Synthesia for Sales Enablement: Roleplay, Seats and Limits

Synthesia gives a sales enablement team two tools in one account. Roleplay Sessions, launched in October 2026, put a rep on a live spoken call with an Interactive Avatar that asks follow-ups and pushes back, and an AI coach then scores the rep against a skill rubric. Around that sits presenter-led video for product explainers and personalized outreach. The number that sets the budget is $29 per learner seat a month billed monthly, or $25 per learner seat a month billed annually, sold separately from the video plan ^[1]. Our verdict is simple. Pilot it if your team coaches cold calls and discovery at volume; skip it for multi-person deal simulations.

Monday: building a roleplay reps will actually finish

The work starts in the Synthesia app, not in the LMS. A creator picks a roleplay type (Cold call, Field sales, Manager feedback, Customer conversation, Discovery call, Media Training or Custom), duplicates a template, or, on Starter and above (or Basic with learner seats), asks Assistant to build one from a prompt plus decks, PDFs, documents and URLs. Most roleplays run 5 to 10 minutes; the help center suggests a 5-7 minute limit for practice and 8-12 minutes for assessments. One rep faces one AI persona. A buying committee on one call is not supported, so panel negotiations stay with human coaches.

The rubric is where enablement leaders spend their hours. Synthesia recommends three to five skills per roleplay, weighted to add up to 100 percent, and warns that more than five dilutes the feedback. Skills come from a prebuilt library or are written in the workspace; each rubric level is split into three sub-levels, so a rep who shows part of the next level moves up one step. You set a passing score, such as 75 percent, or switch pass/fail off, and decide whether retries are allowed. Discovery skills like qualifying questions and next-step commitment fit this structure well; vague goals like executive presence do not.

Roleplays run in 14 language options, one per session, while skill names and rubrics stay in English; English roleplays can give the avatar and the coach a US, UK or Australian accent separately. Learners can tailor the persona's job, background and personality to their own territory, but cannot change the success criteria. That keeps a field rep and an inside rep on the same scorecard. Avatars come from the Interactive tab, which holds every Express-3 stock avatar and most Express-2 ones. Roleplay scenes use no video credits.

Free plays or learner seats: the choice that decides tracking

Every plan can create and share roleplays without limit, but plays are capped. A play counts each time someone loads the roleplay on the share page, so the free tier works as a demo for a manager, not as a program for a sales floor. Seats bill separately from the Studio plan, can follow a different billing cadence, and need no paid Studio plan at all; Basic customers who buy seats also get Assistant ^[2].

Seat reductions take effect at renewal with no prorated refund, so annual seats suit a stable team and monthly seats suit a cohort that is still changing. Only enablement admins and organization admins can assign, and only after the roleplay is published. Self-Serve teams add learners by email invitation or CSV upload; syncing learners from an HR system is Custom-only. Reps get an email, a chat direct message if that integration is on, and reminders on Day 1, Day 3 and Day 7. Unless retries are switched off, they can retake as often as they like and see only their own scores. A microphone is required, a camera is not, and there is no pause.

Friday: reading the scores without misusing them

Each attempt in the Learner Portal comes back with Summary, Skills and Delivery tabs plus a timestamped transcript. Default ratings run from Not demonstrated through Below and Meets expectations to Exceeds expectations, and a creator can set up to six levels per skill. Delivery metrics cover weak words, talk ratio, median sentence length, sentence openers and repair rate. Managers and admins see individual and aggregate data, and scores roll up by individual, team and organization.

Roleplay Insights adds themes, takeaways and evidence quotes, but only once at least 10 learners with seats have made attempts. A team of eight sees none. It is a once-daily snapshot with at most six themes. If training runs through an LMS, know the SCORM limit: it carries only completion, overall score and pass/fail, with no skill breakdown or transcript, and SCORM and portal completions are not reconciled. Pick one system of record before launch.

Synthesia's Acceptable Use Policy prohibits using its outputs as a basis for employment decisions such as recruitment, promotion, termination, pay or monitoring work performance, while clearly labeled internal training and role-play are not restricted. In practice, a roleplay score can steer a coaching conversation. It cannot feed a commission plan, a performance improvement plan or a promotion case. Coach with the score. Keep it out of reviews.

What the early deployments report

Synthesia announced a partnership with Hays in October 2026, and Synthesia's case study on the Hays pilot reports over 70% voluntary engagement, a 29% average skill uplift and 60%+ of participants saying they would use it regularly ^[3]. Synthesia also says that across early deployments 78% of learners tried a roleplay more than once and 83% of those improved later. These are vendor numbers. Your own rubric on your own reps is the test that counts.

Between sessions: explainers without a studio day

The video side covers product knowledge. Assistant, the chat-based builder launched in July 2026 on Starter and above, turns a deck, a PDF or a URL into an outline and then an avatar video. Synthesia's case study says TD SYNNEX cut the time per enablement asset by 80%, from three days to two hours, and another says Infinite Peripherals produced four role-specific explainers in 48 hours and booked 35% more meetings. MP4 download starts on Starter; Basic videos carry a watermark and are shared by link or embed, with no MP4 download.

A familiar face helps adoption, and a Personal Avatar of your sales leader is possible on paid plans: three on Starter, five on Pro. Synthesia creates a personal avatar or voice clone only with a live consent recording from that same person, who must be of legal age; it refuses celebrities, public figures, deceased people and anyone who has not consented, and you cannot create a photo avatar of someone else even with their permission. Stock avatars cannot endorse products or act as spokespersons, so a persuasive pitch belongs to a personal avatar. Every video passes content moderation, and only paid plans can appeal.

Prospect-facing: personalized video and a website sales advisor

Bulk personalization sits on Pro and Enterprise only. You save a video as a template with variables for script, text, media or avatar, upload a CSV or XLSX file, and Synthesia generates the videos and emails a CSV of links. It draws 180 credits a minute (help center, October 2026), against about 100 per finished minute for a regular avatar video. Clickable CTA buttons, branching and a branded share page with your logo and links are also Pro and Enterprise features. Interactive videos cannot be downloaded as MP4, and analytics exist only for videos played in the Synthesia player.

Stock avatars may not appear in paid advertising, paid social included, without Synthesia's written consent; organic posts and website videos are fine, and Personal and Studio avatars are not bound by that rule. For the website, Interactive Avatars, available on all plans since October 2026 and labeled beta, include a sales advisor starting role. You load a knowledge base of PDF, Markdown or TXT files up to 10 MB each, then deploy a widget, an embed or a hosted link. The full widget uses 20 credits a minute and runs only on websites you allow.

What the budget looks like

Pro is the plan for prospect-facing work: $89 a month billed monthly or $59 a month billed yearly ($708 a year), per account ^[4]. By our arithmetic, yearly billing saves $360 a year; at list price it pays off if you would stay eight months or more, and with the first-month offer a new buyer needs nine. Its 6,000 monthly credits cover up to about 60 more minutes of avatar video on top of the 10 included minutes. It has one editor plus five guests who only view and comment, so a team of several builders needs Enterprise for shared editing. Learner seats are their own line. Credits reset each period and purchases are non-refundable, as our Synthesia pricing guide explains; for training beyond sales, see Synthesia for employee training.

The bottom line

Synthesia earns a pilot from enablement teams that run repeatable call types and want scored practice without booking a manager for every mock call. Test one Discovery call roleplay on the free plays, then seat a single cohort of at least ten reps so Insights switch on. Add Pro only when outbound personalization or branded share pages are on the roadmap. Keep the scores in coaching rather than reviews, and treat the vendor's pilot numbers as a hypothesis your own team still has to prove.