Reading time: 16 minutes
Internal training without agencies: reduce your video production costs by 80% with AI
You have a training video that needs updating. The protocol changed, the product did too, or the material simply no longer reflects how the team works today. You call the agency, spend two weeks negotiating the budget, wait another four to six for delivery, and when the video finally arrives, you ask for three rounds of corrections. Total: two months and between €5,000 and €15,000 for content that needs another revision in six months.
That's if you have the budget. If you don't, the old material stays there, outdated, and nobody watches it.
This article explains what it really costs to depend on agencies for training video production, what changes when you produce internally with AI, how to take the first steps without needing a production team, and — above all — how to move from internalizing production to internalizing the whole L&D function, including the dependency on the consultancies that create the content.
The price of a professional corporate video in the Spanish and European market ranges between €3,000 and €5,000 per finished minute of content¹. A five-minute training module can therefore cost between €15,000 and €25,000.
Those are the direct costs. The indirect ones pile up quietly:
If your company has between five and twenty active training videos and updates them every one or two years, the cumulative cost can exceed €150,000 annually in outsourced production.
A company with 20 active training videos that updates its content every two years can spend more than €150,000 a year on outsourced production, not counting coordination costs or internal team time.
The financial cost is the easiest to see. The hardest to quantify is the cost of dependency: not being able to update content when you need to because doing so requires an external vendor, weeks of waiting, and a specific budget.
This paralysis has a name: Document Inertia. It's what happens when a company's operational knowledge gets trapped in formats that can't be maintained efficiently. PDF SOPs that nobody updates. Onboarding videos featuring employees who no longer work at the company. Compliance modules recorded before the last regulatory change.
The direct consequence is that teams work with outdated information. And that's not just a training quality issue: it's an operational, safety, and compliance risk.
The scale problem makes everything worse. A company with ten locations or with teams distributed across multiple countries can't afford to record a video per protocol, per language, and per update. The outsourced production model doesn't scale.
When operational knowledge gets trapped in formats nobody can keep current, distributed teams end up training on material that no longer reflects how they actually work. Depending on an agency for every update doesn't scale.
According to Training Magazine's 2025 Industry Report, U.S. companies spend an average of $874 per employee per year on training². At those volumes, content production becomes one of the main bottlenecks for L&D teams.
AI video production doesn't require a camera, a studio, or a creative team. The basic workflow is: script, avatar, voice, video. A training manager can create a complete module from their computer in the time it used to take to write the brief for the agency.
The impact on timelines is concrete. According to vidBoard.ai data, a 60-second video that required 13 days with traditional production can be completed in 27 minutes with AI tools³. A five-minute module that might take six to eight weeks with an agency shrinks to a few hours of actual work.
The cost impact is proportional. Industry analyses put the reduction in production costs across a wide range — often above 70% — when moving from outsourced to internal AI production³. But the advantage that usually matters more isn't the one-off saving: it's that every future update costs roughly the same as the first version, instead of opening a new project. You can see the full calculation in our analysis of ROI of AI training videos in industry.
The capabilities that matter for corporate training are available without a technical team:
With AI tools, producing training video stops being a commissioned project: production cost drops noticeably and any module can be updated in hours instead of weeks. Operational knowledge stops being a frozen asset.
An example makes it tangible. A plant changes the procedure for using a piece of protective equipment. Before: L&D reopens the brief, sends it to the agency, and waits for the new version while operators keep watching the old video for weeks. After: you rewrite only the affected block of the script, regenerate the module with the same avatar and voice, and the corrected version is published to the LMS that same afternoon.
The table below compares estimated costs for a typical volume of training video production in a mid-size company.
| Scenario | With external agency | With in-house production (AI) | Estimated savings |
|---|---|---|---|
| 5-min video (1 unit) | €15,000–25,000 | €2,500–4,000 | ~80% |
| Annual plan (5 videos) | €75,000–125,000 | €12,500–20,000 | ~83% |
| Update to existing video | €5,000–15,000 | €500–1,500 | ~85% |
| Catalog of 20 videos (2 years) | €150,000–250,000 | €25,000–40,000 | ~83% |
In-house production costs include the platform and team time. They don't include agency review and coordination costs, which on the left column can add 15 to 25% to the total.
The ROI expands when you factor in training's impact on talent retention. According to the Association for Talent Development (ATD), companies that invest in employee development see 46% lower turnover than those that don't⁴. With an average replacement cost per employee ranging between 50% and 200% of their annual salary, keeping training efficient and up to date is a direct lever on HR costs.
This isn't a change you make in a week, but it doesn't require an 18-month digital transformation project either. The process has three concrete phases.
Before changing anything, map what you have. How many training videos exist in the company? What's their creation date and when was the last update? What's the cumulative production and revision cost for each one?
This exercise reveals two things: the total investment already made in outsourced production, and the percentage of content that's outdated or unused. In most companies, between 40% and 60% of the training catalog hasn't been updated in more than two years.
With that information, you have the economic case to justify the change internally: it reveals how much has already gone into outsourced production and what share of the catalog is outdated or unused.
Most training content exists in PowerPoint or PDF before becoming a video. The fastest transition is to take those documents and convert them directly to video with AI, bypassing the agency entirely.
The L&D team handles editorial review, not technical production. Production time is measured in hours, not weeks. The result isn't a cinematic video — it's a functional, clear, and updatable one that does its training job without a film production budget.
You can see how this workflow plays out in our article on how to cut training content costs without losing quality.
Once existing materials are migrated, the next step is defining the workflow for new content: who writes the script, who validates the technical content, who approves before publishing.
This workflow doesn't require a video team. In most cases, it's managed by the same L&D manager who previously coordinated with the agency — the difference is they now produce too. The full cycle, from script to video published in the LMS, can be done in less than a day's work.
With this in place, every time a process, product, or regulation changes, the corresponding video gets updated the same day, not the same quarter.
Producing video in-house is the entry point, but the deeper shift is broader: moving from commissioning training to owning the function that creates and maintains it. Producing internally solves the cost; internalizing the function solves the structural dependency.
The difference is who owns each decision. When you outsource the whole function, you depend on a third party not just to record, but to decide what gets taught, how it's structured, and when it's updated. Internalizing the function means taking those decisions back and leaning on AI so a small team can sustain them without needing a vendor for every change.
Not everything has to come in-house. The key is to separate what's an advantage to own from what still makes sense to outsource.
| L&D activity | Recommendation | Why |
|---|---|---|
| Video production and updates | Internalize | It's the most recurring work and what used to create dependency and cost |
| Capturing knowledge from internal experts (SMEs) | Internalize | Operational knowledge already lives inside; the bottleneck was turning it into content, not having it |
| Instructional design of complex programs | Mixed | For long or certifiable learning paths, occasional external support can pay off |
| L&D strategy and stack architecture | Internalize | Deciding which tools you use and how they integrate is a business decision, not a task to hand off |
| High-end audiovisual production (brand, campaigns) | Outsource | Low frequency, high production demands; it isn't operational training |
Governance matters as much as production. Internalizing the function means defining who keeps the content alive and at what cadence, not just who creates it the first time. Without that maintenance routine, agency dependency is replaced by an in-house archive that ages just the same, because nobody is assigned to keep it current.
That architecture — which layer distributes, which layer produces, and which tools you choose — is a decision best made in-house. We break it down in our decision guide between an LMS and an AI video platform.
When we talk about "outsourcing training", we tend to lump together two services that are actually very different and create different dependencies:
Many companies depend on both at once without having consciously decided to. And the content one is usually the most expensive to sustain, because every time a process or a regulation changes, you have to commission the design of the update all over again.
What matters is where that dependency comes from. It's almost never a lack of knowledge: the operational know-how already lives inside, in the technicians, the quality managers, and the plant supervisors. The bottleneck was turning that knowledge into structured content and keeping it current — a task that until now required a profile few L&D teams had in-house.
Dependency on a content consultancy rarely comes from not having the knowledge, but from not having an agile way to turn internal knowledge into maintainable training.
That's where AI-supported capture of internal experts changes the math. When a training manager can sit down with an SME, structure what they know, and turn it into an updatable module without going through a vendor, the consultancy stops being indispensable for the bulk of operational content. And when every regulatory change no longer forces a fresh external commission, keeping mandatory training current stops being a recurring expense: it's the condition for getting the whole workforce to complete mandatory training without depending on a third party for every update.
This doesn't remove the value of expert instructional design in long or certifiable learning paths, where occasional external support still makes sense. What it removes is the default dependency: paying a consultancy for content the organization already knows, simply because it had no way to structure it.
In-house AI production doesn't replace everything. There are cases where outsourced professional production is still the right call.
High-visibility institutional communication videos, brand spots for major campaigns, or productions requiring on-location shoots are projects where the investment in professional production is justified. The same applies to employer branding content designed for mass external distribution.
In-house AI production is built for volume: operational training, SOPs, onboarding, compliance, product tutorials, process updates. Content that needs to be accurate, current, and accessible — not necessarily cinematic.
The most efficient combination for many companies is using external production for high-visibility projects and in-house AI production for the bulk of the training catalog — the part that needs speed and scalability.
Depending on an agency to produce training videos was the only available option five years ago. Today it's a choice — and an expensive one.
Companies that have internalized their training video production haven't reduced the quality of their content. They've gained control over their knowledge infrastructure: they can update in hours, produce in any language, and scale without increasing their production budget.
The savings aren't the goal in themselves. They're the result of no longer paying in each production cycle what a single project used to cost.
If you want to see how the main platforms compare, our guide to AI tools for corporate training in 2026 has an up-to-date comparison.
A five-minute module can be produced in two to four hours of actual work, including script writing, video generation, and revisions. With outsourced production, the same video typically takes four to eight weeks from briefing to final delivery.
The most common in corporate training are onboarding modules, SOP and operational procedure videos, compliance and safety training, tool or product tutorials, and process updates. Any content that requires frequent updates or distribution to distributed teams is a good candidate for internal production.
The onboarding process for platforms like Vidext includes a kick-off session and hands-on training with the team. Most L&D managers without prior video production experience generate their first module within three to five days of use. No technical background required.
For operational training, yes. Videos with AI avatars and generated voices have a presentation quality comparable to a standard corporate video. They don't have the cinematic quality of a high-budget production, but that level of production isn't necessary for a compliance module or product tutorial to be effective.
Yes. Platforms like Vidext generate videos in over 120 languages with standard and professional voices, including regional Spanish dialects (Catalan, Galician, Basque). The same content can be adapted to different markets without re-recording.
Programme to promote permanent employment of qualified young people within the framework of the National Youth Guarantee System. Vidext has received a grant of €25,401 awarded by LABORA (Valencian Employment and Training Service) for the permanent hiring in 2024 of qualified young person(s) registered in the National Youth Guarantee System, an action eligible for co-financing by the European Social Fund Plus (ESF+) 2021-2027 or any other European Union fund. Expediente ECOGJU/2024/550/46. This publication is made in compliance with the transparency obligations established by Spanish Law 19/2013 of 9 December.
