At Ascent, we work on some pretty cool engineering projects.
But the development and implementation of a project is one of the last waypoints in the lifecycle of an idea. The journey starts much earlier, and our customers have to cover a lot of ground before their idea becomes a real-world solution. There are stakeholders to enrol, budgets to be won and business impact to be assessed in addition to the fundamental technical questions: ‘Is it feasible? Is it realistic? Will the technology deliver what we need? Will the solution solve the business problem? Will it work?’.
This is where the Proof of Concept (PoC) comes in. It’s a time-bound (usually <12 weeks), fixed scope exercise to test the feasibility of a concept against a clear set of success metrics, and we do a lot of them - to support our customers in bringing their ideas to life, and to help them gain business buy-in and investment. To us, a great PoC comprises lots of things - but for the customer, it really just has to be conclusive and affordable. And affordable Proof of Concepts are only possible because of the benefits that cloud ecosystems like Azure bring to the table.
Business value first and foremost.
When running a PoC we want to be able to focus on the business value that can be derived from the implementation of the solution. Platform as a Service offerings on Microsoft Azure make this possible because the technical heavy lifting has already been done by Microsoft in the background. Gone are the days where you had to install and configure servers before you could start development. With services like Azure SQL, App Service and AKS, we’re up and running in minutes.
Some services have evolved into even more powerful versions of themselves and allow us to achieve functionality that was only previously possible with months of development. Technologies like Azure Cognitive Search, an evolution of Azure Search, has augmented the capabilities of these services with AI features that are readily applicable to business problems. Pre-trained models for common business problems like invoice digitisation are available as ready-made building blocks that can be woven into a solution with ease. Of course, these can always be customised, but in the context of a PoC, this can save a lot of expensive time and get you to the next decision point with more confidence.
Azure Cognitive Services in action.
Challenged with a huge influx of emails hitting their public facing inbox during the pandemic, Customer A wanted a solution to automate the classification and redirection of incoming requests into their relevant departments.
This inbox, managed by a handful of individuals was reliant on their business domain knowledge for this manual sorting to take place. The proposed solution was to leverage Azure Cognitive Services to classify emails based on their content. We delivered a PoC that increased the confidence in the solution across the board, surfaced persistent issues and reduced the risk of progressing the project to the next phase and developing an MVP (Minimum Viable Product).
Interested in simplifying their fulfilment procedure for quotes, Customer B realised that considerable time could be saved through automation. We used Form Recogniser with AI capabilities to process incoming quotes in PDF format and compile an Excel sheet with the order items.
The PoC determined that the quality of scanned PDFs did not meet the level of accuracy needed to fully automate the process. Instead, the project would need to focus on building an AI-assisted process. Having such clarity with a small initial investment, as opposed to halfway through an expensive project, can really make the difference between a successful initiative and wasting both time and money.
Efficient subscription and tiering options.
Development and Test subscriptions can also help drive down the cost of development and cost of ownership for a PoC. This special pricing option removes the production-grade SLAs that Microsoft offer on their standard subscriptions, which is less of an issue in a PoC. This option allows your dev teams to spin up more environments and play around with expensive resources at a fraction of the cost. The tiered approach to provisioning of each service can also be leveraged during PoC development. Unless the higher tier is relevant for the concept under test, we can typically develop against the basic tiers and keep the consumption bill very low during development.
Although it might feel counter-intuitive, we develop PoCs using Infrastructure as Code templates and release pipelines to provision and deploy the solution. Azure DevOps has made this so easy for developers that the overhead of configuring these pays off even from short development cycles of a few weeks. This allows the customer to also start getting used to a modern development environment, governed by a DevOps mindset where the time from idea to production is measured in days not months – and this to us is what success is all about.
We’re always interested in opportunities to share our expertise. If you have an idea or a project in mind, get in touch - we’d be delighted to help.