This is original Revspire data: 20 B2B companies, more than 2,000 scored AI role-play sessions, October 2025 to September 2026. The figures are aggregated and anonymized, and they describe patterns in Revspire customer data rather than proving cause and effect.
Which sales skills do reps score lowest on?
Closing technique (3.0 out of 5) and objection handling (3.2) were the weakest skills across all scored role-plays, followed by value proposition delivery (3.7) and active listening (3.8). Reps were strongest on product knowledge (4.4) and discovery questioning (4.1). The pattern is consistent: reps know what they sell and can open a conversation, but lose ground when they have to handle resistance and ask for the next commitment.
- Product knowledge4.4
- Discovery questioning4.1
- Active listening3.8
- Value proposition delivery3.7
- Objection handling3.2
- Closing technique3.0
Source: Revspire customer data, October 2025 to September 2026.
- Each session is scored by Revspire's AI against the rubric the customer configured; scores are averaged across sessions.
Which sales objections are hardest for reps to handle?
The status quo objection ("our current process works fine") was the hardest, with an average score of 2.9 out of 5. Price ("it's too expensive") came next at 3.2, then competitor (3.5) and timing (3.8). Security concerns were handled best, at 4.1, probably because most companies give reps a prepared answer and documents for them. The weakest two are the ones with no ready-made answer: they need the rep to build a business case for change.
- Security: "We have concerns about data security."4.1
- Timing: "We're not ready yet."3.8
- Competitor: "We're already using another solution."3.5
- Price: "It's too expensive."3.2
- Status quo: "Our current process works fine."2.9
Source: Revspire customer data, October 2025 to September 2026.
- Objection type is the scenario the session was built around.
How many attempts does it take a rep to pass a sales certification?
Reps needed a median of two to three attempts to pass a role-play certification. A first-attempt pass is the exception, not the rule, which is a reason to design certification as practice with feedback rather than a one-shot exam, and to set the retake policy before launch.
- Attempts are counted per rep per certification scenario until the first pass.
How to use this benchmark
- Train where the scores are lowest, not where content is easiest to make. In this data that means closing and the status quo objection, ahead of more product training.
- Build a business-case answer for "our current process works fine": the cost of the current process, the risk of waiting, and a small first step. Then have reps practise it until they score above 3.5.
- Plan for two to three certification attempts per rep when you set timelines and manager time for a launch or onboarding cohort.
- Compare your own rubric scores by criterion against this table to find the team-wide gap, then coach individuals on their own lowest criterion.
Method and sample
- Organizations
- 20 B2B companies
- Records
- More than 2,000 scored AI role-play sessions
- Period
- October 2025 to September 2026
The figures come from more than 2,000 AI role-play sessions scored in Revspire AI Sales Training & Coaching across 20 customer organizations between October 2025 and September 2026.
Scores use a 1–5 scale. Revspire aggregated them before analysis; no figure identifies a customer, a rep or a session, and each figure is based on at least 10 organizations.
Internal, demo, trial and test workspaces are excluded. Customer contracts allow aggregated, anonymized benchmarks.
Role-play scores measure performance in practice conversations, not in live deals. They are observed patterns, not proof that practice caused a sales result.
Limitations
- Rubrics are configured by each customer, so criteria with the same name may be weighted or worded differently between companies.
- Teams choose which scenarios to practise, so harder or newer scenarios may be over-represented.
Frequently asked questions
What sales skills are most often missing?
In 2,000+ scored AI role-plays, closing (3.0 out of 5) and objection handling (3.2) scored lowest, while product knowledge (4.4) and discovery questioning (4.1) scored highest.
What is the hardest sales objection to handle?
The status quo objection, "our current process works fine", averaged 2.9 out of 5, below price (3.2), competitor (3.5), timing (3.8) and security (4.1).
How many attempts should reps get to pass a sales certification?
Reps in this data needed a median of two to three attempts to pass, so a policy that allows at least three attempts, with coaching between them, matches how people actually learn.
Can AI role-play identify objection handling gaps?
Yes: scoring each session against the same rubric shows which objection types a rep or team handles worst. In this benchmark it separated the status quo objection (2.9) from security concerns (4.1).
Can I use this benchmark in my own report?
Yes, cite it as "Revspire sales skills gap benchmark, 2026" with a link to this page, and include the sample and time window shown in the method section.
How to cite this benchmark
Revspire, "Sales skills gap benchmark: what 2,000+ scored AI role-plays show", 2026. https://www.revspire.io/research/sales-skills-gap-benchmark
Related on Revspire
- Sales objection statistics: the published researchThird-party data on the most common objections and the ones that lose deals.
- How to build a sales certification programFormats, pass marks, attempts policy and recertification.
- Sales training needs analysisTurn skill scores into a prioritised training plan.
- Revspire AI Sales Training & CoachingAI role-play, objective scoring and readiness tracking.