Guillaume and I started Rise Up in 2014 with a small idea that turned out to be a big one: L&D teams needed a better way to design and deliver learning at scale. Ten years later, Rise Up is trusted by 600+ enterprises worldwide, with 4.2M learners. We’ve shipped the full LMS-plus-LXP-plus-adaptive-pathways stack, and we’ve seen it work, in hospitality, retail, finance, healthcare, industry, energy, public sector, NHS trusts, ministries, operator networks.
We also saw, very clearly, what ten years of enterprise learning software cannot fix. Forge is the answer to that. Here is what we learned, and what drove us to build a new category.
Lesson 1, Completion is not the same as performance
Every LMS in the industry reports completion. Every enterprise above a certain size is sitting on 90-95% completion rates on mandatory training. And yet: audit findings persist, sales win rates don’t move, compliance violations keep happening, onboarding ramp times stay the same.
The reason is simple. Completion measures that someone watched the video. Performance measures whether they apply the knowledge the next time it matters. Those are two different phenomena. And for ten years, we’ve been running the industry on the first metric because it is the one we could measure.
Lesson 2, Learning disappears between modules
Forgetting curves are not a theory. They are a law. Most of the content a learner consumes in a training module is gone within 72 hours. Spaced repetition helps. Microlearning helps. Nudges help. But none of them fix the fundamental problem, which is that the moment the skill actually applies is almost never the moment the training was delivered.
That gap, between “taught” and “applied”, is the single largest value leak in enterprise L&D. It is also the gap an AI Performance Officer closes.
Lesson 3, Content is no longer the bottleneck. Context is.
A decade ago, content was scarce. Getting a course built was a 6-week engagement. Now, with generative AI, content is cheap, a competent L&D designer with the right tools can produce a module in an afternoon. Every LMS vendor has added an AI content generator.
The bottleneck has moved. The new bottleneck is: delivering the right piece of that content to the right person at the right moment. And that is a context problem, not a content problem. Generating a thousand training assets doesn’t matter if none of them arrive at the moment the skill is needed.
Lesson 4, L&D cannot own the business KPI, but it should be tied to it
In nearly every organisation we served, L&D was held to training-centric metrics, completion, satisfaction, NPS, hours delivered. Never to the business KPIs the training was supposed to move: win rate, time-to-productivity, audit pass-through, safety incident rate, ticket resolution time, customer CSAT.
That disconnect is painful, L&D is a vital function, but in the wrong frame, it looks like a cost centre. The future of L&D isn’t about producing more content; it’s about owning the handoff from knowledge to performance. That handoff is what an AI Performance Officer manages.
Lesson 5, Enterprises don’t want another platform
This is the hardest lesson, and it is the one that shaped how Forge is built. Our customers do not want a new tab to check, a new login, a new tool to integrate, a new vendor to manage. They have 250 SaaS tools already. Adding one more is a negative proposition unless the new tool reads from the tools they already use, fits into their existing flow, and disappears when the moment has passed.
Every AI coaching product we’ve seen launch in the last two years makes the same mistake: it’s a beautiful standalone app. Users don’t open it. The product dies of neglect after 90 days.
Forge reads Salesforce, Teams, Slack, HRIS, the ticketing system, the LMS, every system that holds a live signal of what the user is doing, and turns those signals into coaching delivered through its own web and mobile app at the exact moment it matters. That is not a distribution choice, it is an architectural commitment. If the coach is not grounded in what is actually happening at work, the coach doesn’t happen.
The shape of the next ten years
We think enterprise learning is splitting into three layers that coexist:
- System of record (LMS), the catalogue, the assignments, the certificates, the compliance trail. Mature. Rise Up, Cornerstone, Docebo, 360Learning, Workday. Don’t replace it.
- Discovery layer (LXP), the curated, learner-initiated content experience. Useful. Degreed, EdCast, Rise Up’s Flow. Don’t replace it.
- Performance layer (AI Performance Officer), the context-aware coach that reads the work, closes skill gaps in real time, proves impact on business KPIs. New category. What Forge is.
The first two layers answer the question “what should I learn?” The third layer answers the question “how do I do my job better, right now, in this specific moment?” Those are different questions that deserve different tools.
Why we built Forge
Because the thing we learned, ten years and 4.2M learners in, is that L&D was never going to close the last mile by itself. The last mile is a software problem. It lives in the flow of work, it requires deep context on the user and the business, and it has to be auditable, compliant, and trusted.
That’s a category. It’s not a feature of the LMS. It’s not a ChatGPT plugin. It’s a dedicated tool, an AI Performance Officer, that every enterprise function will run alongside the LMS within the decade.
We’re building it first. Built by the team behind Rise Up, the same team, the same enterprise-grade security posture, the same multi-language depth, the same 10 years of customer scars. New category, same pedigree.
If that resonates, come meet us at Learning Technologies London on 30 April, or book a 30-minute demo and we’ll show you.
, Arnaud
