So You Want to Build an AI Solution?
App Dev Manager Rich Maines shares insights from recent engagements with customers using Azure Machine Learning.
I recently engaged a customer on building out various machine learning scenarios using Machine Learning Studio (classic). I was impressed with the speed at which our consultants and engineers were able to build out solutions alongside our customer’s technical resources. Azure Machine Learning has quickly become the de facto tool for those requiring Python and R SDKs and the “drag-and-drop” designer to build and deploy machine learning models. Rapid development and deployment of AI-infused services has never been easier.
The most valuable lesson from our recent work was the time invested in identifying opportunities on how best to leverage AI to enhance our customer’s solutions before delving into architectural decisions and proceeding with a proof of concept. What follows is a blueprint of sorts, on how Microsoft Premier Developer engages our customers when building strategic solutions to identify AI scenarios that are relevant to your business objectives, drive readiness for your developers through education services, and provide best practices for designing and implementing Microsoft AI services while providing ongoing consultative guidance.
Customer Business Overview and Strategy
Before we dive into architecture and technical readiness, we must first understand the customer’s needs. Understanding their business overview, strategy, market/industry forces and technical maturity allows us to not only gain a deeper understanding of our partners/customers, it provides a holistic view of the customer.
Review of Existing Offerings and Solutions
We then dive deeper into their current solutions. Typically we find that customers already have a good idea of how they would like to apply AI to existing/new solutions, but it’s helpful to explore other ideas to refine/correct initial assumptions.
Roadmap, Gaps and Needs
By now, we have a better understanding of the underlying motivations, so how likely are the initial ideas to be marketable? What problems are they attempting to solve? If it’s a new offering, what does their customer pipeline look like?
Candidate Scenarios List
Here’s where the magic happens. We work with the customer on creating an initial list of potential use cases, or at the very least, refining and expanding upon the ideas they brought to us. Below is an example of one of our exercises:
AI Vision, Strategy and Platform
We can now provide an overview of the Microsoft AI Platform with demos focusing on Azure Machine Learning, Cognitive Services, bots and other solutions.
The remaining time is spent on deep dives on:
- Microsoft’s Approach to Ethical AI
- Design Principles and Best Practices to Drive Adoption
- Customer specific deep dives where we expand on the scenario list created earlier
- Planning and Next Steps where we scope, plan to build PoCs and determine ongoing cadence (4-6 week sprints)
Everything we do is guided by our values and principles:
- Enable people – we don’t believe AI will replace humans, but rather, amplify their ingenuity
- Inclusive design – technology should help lower the barrier to entry for everyone – AI has the power to bring technology to more people at a vastly different scale
- Build trust in technology –the entire technology industry should come together to create a set of principles and ethics that drive AI development in an ethical way
So while your developers may learn how to develop, deploy and support applications using Azure Machine Learning, your organization is best served by working closely with your development team on identifying scenarios to take full advantage of the platform. If you wish to accelerate the adoption of these technologies, direct engagement with a Microsoft Premier Developer consultant will ensure a comprehensive strategy that provides practical experience with guided scenarios and a complete end-to-end solution.
Light
Dark
0 comments