SailPoint has published the latest edition of its annual Horizons of Identity Security research, a study that has become one of the more closely read benchmarks in the identity governance and administration (IGA) market. Its value to practitioners is not the vendor framing but the maturity model underneath it, which describes how organisations progress from basic provisioning and directory hygiene toward governed, automated and eventually adaptive identity controls.

The problem with maturity self-assessment

Most identity governance programmes believe they are further along than they are. The reason is structural rather than dishonest. IGA capability tends to be measured by artefact: a tool has been purchased, a certification campaign has been scheduled, a joiner workflow has been automated for the largest population. Those artefacts are real, but they describe implementation activity rather than governance outcome.

The harder question is whether entitlement decisions are actually informed. A certification campaign that reviewers approve at volume, without supporting context about what the access enables, produces evidence of activity and very little evidence of control. The same is true of role models that were built once and never reconciled against how work is really performed. Maturity research is useful precisely because it separates those two conditions.

Three signals worth taking seriously

First, the gap between identity security leaders and the rest is consistently described as widening rather than narrowing. That pattern matters because identity governance is not a capability that scales incrementally. Organisations with mature programmes typically run cleaner joiner, mover and leaver processes, which means their entitlement data is trustworthy enough to automate against. Those without it face a data quality problem that every additional automation makes more expensive to unwind.

Second, the research consistently shows that identity governance value concentrates in a small number of high-impact controls. Comprehensive lifecycle management across the whole estate is aspirational for most organisations. Governing the identities and entitlements that matter, particularly those with privileged reach, administrative rights or access to regulated data, delivers a disproportionate share of the risk reduction for a far smaller programme footprint.

Third, the emergence of non-human and AI agent identities reshapes the maturity curve. Traditional IGA maturity assumes identities are created by human process, reviewed by human managers and terminated by human action. Agent and workload identities are provisioned programmatically, often by teams outside the identity function, and they can outnumber human identities in cloud-native estates. Programmes that were mature on workforce identity governance can find themselves back at an early stage of maturity for machine identity without recognising the regression.

How to use the model honestly

The most useful application of identity security research is diagnostic rather than strategic. Take one control area, such as access to a regulated system of record, and test whether the programme can answer four questions: who has access, why they have it, who approved it, and when it was last reviewed with real evidence. If any of those answers requires a spreadsheet exercise, the maturity level is being overstated regardless of what the tooling claims.

Identity governance administration roadmaps built from that kind of honest baseline tend to prioritise data quality and ownership before expansion. Identity lifecycle management automation, segregation of duties analysis and agent governance all depend on the same foundation: an entitlement model that reflects reality and named owners accountable for it.