The only price Skillsoft publishes anywhere is a self-serve tier that stops at 300 seats. Above that line there is no list price, no published volume discount, and a negotiated contract for a different product. If your company employs 50 to 500 people, the tier boundary runs straight through you — which means the Skillsoft most review sites describe is not the Skillsoft you will be quoted.
A declaration first. SFAI Labs runs facilitated AI enablement sessions teaching teams to work with assistants on their own documents, so we compete with part of what Skillsoft sells. We also think a governed content library is the right purchase in several real situations, named below. What follows is read from Skillsoft’s own documentation and filings rather than from star ratings.
What you are actually buying
Three things carry the Skillsoft name, and buyers routinely conflate them.
Percipio is the platform: content library, learner interface, assignment and reporting layer, and the recommendation engine that suggests courses by role and skill. A next-generation version reached general availability in the fiscal year ended January 2026.
The AI content inside Percipio is a catalog: generative-AI courses, role-based journeys, and Skill Benchmarks that score a learner against a skill area before and after. This is where “AI training” usually points when a Skillsoft seller uses the phrase.
The AI Skill Accelerator is the wrapped program, built with Microsoft and the vehicle for the joint generative-AI skilling program the two announced in July 2024: on-demand courses, coaching, live instructor-led training, and practice labs, with Skill Benchmarks for measurement. Skillsoft says it serves over 60% of the Fortune 1000 across 160 countries — the scale this program was designed around.
Watch that third item in a proposal. The coaching hours, instructor-led days, and lab access are what produce practice rather than viewing, and they are priced separately from library access. A quote that reads as one number often contains one line of real practice and a great deal of catalog.
The 300-seat line
Skillsoft does not publish enterprise pricing. Independent pricing research that checked the checkout directly in 2026 found a single published figure: the self-serve Skillsoft Teams tier at $319 per user per year, available for 2 to 300 seats, after a 30-day trial. No volume discount is published at that tier, so the arithmetic is linear.
| Seats | Published self-serve cost, license only |
|---|---|
| 50 | $15,950 |
| 150 | $47,850 |
| 300 | $95,700 |
Above 300 seats, the price is whatever your negotiation produces. Third-party aggregators put enterprise rates at $200 to $350 per user for deployments of 500 to 1,000 learners, but those are modeled figures from spend-data vendors whose own research flags roughly 30% variance against real outcomes. Treat them as a sanity check, not a quote.
Two consequences for a company of 50 to 500 people. First, you may be eligible for the published tier and not know it, because a seller’s default motion is the enterprise contract. Ask whether your seat count qualifies, and what the enterprise contract adds for the difference.
Second, the two are not the same product. Custom learning paths, deeper analytics, API and SCORM integration, assigned customer success, and the AI simulator below sit on the enterprise side. Reading a review of enterprise Percipio and then buying self-serve produces a predictable disappointment.
What CAISY is, and what it is gated behind
CAISY is Skillsoft’s conversation AI simulator and its most distinctive AI feature. The learner documentation describes it as interactive, scenario-based content powered by artificial intelligence, in three modes: Practice, where the learner supplies their own wording and is scored; Role model, where the simulator demonstrates the skill while the learner takes the opposing role; and Guiding, where an AI Coach comments after a few exchanges. Only Practice is scored. A learner can complete up to 50 scenarios a day, and the feature must be enabled for your site before anyone sees it.
Two facts from the administrator documentation that review pages omit.
It is license-gated. Skillsoft’s FAQ states that CAISY can be made visible to learners if you hold any one of the Expert collections or the Skillsoft Leadership Development Program, and that scenarios are provided to customers licensing Business or that program. Compliance-only customers see four compliance simulations, reachable solely through direct library access. If a demo showcased CAISY, confirm in writing which collection in your quote carries it.
It practices conversations, not assistants. CAISY rehearses difficult workplace exchanges — feedback, negotiation, escalation. That is a legitimate and well-built thing. It is not practice at prompting an assistant against your own documents. A buyer who approves Percipio to fix “our people hold AI licenses and still produce weak output” is buying a roleplay simulator plus a video catalog, which addresses a different problem. Independent analysis agrees: CAISY is a practice tool for specific scenarios, not an always-on coaching system.
The governance answers, in fairness, beat the marketing. The FAQ states that the model does not store conversation data, audio recordings are not saved, learner-side conversations are logged only with the admin setting enabled and no learner opt-out, access is limited to a small set of Skillsoft employees, and log data can be deleted on request. Hand those sentences to counsel; they sit in documentation, not on any comparison page.
How to read Skillsoft’s AI numbers
Skillsoft reported the following year-over-year changes in its AI learning activity for the fiscal year ended January 2026:
| Reported metric | Change |
|---|---|
| AI skill benchmark completions | +994% |
| AI content completions | +261% |
| AI journey completions | +222% |
| CAISY learners | +146% |
| CAISY launches and engagement | +341% |
Read the nouns. Completions, learners, launches. Every figure counts an action taken inside the platform, off a base small enough that a four-digit growth rate is unremarkable. None says anyone’s work changed.
These numbers travel. A platform owner justifying a renewal will pull the same report shape for their own account, and a board will read it as capability. It is attendance. The test we apply instead, set out in our guide to proving that capability moved, is simple: name one artifact produced faster or better than it was in March, and name the person who produces it.
Skill Benchmarks are the closest thing in the product to a real capability measure and are worth using, provided nobody retakes one until the number looks good.
The segment risk nobody in the reviews mentions
This sits in Skillsoft’s filings and in none of the review aggregators, and for a buyer who wants blended delivery it is the most important thing here.
Skillsoft reports in two segments. Talent Development Solutions is the Percipio platform and content business. Global Knowledge is the instructor-led business — live classes with instructors, the part a buyer leans on when they want more than video.
For the fiscal year ended January 2026, total revenue was $513 million against $531 million the prior year. Talent Development Solutions was roughly flat at $404 million against $406 million. Global Knowledge fell to $109 million from $125 million. Net loss widened to $140 million from $122 million. The company has opened a strategic review of Global Knowledge and states that it is ongoing, and its fiscal 2027 outlook covers Talent Development Solutions only — revenue of $388 million to $406 million — with no guidance for Global Knowledge.
Translate that for procurement. The live-instruction arm is shrinking fastest, is under formal review, and was not forecast. If your reason for choosing Skillsoft over a pure content library is that it can also deliver instructor-led days, you are relying on the one part of the business whose future management has not committed to in writing. That is not a prediction of failure; it is a question for the contract. Who delivers the live component, under what notice, and what happens to your scheduled days if the segment changes hands? Multi-year terms deserve a continuity clause, the kind we cover in our piece on the contract terms that decide whether training lands.
What the reviews agree on
Across G2, Capterra, Gartner Peer Insights, PeerSpot and Info-Tech, Skillsoft sits between roughly 4.1 and 4.3 out of 5 on the high-volume sites, with curated enterprise panels higher and composite indices lower. Review volume is healthy on the aggregate product pages and thin on individual AI courses, several rated on one or two reviews. Do not read a five-star AI course rating as a signal.
The praise is credible: library breadth, role-based recommendations reviewers find genuinely useful, Skill Benchmarks for gap diagnosis, AI translation for distributed teams, and integration with systems like Workday.
So is the criticism:
- A dated interface is the most frequently cited complaint across sites.
- Content overload. The catalog’s size makes the platform hard to navigate — a real adoption tax when nobody curates.
- Passive, video-first delivery. One reviewer’s framing: the format gives you no way to ask a question. For AI skills, learned by producing output and having it corrected, this is the most consequential limitation here.
- Credential weight. Reviewers report that Skillsoft certificates are not treated as industry-standard credentials. For an administrator or engineer, a Microsoft or cloud-provider certification carries more labor-market signal.
Where a library genuinely wins
We sell against part of this, so here is the honest version. A Percipio-class library is the right purchase when you need governed, auditable coverage at scale across compliance, security awareness and policy, in many languages, with records that survive an audit; when demand is heterogeneous enough that no facilitated program covers the spread and buying four vendors is worse; and when your people sit across time zones that make synchronous sessions punishing.
It is the wrong purchase when a specific team cannot get good output out of assistants it already pays for. That is a practice problem, and catalogs do not fix practice problems. The same trade appears with every self-paced provider, which is why the platform-versus-program question keeps producing the same answer.
The comparison the aggregators get wrong
Every page-one result frames this as platform against platform: Skillsoft against LinkedIn Learning, against Pluralsight, against Coursera for Business. That is the right frame if you are replacing a library.
It is the wrong frame for the buyer this is written for — the person who already approved AI licenses, is being asked what came of them, and is reading a Skillsoft proposal because a library is the familiar shape of a training purchase. For that person the real comparison is catalog access against facilitated practice on their own work, and it usually resolves to a sequence rather than a choice. Run a small facilitated engagement on the documents your team actually produces, use what it exposes to decide what the catalog is for, then buy the catalog scoped to that. We work through that ordering decision in the buy-versus-build guide, and the broader case for why licenses are not capability sits in the enablement manifesto. Market rates for the facilitated side run roughly $2,000 to $15,000 per engagement, inside the rounding error on a 300-seat library contract.
Questions to put in the pilot
Get these answered in the proposal rather than on a call:
- Does our seat count qualify for the published self-serve tier, and what does the enterprise contract add?
- Which collection in this quote includes CAISY, and will you name it in the order form?
- How many coaching hours, instructor-led days, and lab seats are included, as separate lines?
- Who delivers the instructor-led component, and what continuity protection applies over a multi-year term?
- Can we run Skill Benchmarks on a 20-person group before committing and see both scores ourselves?
- What does your reporting show about application rather than completion, and can we see an anonymized example?
If the last one produces only completion dashboards, you have learned what the platform measures.
Frequently asked questions
How much does Skillsoft cost?
The only price Skillsoft publishes is the self-serve Skillsoft Teams tier at $319 per user per year for 2 to 300 seats, verified at checkout in September 2026. Enterprise Percipio above 300 seats is quote-only, with no published list price or volume discount. Third-party spend data suggests $200 to $350 per user at 500 to 1,000 learners, with wide variance.
Is Percipio worth it for a company of 50 to 500 people?
It depends which problem you are buying against. For governed coverage across compliance, leadership and technical topics in many languages, yes. For making one team fluent with assistants they already license, no — a catalog does not produce supervised practice. Check first whether your seat count qualifies for the published tier.
What is CAISY?
CAISY is Skillsoft’s conversation AI simulator: scenario-based practice where a learner rehearses a workplace conversation with an AI counterpart and receives feedback. It offers Practice, Role model and Guiding modes, limited to 50 scenarios per learner per day.
Is CAISY included with every Percipio license?
No. Skillsoft’s administrator documentation states it is available to customers holding any one of the Expert collections or the Skillsoft Leadership Development Program, and that scenarios are provided to customers licensing Business or that program. Compliance-only customers get four compliance simulations through direct library access. Confirm the collection named in your quote.
Does CAISY teach people to use AI assistants?
Not directly. CAISY rehearses human conversations — feedback, negotiation, escalation. Practicing prompts against your own documents is a different exercise, and if that is the gap you are closing, CAISY will not close it.
Are Skillsoft certifications recognized by employers?
Reviewers across multiple platforms report that Skillsoft’s own certificates are not treated as industry-standard credentials. For technical roles, vendor certifications from Microsoft or a major cloud provider carry more weight. Skillsoft’s value as preparation for those exams is a separate and stronger claim.
Does Skillsoft still offer live instructor-led AI training?
Yes, through Global Knowledge, but that segment reported $109 million for the fiscal year ended January 2026 against $125 million the prior year, is under ongoing strategic review, and was excluded from fiscal 2027 guidance. If live delivery is a reason you are choosing Skillsoft, ask for continuity terms.
Is our data safe in CAISY?
Skillsoft’s documentation states that the model does not store conversation data, audio recordings are not saved, learner-side conversations are logged only with an admin setting enabled and no learner opt-out, and log data can be deleted on request. Retention otherwise follows your contract.
Will a content library make our team AI-fluent?
Rarely on its own. Fluency comes from producing real work with an assistant and having the output corrected, which a video catalog is not built to do. What works is a facilitated engagement on the team’s own documents, then a catalog to maintain it.
How should we measure whether any of this worked?
Measure output, not activity. Pick two or three artifacts your team produces weekly, record how long they take and how many revision rounds they need before you start, then read the same numbers 60 days later. Skill Benchmarks before and after add a defensible score, provided retakes are not permitted until the number improves.
Key takeaways
- The only published Skillsoft price is a self-serve tier at $319 per user per year capped at 300 seats. Companies of 50 to 500 people straddle that line, so the product reviewed and the product quoted are often different.
- CAISY, the flagship AI feature, is license-gated to Expert collections, Business, or the Leadership Development Program. Get the collection named in the order form.
- CAISY rehearses human conversations, not assistant use. It does not address the gap most license-holding teams actually have.
- Skillsoft’s reported AI metrics — completions up 994%, 261% and 222%, CAISY launches up 341% — are activity counts off a small base. They measure attendance.
- Global Knowledge, the instructor-led segment, fell to $109 million from $125 million, is under strategic review, and was excluded from fiscal 2027 guidance. Blended-delivery buyers need continuity terms.
Holding a Percipio proposal and unsure whether it will change what your team produces? The cheapest next step is not a longer evaluation. Book a free AI-readiness call and we will read the quote with you — including the case where the honest answer is that the library is the right buy and a facilitated session is not what you need.
Arthur Wandzel
