About business outcomes, impact on learning, and AI
Saravana Sivanandham is Chief Product and Marketing Officer at Absorb Software, where she leads the strategy and execution of Absorb’s product, marketing, AI, and growth teams to deliver market-leading solutions that help organizations build critical skills, transform their workforce, and drive measurable business outcomes. Outside of work, I enjoy spending time with my family, running in the Texas hills, and playing ping pong.
Today, we discuss the use of AI in its impact on learning, business outcomes, and where performance comes into play.
The industry has long promised to prove the impact of learning on business, but the results have been mixed. Can AI actually solve that, or will it just create more activity to measure?
yes. Now that you have solved the three things that have always been missing, you can truly close the gap for the first time. You can actually get to the data on which the work is done, reason about it generatively, and then loop back to the results. The caveat is that you only quit when you measure results, not activities. Otherwise, AI will just industrialize the same vanity metrics faster.
The old Gap was never about ambition. It was about plumbing work. Since the learning system had no visibility into the work, completion and quiz scores were the only signals, and impact was inferred rather than observed. Technologies like MCP and A2A change that. Agents can now use their own permissions to read where the functionality gaps actually exist from the systems where work is done, such as CRM, support, code, and conversations, without requiring a 12-month data lake project. Generative AI translates that signal into specific interventions and reads whether performance has changed. This is a closed loop, not another dashboard.
The risks you listed are real. AI makes it very easy to generate more content and track more activity, but most platforms end up falling into exactly that trap. This discipline is about focusing on outcomes that the company already values, such as ramp-up time, win rate, retention rate, etc., and making sure that everything else contributes to that. It is clear that even the most rigorous skills measurement approaches measure competencies, not business outcomes. Measurement is a means. Proving the results is the end, and that last step is what the AI will ultimately do.
What are two or three questions you should ask someone who is currently evaluating an AI-powered learning platform?
There are three questions that separate an AI native platform from an AI feature bolted onto an LMS. Where does AI get its data? Can it act or just answer? And can it prove business results?
First, where does AI get its data from? Since everyone has access to the same underlying model, the model itself has no advantage. The data is. The most powerful systems are based on two things at once. One is the provider’s own learning data. This means who learned what, what they were able to do, and what actually worked. The other is a live connection to the business systems where work is done, such as CRM, support, and HR. Please be careful here. No vendor can learn a business performance recording system, so be wary of any vendor who claims to be able to do so. What matters is whether the platform is a system of record for capabilities and readiness, and whether it can read context from the system that performs its own performance. A wrapper for a public model that contains neither is not a demo.
Second, does it act or just respond? A chatbot answers your questions. The agent detects gaps, performs interventions, and follows up to see if it worked. Ask to see the complete workflow that will be performed, rather than a chat box.
Third, can you prove that it worked in the language that CFOs already use? Think about how you can connect learning with metrics that companies are already measuring, and how you can show cause, not just correlation. If the answer is engagement and completion, it’s an old game in a new package.
An easy way to test all three at once is to ask to see the architecture and public changelog. A truly AI-native platform provides visibility into how systems are built and shipped. You can’t do anything else.
What questions are your business customers bringing to you today that they weren’t asking a year ago?
A year ago, much of the conversation was still about features and systems. Customers asked if the platform could do certain things, or asked us to help them stand up new systems or build maps. Today’s question is about work-integrated results. Across every use case, employee development, customer education, partner enablement, and compliance, customers want the same thing in different words. Will this actually change what your employees can do? Can you prove it?
The most obvious example is skills. A year ago, a company asked us to build a skills taxonomy. Today they want almost the opposite. Rather than mapping every skill, fill in the gaps in your work flow that actually drive your business. Top-down skills projects that seek to catalog all skills, map all roles, and close gaps have become largely theoretical exercises. The map took a year to build, and from the moment it was completed it began to decay and learners never actually saw it. Skills became a means to an end.
What customers want now is what skills have always been a substitute for. The people who can do the work, and the people who can prove that the work worked, are delivered as workflows built into the work that improve themselves, rather than as separate catalogs that must be maintained. That’s what ambient, context-aware systems do. We understand the learner and business context, so we can lay the foundation for development to move your business forward. Deliver the skills that intelligence was trying to do in the right way.
Every learning platform claims to be an AI-powered platform. What does that actually mean and what should buyers be suspicious of?
Today’s use of AI covers everything from thin wrappers for public chatbots to systems that prove diagnosis, behavior, and results, so labels themselves mean little. What matters is the underlying architecture, which is divided into three honest layers. The first is AI features added to the LMS, such as content generators and question-and-answer bots. It’s convenient, but it doesn’t change your job. The second is AI-assisted, where the system provides recommendations and insights for humans to act on. It’s better, but still at a human pace. The third is AI-native, or agent ticking, where an agent detects gaps, acts on them, and measures the results across a closed loop. This is the only layer that changes the outcome rather than the effort.
Buyers should be skeptical about a few things. The claim of innate AI without its own data. Demos are impressive, but they don’t provide tangible results for your business. AI turns out to be a feature rather than a system. And the same goes for vendors who don’t publish their architecture or public change logs.
There’s also reframing, which most buyers miss. In AI, incumbency can be an advantage rather than a hindrance. It’s not the model that’s difficult. Because everyone has the same model. The hard part is having your own learning data, a system of record for capabilities and readiness, and working range within the tools where the work is already being done. Platforms that have been running enterprise learning for years have the learning data and installed base that are exactly what new entrants lack. AI is only as good as the data and the context in which it runs.
Learning, upskilling, compliance, customer training, partner assistance – businesses manage it all with a patchwork of disconnected tools. What actually is a better model?
Most companies run four or more learning systems. One for employees, one for customers, one for partners, and one for compliance. This fragmentation is the single biggest reason why learning fails to prove its effectiveness. You cannot build business-grade evidence from systems that are intentionally isolated.
Every system is a separate record, a separate budget, a wall that no intelligence can cross. The data needed to prove impact is distributed by design. A better model would provide one platform for all the audiences your business depends on, including employees, customers, partners, and vendors, all running in a single layer of intelligence and based on your company’s unique context.
It must also go beyond formal courses. Most of what an organization knows exists outside of the LMS, in places like SharePoint, Confluence, Google Drive, support tickets, and recorded calls. Modern systems connect to knowledge that already exists, so learning is based on how companies actually work, not just courses. This is also where learning systems break with horizontal knowledge and search tools. Find answers with enterprise search. Only a learning system can prove that a person is now able to do the job.
When learning is finally concentrated in one place, the system can see the big picture. How customer education impacts renewals, how partner readiness impacts channel revenue, and how employee upskilling impacts productivity. That’s not such a great integration story. It’s the difference between managing a tool and actually understanding what your employees can do.
We recently launched Absorb Aura, an agent learning system built to connect every learning interaction to business outcomes that companies are already measuring. What can L&D teams do that wasn’t possible before?
For the first time, L&D can answer questions that have been avoided for 20 years. Did it work?
Aura is the intelligence layer of Absorb’s agent learning system that connects every learning interaction to outcomes that the enterprise is already measuring. Read where capability gaps actually exist from the systems in which work is done, provide appropriate interventions, and read whether performance has changed across employees, customers, and partners. Architecturally, it is a closed loop with four systems. A competency and readiness tracking system that answers whether a person is ready to perform a specific task right now. A system of actions that intervene in the flow of work. Intelligent systems that learn what actually works. and measurement systems that link those results to business outcomes. This combination enables agent functionality rather than just AI assistance.
What is changing is the work itself. Rather than delivering programs and reporting completion, teams run agent-driven workflows to surface gaps, fill them, and prove results. Administrators who used to spend Monday tracking compliance lists can instead spend their time building next quarter’s skills strategy. And when learnings like ramp-up time, retention, and revenue finally start showing up in the CFO’s words, L&D stops sticking to the budget and starts earning a seat at the strategy table. That’s the shift.
Looking ahead three to five years, what are you most excited about where the learning industry is headed?
Two shifts, both of which move learning from the periphery of the company to the center of the company’s performance.
First, learning becomes a differentiator rather than a support function. People will do more and lead more than ever before, as AI increases productivity for everyone and expands the reach of every manager. Traditional human-based training and apprenticeships aren’t cut out for that. In that world, how quickly an organization can build capabilities will be a key competitive advantage. Learning ceases to be a background function and becomes the core muscle of an organization, perhaps the fastest growing one.
Second, we provide one-on-one coaching for all learners. We’ve always known that one-on-one learning is best, but there were never enough tutors, so we invented classrooms, books, and courses. They are all one-to-many compromises. AI removes that constraint. Every learner has a coach who gets to know them, understands the needs of their organization, and is dedicated to their success. This is the most human thing technology has done for learning in a century, and it’s exactly the model Aura is built on.
summary
A big thank you to Saravana Sivanandham for sharing her expertise in translating learning into real business outcomes by embedding AI where work actually happens. If this topic interests you, check out Absorb’s AI in learning report for exclusive insights: How L&D leaders can turn AI into business impact.
Share with
