Falkonry (IFS) review: scores, strengths, limits
Last verified September 1, 2026 · how we score
Falkonry built genuinely strong self-supervised anomaly detection for high-frequency industrial signals — the kind of ML that finds pattern shifts humans cannot see — and its 2023 acquisition by IFS folded that capability into a broader enterprise asset-management suite.
Post-acquisition, the technology increasingly reaches the market as part of IFS's EAM/ERP motion rather than as a standalone plant purchase, which changes both the buying process and the pricing conversation.
Technically credible; commercially now an IFS-ecosystem decision.
Scores
| Criterion | Weight | Score | |
|---|---|---|---|
| Time-to-value How fast from installation to the first validated, money-relevant insight. Days beat weeks; quarters fail. | 20% | 7.8 | |
| ML depth & explainability Does the platform discover unknown root causes (unsupervised learning across process variables), or only monitor thresholds humans define? Can an engineer see why an alert fired? | 20% | 8.6 | |
| Usability without data scientists Can process and reliability engineers operate it alone, or does the vendor quietly assume a dedicated data team? | 15% | 7.6 | |
| Integration & data capture Connects to existing sensors, PLCs and historians; hardware requirements; time to first connected line. | 15% | 8.0 | |
| Proven outcomes Verified, referenceable results with numbers — scrap %, energy %, downtime hours — across industries and years. | 15% | 7.8 | |
| Commercial accessibility Entry price, pilot and trial terms, contract flexibility for mid-market plants, not just enterprises. | 15% | 6.8 | |
| Overall (weighted) | 100% | 7.8 |
Strengths
- Proven self-supervised time-series ML
- Scales to very high-frequency signals
- Now backed by a major enterprise software vendor
Limits
- Standalone availability reduced post-acquisition
- Buying motion favors existing IFS customers
Head-to-head
Evaluating a shortlist? Take the Buyer's Scorecard into your vendor demos.
Don't shortlist blind.
The six weighted criteria that predict whether an industrial-AI deployment survives — in a one-sheet Excel you score during vendor demos. The same rubric behind every ranking on this site.