Key takeaways
The 2026 KuppingerCole IGA Leadership Compass reveals a structural gap across all 31 evaluated platforms: disconnected legacy applications that lack modern APIs remain ungoverned. Risk concentrates in these systems—mainframes, terminal-based platforms, and custom tools—where access drifts and manual workarounds replace true governance. RPA-based connectors fail at scale when application interfaces change.
Industry analysis.
KuppingerCole published its 2026 IGA Leadership Compass in May. It evaluated 31 vendors. It named Overall Leaders. It ranked platforms on product depth, innovation, and market presence. It is the most-referenced independent evaluation in the identity governance market, and every serious IGA vendor spends months preparing for it. What it also did – quietly, in the challenges section of vendor after vendor – was document the same structural gap. Not a gap belonging to one platform or one architecture. A gap that belongs to the category.
The analyst named the IGA connector gap. Every platform has it.
The 2026 report is direct about what matters in IGA. “Connector depth and reconciliation capabilities,” the analyst writes, are critical for managing complex entitlement models in line-of-business applications. That criterion appears in the required capabilities section. It is one of the things every platform is measured against.
Then the individual vendor profiles arrive, and a pattern forms.
One Overall Leader is noted to rely on RPA-based connectors to extend coverage to systems outside its native library. Another has legacy connector gaps listed as a formal challenge. A third – a platform recognized for its broad SaaS connectivity – acknowledges that disconnected applications can only be included in access reviews through ticketed and workflow-based approaches. In analyst language, that phrase means someone is handling it by hand.
This is not a coincidence of evaluation methodology or a sign that any single platform is falling short. It reflects something structural about how IGA was built. The connector ecosystems that exist today were designed around the applications that were willing to cooperate: modern SaaS platforms with documented APIs, directories with standard protocols, HR systems built to integrate. Those applications are well-governed. The problem is they are not where the risk lives
Risk concentrates in disconnected applications your IGA platform cannot reach.
Ask any identity team which applications they worry about most, and the answers are consistent. The mainframe environment that handles financial settlements and predates the current compliance regime by two decades. The clinical platform that runs on a terminal interface and has never had an API, not because no one wanted one but because the vendor stopped shipping new versions before SCIM existed. The line-of-business tool a single team built internally and now controls access to a sensitive data environment that auditors have been circling for years.
These are not marginal applications. They are the ones that concentrate privileged access, accumulate orphaned accounts, and resist the governance motions that work cleanly everywhere else. And the same characteristics that make them hard to connect – age, custom architecture, no modern interface – are the characteristics that let access drift and stay drifted.
The standard responses to a disconnected application each run out of road in the same place. Waiting for an API assumes a vendor who is still actively developing a product that, in many cases, shipped before the current SCIM specification existed. Building a custom connector works until the application changes, which it will, and until the IAM team that built it has moved on, which they will. Managing the application by hand – spreadsheets, email approvals, quarterly certification campaigns where managers are approving access to systems they do not fully understand – is not governance. It is the appearance of governance, maintained at significant operational cost, and it is the first thing a sophisticated auditor finds.
The 2026 KuppingerCole report documents this reality without framing it as failure. It frames it accurately: as a gap that has persisted across the category, and that RPA has not closed.
Why RPA-based IGA connectors keep appearing – and why identity governance connectivity breaks down
The 2026 Leadership Compass mentions RPA in the context of connectivity for at least three vendors, including one Overall Leader whose entire connector architecture is built on it. KuppingerCole notes in a separate profile that “bot or RPA certifications remain limited.”
This is not a condemnation of the intent. It is an accurate description of what happens when a task automation technology gets applied to an identity governance problem. RPA is built to automate repetitive tasks through a scripted sequence of interface interactions. For a specific application, on a stable interface, with a team available to maintain the script, it works. The governance problem arrives when the interface changes – after an application update, a UI refresh, a vendor patch – and the script breaks.
Provisioning queues silently. The Leaver event that should have triggered deprovisioning across a set of legacy applications does not. The audit trail shows a gap that may not be discovered until the next certification cycle, or later.
Scaled across twenty or forty disconnected applications, each with its own update cadence and interface quirks, the operational cost of keeping RPA connectors current becomes the work itself. The IAM team that was supposed to be expanding governance coverage is instead doing connector maintenance. The gap that was supposed to be closing is staying
roughly the same size.
That is the context in which the 2026 report’s language should be read. When the analyst describes ticketed workarounds and RPA-based coverage extensions as challenges, it is describing a category that has not yet found a durable answer to its hardest connectivity problem.
How READI extends IGA provisioning automation to legacy and disconnected applications
READI is not an IGA platform. It is the connectivity layer that allows IGA platforms to govern the applications they were not designed to reach.
The Smart Connector uses a combination of a language model and a computer vision engine to operate applications through the interface they already have – the screen. Plain[1]English instructions describe what a provisioning action should do. The language model translates those instructions into structured intent. The computer vision engine reads the
actual interface at runtime to identify the elements it needs to interact with. The result is a connector that works across web applications, Win32 desktop applications, and terminal systems without requiring an API, without requiring scripting, and without breaking when the application interface changes. When a UI update shifts a button or relabels a field, the connector self-heals.
The bidirectional architecture matters as much as the coverage. READI aggregates identity data back from disconnected applications into the IGA platform, not just provisioning out tothem. That means the governance platform has visibility into what access exists, not just the ability to act on it. The Leaver event that previously generated a ServiceNow ticket and a
manual workflow becomes an automated, auditable, policy-enforced deprovisioning action – including across the mainframe terminal system, the legacy clinical platform, and every other application that was previously outside the program’s reach.
The IGA platform remains the system of record and the policy authority. READI extends its execution reach to the applications that were always in scope from a risk perspective, but never in scope from a connectivity one.
What the KuppingerCole challenges sections tell every identity governance team.
If your organization is evaluating IGA platforms using the 2026 KuppingerCole Leadership Compass – which is a reasonable place to start – read the challenges sections as carefully as the strengths. The leaders are well-documented and genuinely capable. The gaps are also well-documented and genuinely structural.
The question to ask your platform vendor at any stage of evaluation or renewal is not whether their connector library covers your SaaS estate. It almost certainly does. The question is what governance looks like for the thirty or forty applications that are not on that list – the terminal systems, the legacy line-of-business platforms, the applications that have been deferred from the IGA program for long enough that they have stopped being discussed. Those applications do not get smaller as a governance problem the longer they wait. They get larger.
READI exists to close that gap, alongside the IGA platform you have already invested in or are in the process of selecting. The 2026 report documented the problem clearly. The answer is already available.