Product

Predictive Solutions: Moving Maintenance from Unplanned to Planned

PREDIKTO ENTERPRISE SUITE

Predikto Enterprise Suite is our SaaS pipeline turning your raw data into actionable outputs.

The platform is designed for complex, distributed systems. Our cloud-based service mean you avoid unnecessary IT overhead, and we can quickly scale to your needs.

process

Journey from Data Silos to Proactive Operations

Now that you have a plan, let’s execute.

Our software and services provide the entire data analytics hierarchy. Delivering results and value at each step along the way.

Let’s follow a fictional company on a journey to operationalization.

Fast Fit makes treadmills with two offerings:
  1. Professional grade to gyms, with SLAs
  2. Personal grade to consumers, with warranty service

Due to bulk of their equipment, all service is on-site by a technician.

They would like to reduce maintenance costs and increase customer satisfaction.

process

Problem & Data Identification

Fast Fit:
  1. Unsure of maintenance cost-drivers
  2. Siloed data sets
Works with Predikto through the Digital Transformation Checklist.
  1. Identifies overall strategy
  2. Identifies quick-win pilot: “repeat service visits and broken drive trains”

Predikto kicks off pilot and transfers Fast Fit data to Predikto’s Enterprise Suite, a cloud-based SaaS.

process

Data Consolidation

Predikto:
  1. Predikto receives the data in it’s online intake environment
  2. Predikto automates the ETL process

Predikto’s back-end software enables it’s employees to set up and monitor these processes in a fast and repeatable manner.

process
process

Business Intelligence Insights

Joined data in Predikto’s Data Lake
  1. Business Intelligence via Predikto RECON’s Data Lake Interface
Fast Fit learns:
  1. Coverage gap in their service tracking Apps in March and April of 2017
  2. Devices with serial numbers in the 200s break twice as often
  3. Anomalous data from their installed temperature sensors
process
process

AXIOM: Rules Engine

AXIOM takes over to create basic physics rules.
  1. Creates rules based on known faults
  2. Useful for identifying emergent faults that need to be assessed quickly

Approximately 3 weeks from the receipt of data, Predikto can present a set of viable rules

process
process

Exploratory Analytics

The data from Predikto Axiom is delivered via Predikto Recon. Rule results, watch lists, and underlying data are available, and updated nightly.

Fast Fit learns:
  1. Drive train should be replaced after 3000 hours of use
  2. Use of the wireless feature causes premature failure of the console

They update their suggested maintenance plans, and their engineering team investigates a new wireless integration.

process
process

Feature Creation

Predikto Enterprise Suite’s MAX Machine Learning Engine:
  1. Creating and scoring thousands of potential features
  2. One data scientist can do the work of ten

Example features: number of days since a technician has visited, average usage per day, temperature variation during a single day.

process
process

Inferred Relationships

Features result in:
  1. Correlations between incidents and indicators
  2. Available in RECON and also to download
Fast Fit discovers failures typically occur when:
  1. The temperature variation is extreme
  2. The unit is used too infrequently (i.e. home use as a laundry hanger)
process
process

Modeling

Predikto MAX uses the scored features
  1. Score and select models to predict failure events
  2. Models are made with lead-time in mind
process
process

Predictive Analytics

Predikto issues predictions:
  1. Delivered via Predikto RECON: interpretations, asset status, knowledgebase
  2. Contextualization enables trust and action on predictions
process
process

Interpretable Models

Predikto has the capability to interpret all predictions made by our platform.

We provide the top features that influenced that specific prediction

By understanding signals:
  1. Users gain trust
  2. Associate signals with failure
  3. Support decisions to act or not act on a prediction
process
process

Operationalize & Live Data

Operationalize:
  1. Turn reactive to proactive through updated procedures
  2. Set up live data feeds
Fast Fit & Predikto:
  1. Set up a maintenance pipeline via Predikto’s APIs
  2. Stream updates and scheduling Fast Fit’s maintenance app
  3. Onsite technicians preemptively fix co-located equipment
  4. Update owner portal with tips & tricks
  5. Update temperature sensor and wireless connection point installed with new units
process
process

Autodynamic Learning

Once in production, Predikto:
  1. Ingests live data feed
  2. Issues daily predictions
  3. Automatically searches for new features and models to keep models running at peak performance
process

Built for every level of your business

We work with and inform decisions for a variety of end users.

Operations / Engineers

  • When will an unplanned event occur?
  • What has happened recently?
  • How can I efficiently schedule proactive interventions?

Executive Management

  • How much waste have I eliminated?
  • What is my ROI?
  • Where can I improve processes?

Data Scientists

  • Where is my data?
  • Can I trust my data?
  • Can I trust the models?
  • What is model performance?
`
Start the journey to predictable operations today. Contact Us