Digital Transformation – Crossing the Proof of Concept (PoC) Chasm

“Make digital transformation tangible” – was the key message at Hannover Messe this year.  The world’s leading industrial technology showcase brought over 6500 exhibitors and organized discussions on topics ranging from industrial AI, production and logistics digitization to 5G applications. While the technology demos from exhibitors looked impressive the key questions for manufacturing executives remain:

• Are these technologies capable of crossing the POC stage to scale up to production deployments?
• Are they easy to use for my operations teams?
• Can they easily integrate with our existing infrastructure and assets?
• Will they deliver a clear ROI and payback?

Preventing IIoT PoCs from Stalling

IHS Markit analysts presented their findings from a recent survey on adoption of Industrial IoT. They found that while , Automotive and Oil & Gas industries are leading in Proof of Concept (PoC) projects, project failure is common.

We discussed the reasons behind why crossing the POC stage for AI and data science PoCs is difficult in Operational Machine Learning vs. Data Science Projects. Putting “ready-to-use”machine learning products in the hands of your operations teams, who have industrial domain expertise has been shown to address these challenges and successfully lead to production deployments. Such an approach is effective because:

• The operations SMEs retain both the control and visibility needed to drive improvements.
• The SMEs are in the best position to determine the corrective action to take.
• The project delivers a shorter time to value since the entire process, from identifying the use case to verifying the results, resides within the operations team

Crossing the POC stage by applying an Edge Strategy to Manufacturing

IHS analysts also discussed the business impact of AI, the importance of time sensitive networks (TSN) and edge strategies in manufacturing.

Falkonry team demonstrated the recently launched Edge Analyzer in the Siemens booth at the Hannover Messe. Edge Analyzer is a portable, self-contained engine that enables predictive models to be deployed in remote, resource constrained environments.

Edge Analyzer adds intelligence to devices at the edge, enables use cases where data needs to be analyzed in real-time and offers powerful benefits to industrial customers:

• Increases efficiency & productivity by eliminating latency for time-sensitive applications
• Improves reliability, resiliency, compliance and security with local processing vs. remote
• Reduces operational cost by filtering data,  sending only useful, aggregated data to central location

These can be important factors in crossing the POC stage and getting to production.

Falkonry experts showed how to build prediction models using Falkonry LRS, deploy them on lightweight edge devices for real-time prediction and explanation.

Here is the link to our demo from Hannover Messe: Click here for Demo Video

Takeaway

The key message for operations executives, manufacturing heads and chief technology officers is clear: think about what happens after crossing the PoC stage from the start. Pay attention to common challenges – scalability, integration with existing infrastructure and deployment options. Leverage edge architecture for best cost, latency and security for distributed low-latency applications. It is important to think about change management – will the operations teams use new product?  Are the new tools easy to use, will they capture SME’s tribal knowledge and foster collaboration? Finally look out for partners who have successfully deployed AI solutions in the production environment for multiple use cases. That’s how you will be able to address challenges, cross the PoC chasm and make digital transformation tangible.