precise reason: it is scored on product knowledge, which a new hire does not have and will acquire in a fortnight, and on polish, which the job needs far less than selectivity. In a real SMB demo the buyer has named two or three problems in a short discovery and has very little attention left; the entire skill is choosing what not to show. A candidate who has memorised the feature list will show all of it and score highly on a knowledge-weighted mock, then feature-dump their way through a real quarter. The second gap is that this cycle closes in one or two conversations, so the written recap and proposal between them carry weight a longer cycle spreads over months — and almost no process reads a candidate's writing before hiring them. The third is the comparison question, which arrives in nearly every SMB deal: the narrow honest answer and the reflexive disparaging one are indistinguishable in an interview, where nobody is sitting opposite the candidate visibly preferring the other product.

An SMB account executive runs a compressed version of a full sales cycle, several times a week. A lead arrives, they run a short discovery call, they demo, they send something in writing, and they close or lose within a fortnight. The distinguishing artefact of the seat — the thing no sibling role in this hub has — is the demo, and it is worth being clear that a demo is not a pitch and not a discovery call. It is a live, partly improvised act of selection performed under a countdown, in which the seller has to decide in real time which three of the many things it could show, in what order, and when to stop and ask a question instead.

That selection is where the top quartile separates. The median AE has been trained on a demo script and follows it, which produces a competent, forgettable session in which the buyer is shown everything and remembers nothing. A strong AE opens by restating the two problems the buyer named, shows the path through the product that addresses exactly those, and deliberately declines to show the rest — often saying so out loud, which is disarming and builds more credibility than any feature could. The second separator is the handling of the competitive comparison. SMB buyers are almost always evaluating something else, often something cheaper, and they ask directly. The answer that works is narrow and specific — one honest dimension on which the difference matters for the problem the buyer described — and the answer that loses deals is a general claim of superiority or, worse, a dig at the alternative, which makes the buyer defend a choice they were previously uncertain about.

The third is discipline about the next step. In a two-week cycle, a deal that ends with "let me think about it and I'll come back to you" is usually lost, and the difference between that and "shall we speak Thursday at 10 with your finance lead on the call?" is a single sentence that many otherwise capable sellers cannot bring themselves to say. Related: the SMB buyer who is enthusiastic and cannot sign is the standard way these deals die, and identifying that in discovery rather than in week three is worth more than any amount of product fluency.

What a hiring manager is really trying to predict is cycle-time-adjusted throughput — how many of these compressed arcs the person can run well in a week — and whether their forecast for a two-week cycle can be believed on a Monday. Ramp is the pressure behind that: the 2026 Bridge Group cohort reports a 6.2 month median ramp across AE seats, and a velocity segment cannot afford to discover a mis-hire two quarters in.

What the job actually needs

How people fail in this seat

What most employers do instead

CV and quota history, a competency interview, and a mock pitch or "demo our product back to us" exercise scored largely on polish and product knowledge.

The mock demo is the step that looks most like the job and predicts it least, for a precise reason: it is scored on product knowledge, which a new hire does not have and will acquire in a fortnight, and on polish, which the job needs far less than selectivity. In a real SMB demo the buyer has named two or three problems in a short discovery and has very little attention left; the entire skill is choosing what not to show. A candidate who has memorised the feature list will show all of it and score highly on a knowledge-weighted mock, then feature-dump their way through a real quarter. The second gap is that this cycle closes in one or two conversations, so the written recap and proposal between them carry weight a longer cycle spreads over months — and almost no process reads a candidate's writing before hiring them. The third is the comparison question, which arrives in nearly every SMB deal: the narrow honest answer and the reflexive disparaging one are indistinguishable in an interview, where nobody is sitting opposite the candidate visibly preferring the other product.

The assessment

About 38 minutes end to end.

The systems it runs in

A CRM deal on a stage pipeline, with contact roles for the buying group, a quote built from the price list on the product sheet, and an e-signature step at the end of it. The authority fork in t1 is scored partly off the record: Priya either exists on the deal as a contact with a role and a reason, or the buying group on this deal is one enthusiastic operations lead who does not control the budget. t3 is composed and sent from the deal record as the follow-up the buyer forwards, with the quote attached at the price the candidate actually named, and the candidate has to move the stage and set a dated next step before the session closes.

Any CRM with a deal or opportunity object, stages, and somewhere to record who else is involved in the decision — contact roles, a buying-group field, or a stakeholder list on the record. Quoting and e-signature can be part of the CRM or bolted beside it; the fixture only needs the price to end up somewhere the buyer can forward. Rebuilt against the buyer's own stage names and quote template where they can provide a sandbox, because the stage definitions are where an SMB sales team's actual qualification standard is written down.

Working speed is not scored. Not scored, and the two forks explain why. The fastest route through t1 is to accept "I'm basically the one who decides here" and book the demo; the fastest route through t2 is to accept "let me have a think and I'll come back to you" and end early; and both save minutes while producing exactly the pipeline this seat is hired to stop producing. A cycle measured in weeks does not reward a rep who got off the call four minutes sooner. Time spent is logged and available to a reviewer, but it is not a criterion and it is not shown as a headline.

What the candidate actually does

TaskWhat happens
Discovery, and who actually signs
live_call · 8 min		Six unscored minutes beforehand with a one-page sheet describing a fictional product — eleven capabilities, a price list, and two things it deliberately does not do. Product knowledge is therefore given, not tested. The AI buyer is an operations lead at a forty-person company who has booked the call themselves. They name two concrete problems if asked open questions, and will happily spend the whole eight minutes on a third, shinier topic that the product handles well and that is not costing them anything. The fork is authority, and it is quiet: asked about process, the buyer says they would "run the number past Priya, but I'm basically the one who decides here." They will not elaborate unless pushed, and the truth is that Priya sets the budget and has not agreed to spend anything. The buyer is enthusiastic and will agree to a demo either way.
The demo, and what you refuse to show
live_call · 12 min		A narrated walkthrough against the same product sheet — the candidate says what they would show, in what order, and why, while the AI buyer interrupts. No live software is required, which is the point: the seat's skill is selection, not clicking. Two forks. Early on the buyer asks to "just see everything, I've got the time" — a warm, cooperative request that leads directly to a twelve-minute feature tour, a pleasant call, and a buyer who remembers nothing. Around minute eight the buyer names a cheaper competitor and says it looks "basically the same for half the price", then waits. The AI is configured to disengage slightly if the candidate disparages the alternative or claims general superiority, and to lean in if given one specific dimension tied to a problem they themselves named. Late in the call, if no next step is proposed, the buyer offers "let me have a think and I'll come back to you", which is available to be accepted.
The recap that gets forwarded to Priya
written_artifact · 8 min		The written follow-up, composed without the transcript, and explicitly for onward forwarding to the person who controls the budget and was not on either call. It must restate the two problems in the buyer's own terms, state price and what it covers, state what was not promised, and propose a dated next step with the right people on it. It is composed and sent from the deal record, and the deal has to be left in a state another person could pick up: the stage moved or deliberately not, a dated next step, and Priya either attached to the deal as the person who controls the budget or conspicuously absent from it. In a two-week cycle this document does the work a longer cycle spreads over months of meetings, and almost no hiring process reads a candidate's writing before the offer.

The mark scheme

Each criterion is scored 1 to 5 against written anchors, and every score is reported with the excerpt that earned it. A criterion marked floored is reported as a finding rather than averaged into the total. The first is open; open any other to read its anchors in full.

Discovery finds the problems and finds who signsweight 0.2Elicits both problems with open questions, and follows the Priya mention until it is clear who sets the budget, what they would need to see, and wheth…
1 Spends most of discovery on the shiny third topic, or accepts "I'm basically the one who decides" without a follow-up question.
3 Elicits both problems, and establishes that another person is involved without establishing what that person controls.
5 Elicits both problems with open questions, and follows the Priya mention until it is clear who sets the budget, what they would need to see, and whether anything has been approved.
Shows three things and declines the rest, out loudweight 0.3Opens by restating the two problems, shows only the path through the product that addresses them, and says explicitly that they are not showing the re…
1 Accepts the invitation to show everything and walks the capability list; the buyer's two named problems are not the organising principle.
3 Covers the relevant capabilities plus several others, in an order the product sheet suggests rather than the buyer's problems.
5 Opens by restating the two problems, shows only the path through the product that addresses them, and says explicitly that they are not showing the rest and why — declining the buyer's own invitation to see everything.
Answers the competitor comparison narrowly and honestlyweight 0.2Names one dimension where the difference matters for a problem the buyer described, concedes plainly where the cheaper option is genuinely adequate, a…
1 Disparages the alternative, or asserts general superiority, or claims a capability the product sheet says the product does not have.
3 Lists several differentiators without connecting any of them to the problems this buyer named.
5 Names one dimension where the difference matters for a problem the buyer described, concedes plainly where the cheaper option is genuinely adequate, and leaves the buyer's earlier judgment intact.
Agrees a next step with a date and the right people on itweight 0.15Proposes a specific date and time, names who needs to be on it including the budget holder, and states what will be decided — and does not retreat whe…
1 Accepts "let me have a think and I'll come back to you", or ends with no next step at all.
3 Proposes a follow-up with a date but no named attendees or purpose.
5 Proposes a specific date and time, names who needs to be on it including the budget holder, and states what will be decided — and does not retreat when the buyer hesitates once.
A recap that survives being forwardedweight 0.15Opens with the two problems in the buyer's own terms, states price and what it covers, states what was explicitly not promised, and closes with the da…
1 Restates the price list, or claims something the calls did not contain.
3 Accurate summary of the demo, written to the person who attended it.
5 Opens with the two problems in the buyer's own terms, states price and what it covers, states what was explicitly not promised, and closes with the dated next step and its attendees — readable and actionable by someone who was on neither call.

How it is scored

Weighted mean of five criteria, 1-5 against the anchors, with excerpts. Two counts are reported beside the score, not inside it: how many of the product's eleven capabilities the candidate showed, and whether the two problems the buyer named were restated before the walkthrough began. A high capability count with a strong overall score is a sign the rubric was applied loosely and should be re-reviewed, because the whole thesis of this design is that showing less is the skill.

Integrity

The log describes what happened. It does not produce a cheating verdict — the follow-up conversation is the control, because a statistical accusation is not something we would ask a reviewer to defend.

What you receive

Who decides

Required before any offer. This is a mid-level seat with a long ramp, so the review is the most substantial in the hub and still costs under six minutes: the reviewer reads the recap in full, then listens to the first ninety seconds of the demo and the two minutes following the competitor question. Before seeing the score they answer two written questions — what did this candidate choose not to show, and would you forward this recap to your own CFO. A candidate scoring 5 on selectivity while the capability count is high means one of the two is wrong and the case goes to a second reviewer. Reviewers who are themselves strong demo-ers tend to reward polish; brief them that polish is not on this rubric anywhere.

What this does not measure

Nothing in this design scores accent, dialect, first language, speech rate, disfluency, vocabulary range, or the cluster usually labelled executive presence, gravitas, polish or charisma — the last group is the specific reason the incumbent mock-demo step is unfair as well as unpredictive, since it rewards a performance style rather than a commercial decision. Every anchor is decidable from transcript text. Two design choices exist to protect this. First, product knowledge is supplied on a sheet and re-seeded per candidate, so nobody is advantaged by having sold a similar product before, and nobody is penalised for lacking the vocabulary of one industry. Second, the demo is narrated rather than clicked, which removes screen-sharing, home broadband and equipment quality from the assessment entirely — a real source of impact when candidates demo from wherever they live. Risks to monitor: the writing task carries fifteen percent and should be marked on content and structure only, never on idiom, register or grammar; and the next-step criterion rewards direct asking, a trained and learnable behaviour that should be disclosed to candidates in advance along with the rest of the rubric dimensions. Do not derive any part of the score from voice, sentiment or video analysis, and do not use video at all for scoring. Publish before the assessment opens how to request extra preparation time, a break between the two calls, or a written variant. Track scores per criterion and by cohort; the selectivity criterion is the one that carries most of the weight and is therefore the one to watch.

The mock demo is the step in this seat's hiring process that looks most like the job and predicts it least, and the reason is specific enough to design against. It is scored on product knowledge, which a new hire does not have and will acquire in a fortnight, and on polish, which the job needs far less than selectivity. So this design gives the product away — a one-page sheet, six unscored minutes to read it, eleven capabilities and two honest limitations — and then makes the entire demo task about what the candidate refuses to show.

The fork in t2 is written to be genuinely tempting rather than obviously wrong. The buyer does not merely permit a feature tour; they ask for one, warmly, and say they have the time. Complying is cooperative, is what the customer requested, fills twelve minutes comfortably, and produces a call that feels good to everyone in it. Declining it — restating the two problems, showing the three things that touch them, and saying out loud that the rest is not relevant today — is the behaviour the role file identifies as separating the top quartile, and it requires the candidate to contradict a friendly buyer with no product tenure to fall back on. The capability count reported beside the score exists so that a hiring manager cannot accidentally reward the tour, which is what a knowledge-weighted rubric does automatically.

The competitor question is the second fork and it is the one that is invisible in an interview, because in an interview nobody is sitting opposite the candidate visibly preferring the cheaper product. Here the buyer says the alternative looks basically the same for half the price and then waits, which is the shape of the real moment. Configuring the AI to disengage on disparagement and to lean in on a narrow honest answer means the branch itself carries information: the transcript shows not only what the candidate said but what it did to the buyer. Conceding where the competitor is adequate is written into the top anchor deliberately, because the concession is what makes the one honest differentiator credible.

The authority fork in t1 is the third distinct thing this design tests, and it belongs here rather than in any sibling role because the SMB cycle is short enough that discovering the missing budget holder in week three is fatal. The mention of Priya is a single clause, offered once, immediately followed by a reassurance that the buyer is the decision-maker. Accepting the reassurance is the default and produces a perfectly pleasant call. The candidate who keeps pulling on it — what does Priya control, what would she need to see, has anything been approved — is doing the thing that makes a two-week forecast believable on a Monday.

The written recap is where those three threads are collected, and framing it as a document for onward forwarding to the absent budget holder changes what a good answer looks like. It cannot assume shared context, it has to restate the problems in the buyer's language rather than the seller's, and it has to include what was not promised. In a cycle that closes in one or two conversations, this document carries weight a nine-month cycle spreads across months of meetings, and almost no process reads a candidate's writing before the offer.

Thirty-eight minutes is the largest budget in the sales hub — eight discovery, twelve demo, eight writing, six unscored product reading, four setup — and it is justified by the arithmetic of this seat rather than by ambition. This is the lowest-volume role of the six and the one with the longest ramp, so the cost of discovering a mis-hire is measured in quarters. It is still under forty minutes because the assessment has to fit in front of every applicant to a velocity segment that hires continuously, not just the three who reached a final round.

Sources

Every figure on this page is traceable. Where a claim could not be sourced it is stated qualitatively instead.

  1. The Bridge Group, AE Models, Motions & Metrics 2026 Research Report, 10th edition (158 B2B companies; data collected Q1-Q2 2026): 48 percent of reps at quota, median quota $960,000, 6.2 month median ramp, https://www.bridgegroupinc.com/research/2026-ae-models-motions-metrics

See what the employer actually receives. A full report for one role, with every score shown beside the excerpt that earned it, conduct findings reported rather than averaged, and a reviewer sign-off required before any decision. No form.

Read a sample reportOr talk to us about this role