AVEVA Predictive Analytics

AVEVA Predictive Analytics is the current name for the product previously sold as PRiSM and as AVEVA Predictive Asset Analytics. The current AVEVA product page uses the name AVEVA Predictive Analytics, so that is what is recorded here; the earlier names remain useful for search but are not current.

Sensor requirement: this is software, and AVEVA does not require you to buy sensors for it. It is designed to sit on top of operational data you already collect. The natural pairing is the AVEVA PI System, the historian AVEVA acquired with OSIsoft, which the page lists as a related product that collects real-time operations data, and the two are paired in AVEVA white paper material. An important honesty note: the Predictive Analytics page does not explicitly state that the product ingests PI System data. The pairing is strongly implied by AVEVA’s own cross-linking, but a buyer should confirm the connector rather than assume it. If you already run PI, that should be the first question on the call.

On mechanism, AVEVA is partially but not fully transparent. The page describes a no-code environment that flags asset anomalies before failures occur, plus fault diagnostics, case management and time-to-failure forecasts with recommended remedial actions. It states that the product ships predefined model templates and that users can bring their own algorithms in Python or a similar language, which is a meaningful and verifiable architectural fact: you are not locked into an opaque vendor model. What AVEVA does not do is name the algorithms behind the templates, enumerate supported failure modes, or state the data history required. Anomaly detection is recorded as supported because the page describes the function and the surrounding workflow; digital twin is recorded as unknown because nothing on the page describes one.

The product does not stand alone commercially. The page positions it inside AVEVA’s wider portfolio, naming CONNECT as the cloud platform and AVEVA Asset Strategy Optimization, AVEVA Mobile Operator and AVEVA Unified Operations Center as related products. That matters: Predictive Analytics is the detection layer, and the reliability-strategy work that turns a detection into a changed maintenance plan sits in a separately marketed AVEVA product.

One further note for 2026: AVEVA has entered an agreement to acquire TwinThread, another product profiled in this category. That signals AVEVA intends to broaden this part of its portfolio, and a buyer should ask how the two predictive lines will be positioned against each other.

Fits: process and power operators that already run the AVEVA PI System or another mature historian and want an early-warning layer over it. It suits organisations with a data science or reliability engineering capability, because the ability to bring your own Python algorithms alongside the shipped templates is only an advantage if someone can use it. Utilities, generation fleets, water, chemicals and mining are the typical sectors. Cloud-first buyers are well served, since the delivery described is cloud-enabled.

Does not fit: organisations with no historian. The product assumes a steady stream of operational sensor data, and there is nothing in the public material about hardware, gateways or data collection, so a site without instrumentation coverage must solve that first. It is also not the right purchase if what you want is reliability strategy, RCM or FMEA work, because AVEVA sells that as Asset Strategy Optimization, a different product. Teams wanting a packaged out-of-the-box model with no configuration will find the template plus bring-your-own-algorithm approach more open-ended than they expect.

Typical use cases: early-warning monitoring of generation assets, pumps, compressors and heat exchangers; fault diagnostics with case management so alerts are tracked to resolution; extending an existing PI System deployment from historian and dashboards into prediction; building custom models in Python against historian tags without standing up a separate data science platform.

Buyer caution: confirm in writing that your historian is a supported source, and establish whether you are buying Predictive Analytics alone or need Asset Strategy Optimization alongside it to act on what it finds.