You are currently viewing Successfully implement an AI/ Cognitive solution

Successfully implement an AI/ Cognitive solution

Yesterday I recorded a video podcast with Derek Russell. As we were getting ready to do it, we delved into how our partnership has grown and the type of Digital transformation and AI projects we have embarked. CloudiX has enjoyed tremendous success in the Cloud Data & AI space and as of 2019 we became a Microsoft Cloud Gold Partner (SI). Microsoft has been great to us and Derek has been a great advocate for our practice and how we jointly influence customers. As the topic went into AI & IoT projects and our approach and success metrics around it, I thought it might be good to share our journey with regards to AI projects.

AI or rather Cognitive solutions in general have been trending in any COO/ CIO’s word cloud. In simplistic terms cognitive computing refers to the process of utilizing self-learning algorithms to assess and function the way a human brain works through data mining or pattern recognition along with a natural language processing piece of code.

The Approach

  • Not all business problems can be solved using cognitive practices. In general, it’s imperative to have a clear idea and understanding of what data you have, how clean the data is, what type of skill-sets you have in order to implement a cognitive solution for your business problem.
  • Also choosing the right problem and the approach will help realize the value from that solution and mandate whether you want to further invest in this solution and build it. Typically, we see many companies rush in to do a solution, rather than evaluating a POC/ MVP approach.
  • For us, a POC is loosely defined as showing a service works. i.e. I take your data, run it through my Business Solution Accelerator or stitch a small application using Azure Cloud Services and show a tangible output. There are no bells and whistles built into this… typically can be achieved in 2+ weeks.
  • An MVP is a technology demonstrator that you can take to your board. You can build a business case and take it to your business stakeholders to secure funding. This is where CloudiX excelsCloudiX can be your trusted partner in working with your engineering team, understanding the domain and the business problem that you are facing, architecting the solution and implementing with Azure based cognitive services. We can typically achieve this within 6 weeks.

The process

During this journey, it helps us to identify the gaps that the organization has – typically we have seen that Managers tend to over inflate the expertise that they have in staff, their capabilities or even try to reach on their data quality. All of this can be identified, and remediable actions taken. Once the MVP is successful, most of the times these go into production, in a phased manner.

Typically the rate of success has been higher when we look at projects where we are able to transcribe or translate and do data recognition and validation and use this transformed data to feed into down level processes that users can use and make their life easier – be it elastic search/ be in updates on market condition or troubleshooting or FAQs or general intelligence on solving the problem at hand has been the most successful.

What are some implementations to consider?

Now coming to successful implementations – as I said before – we see businesses wanting to solve the sky. Ideally, I would recommend AI projects, where you can automate your typical BPO activities. i.e. look at the monotonous repetitive work and explore how it can be automated and can be done quicker and in a more efficient manner, thus helping reduce the rate of failure, increasing the rate of success and speeding time of delivery to your business constituents.

Some of the successful projects have been around:

  • trying to drive efficiency in your call center operations,
  • when you want to improve the quality of Natural language and conversational abilities for your customer interactive bots (could be Support bots, information bots, order tracking bots, scheduling bots),
  • data extraction processes; for example – ensuring compliance in contracts, validity of charges in invoices, extracting financial data from statements, co-relating invoicing to contracts and helping with dispute tracking,
  • planning for future business expansion based on contracts, ML based prediction for customer behaviors – these are some of the activities that our teams are actively working on.

Hope you enjoyed it and I look to writing such snippets from time to time. If you are interested or would like to learn more about our services, please feel free to drop me a note or comment and would be happy to engage and demo with a variety of solution accelerators that we have built.

Leave a Reply