Let's Stop Making 'Non-Interoperable Crap' in EdTech

“Stop making non-interoperable crap in edtech.” ?

Those were the opening words from Dr. Rebecca Winthrop (Senior Fellow and Co-Director at the Center for Universal Education at The Brookings Institution ) at the plenary session of last week’s HolonIQ Global Impact Summit in New York City. ?

Throughout the day-long event attended by hundreds of the world's most prominent edtechs and impact organizations, the pain point of #interoperability emerged repeatedly, just as it had at the week prior at the HR Tech in Las Vegas, attended by thousands of the world's human capital and technology leaders.

Interoperability, or the capacity for systems to "talk" to one another, appears to be a clarion call -- if not a mandate -- for #learning and #skills companies across their customer base. So, why are we still developing solutions that can't share skills data easily, and more to the point, are we doing it to the detriment of learners? ?

In response to unprecedented demand (and venture capital) for upskilling, remote learning, and L&D solutions, education and HR tech companies have developed scores of products that don’t talk to each other, many of which maintain proprietary skills taxonomies used to infer learners' skills, influence content recommendations, and guide the user experience.

How does this serve learners?

Many of these mix-and-match taxonomies and ontologies are anchored to static or infrequently updated labor market information (LMI), which is often additionally “massaged” by humans-in-the-loop, introducing subjectivity and embedding human bias into the machine intelligence and AI which feeds off it.

How does this serve learners?

In addition, much of this proprietary skills AI is divorced from the real-time dynamics of the global labor market, an input that matters greatly for the majority of learners who are trying to acquire skills and leverage skills currency in an uncertain and dynamic economy to achieve a pay raise, advancement, or change jobs or careers. ?

How does this serve learners?

The answer is that it doesn't: the intent of most proprietary taxonomies is purposeful non-interoperability. Companies who develop them often hope to monetize their learners' data and/or create a "moat" that discourages competitors. In their attempts to corner an exploding #edtech market, these players have unwittingly transferred complexity onto the shoulders of their customers in education, HR, and workforce development (often referred to as "partners"), adding both cost and integration effort for them to realize value from their purchase without compensatory improvements to the learner experience or outcomes. ?

Is this any way to treat a partner?

I recently met an executive from an edtech acquired by a leading LMS who was bemoaning the fact that as part of post-merger integration they would have to perform a laborious manual skills mapping to cross-walk the taxonomy from the acquired company's catalog to its new parent, an effort that wouldn’t illuminate much about skills and would likely stagnate from the moment the exercise was completed. In a world where skills evolve at the speed of disruption, they asked, how is this a valuable use of resources?

If the idea of conducting manual skills mapping between enterprise applications seems like throwaway work, consider that it also forces BOTH edtechs and their customers outside of their core competencies and thrusts them into the challenging realms of AI and machine learning, where they likely don’t have the infrastructure, computing power, or ML and data engineering capabilities to become a category leader.

Even for the global edtechs with tens or hundreds of millions of users, their learner populations in their data are far from representative of the global labor market of 3.5 Billion people and its fluctuations across geographies, sectors, etc. Attempting to win in both AI and education realms for most of these players is to become a jack of all trades...which is to say, a master of none. ?

As the economy slows and enterprise budgets tighten, companies and organizations will likely continue to experience skills gaps that require them to invest in edtech, but they will also likely become more reluctant to invest in solutions that contribute additional layers of non-interoperable functionality into the tech stack and contain hidden costs to realize value.?

A ray of light for skills interoperability

A rising solution set to address this universal pain are skills technologies that incorporate platform-as-a-service (PaaS) offerings, which use scalable API technology to help non-interoperable applications “talk” to one another. With platforms to translate across data silos, no company has to abandon their beloved skills framework or taxonomy (into which they’ve often invested a lot of sweat equity), nor do they have to adopt another SaaS company's UX interpretation of their users’ needs.

Case in point: SkyHive Technologies, who in the past two weeks unveiled SkyHive Platform as a solution, a working deployment powering a hiring engine for the U.S. nonprofit Opportunity@Work, and an enterprise partnership with HCM giant Workday to translate skills across the enterprise. Organizations that invest in skills can now leverage an AI innovator's computing power and fluid skills ontology as a "translation table" across previously non-interoperable applications, harmonizing them to a common skills language that also reflects the real-world job market that learners are skilling towards.

In a world where all of us must now upskill continuously to stay relevant and employable, this kind of flexible interoperability helps the entire field of edtech applications serve their learners with greater distinction, focus less on trying to be something they're not, and concentrate their creative energies on the pedagogy, instructional design, and compelling learner experiences that support persistence, completion, skills mastery and outcomes.

The real question is for those who don’t deploy it

Are you truly serving your customers, and more specifically, your learners? Or are you creating another “layer” for them to sift through in their quest to upskill?

We haven't solved the problem by any means. But knowing that more edtech and skills applications are aligning to the global labor market as it evolves (versus any one static or "proprietary" taxonomy developed in-house) should give folks like Dr. Winthrop some hope that we are slowly -- but surely -- moving towards a greater degree of interoperability, coupled with a more thoughtful consideration for edtech buyers, their learners, and the ROI both seek from investments in skills and learning. ?