The evolution of AI: 7 lessons learned for effective implementation

Robot teaching robot students

Artificial intelligence is one of the main topics of focus in the technology sector and in boardrooms at the moment. Its attraction comes in its ability to drive multiple benefits in parallel – whether enhancing customer experience, improving productivity, greater consistency or reducing time to process. Machine learning, deep learning, visual recognition, natural language processing – these are all capabilities in the AI domain that can help organisations take advantage of the explosion of structured and unstructured data. Knowing what and where to start can be daunting, but critical, in enabling an augmentation of human intelligence. Below are some of our learnings as we deliver cognitive projects for our customers:

7 lessons we have learned about implementing cognitive capabilities

  1. Be clear on the value – Don’t get lost in “cool” technology. Have clear line of sight to the value it needs to deliver for your organisation. For example, in a contact center, will it impact average call handling time, net promoter score, customer retention or a combination of some or all of the above? The value should also be understood in quantitative and qualitative terms.
  2. Give yourself permission to rethink business processes – Don’t just think that incremental improvements are the answer when a transformation could be in order. While cognitive capabilities could improve the way an existing call centre operates, the real opportunity could be enabling and augmenting a customer self service capability and subsequently eliminating the call process for certain transaction types.
  3. Organisational buy-in is key to success – AI comes with a certain trepidation – will my job be replaced? In order to get all levels of the organisation bought into a cognitive journey, it is best to translate “day-in-the-life-of” scenarios to show how cognitive will either replace, augment or enhance working patterns.
  4. Outline a value roadmap early – The vision of a new way of working will get people driving towards a common goal, but it is also important to deliver incremental value along the way. Our most successful projects have placed focused execution in context of a broader journey.
  5. Be user driven and technology supported – The cognitive age relies even more on business insight to drive value. Real business users should be on the project team to understand the problem, brainstorm the potential solutions, and test if value has been delivered. Most of our best insights and innovations come from real business. In the most effectively balanced teams, the user is the driver and the technology team is in the support role.
  6. Training is iterative – As we move from a programmatic and rules-based era to probabilistic one, training becomes key and continual. This requires focus from a subject matter expert to do periodic reviews and train the model, but the investment will pay off with greater confidence levels and automated decisioning.
  7. Capabilities are rapidly evolving – AI is such a rapidly evolving space, and it pays to stay on top of the latest capabilities to see how your existing or backlog use cases can be tackled in new and exciting ways.

In the experience we have gained over multiple cognitive initiatives for clients, we hear a common theme: “help me reduce time to value and show clear value.” Thus, many of the initial steps into AI journeys are often focused on somewhere between 4 and 8 week efforts.  We have realized the best results when organisations start small, create proof points, and expand quickly. This is also why it is important to understand the overall journey; so that the next steps after the initial piece is completed continue quickly. To help balance “time to value” with “practicality”, we answer the following questions:

  • What is the business problem are we solving? An example in a call center use case might be improving overall customer satisfaction while optimizing costs.
  • Who are we assisting?  Is it “the customer service agent” or “the end customer for self-service”?
  • What (or how) are we trying to help them with ? Are we improving the accuracy, consistency, and speed of answers for the requestor?
  • Where does the data come from to solve the problem? It might be a combination of structured data from back end admin systems, unstructured policy and product information, and externally available unstructured data providing industry insights.
  • Is there a sufficient business case to support the investment/expenditure? This is where the specificity comes in: how many agents, how much of a call reduction time can you achieve, where are the areas in the call that are opportunities to reduce time spent, how many calls can utilize self-service?
  • What are the Key Performance Indicators (KPIs) that the solution will impact? Will it increase customer satisfaction, drive new revenue, reduced call handling time, number transactions processed, etc.?

AI projects are able to deliver unique outcomes that are transformative in nature – thus their appeal – but they also have their own pitfalls and need to be approached accordingly. Besides, an AI project is normally only a step within a company’s broader digital transformation. Therefore, it is essential to embark on your AI journey with a team that has not only the required breadth and depth of skills, but also the right experience to guide you safely through the project and beyond.

Good luck with your AI journey!

This article is the fourth of a series about AI from Focus the Way Forward. Watch out for our next installment!


Evan Salop is currently the Cognitive Practice Partner for DXC, where he works with clients and offering teams to drive business outcomes through extending and enhancing applications with Cognitive capabilities.  Specifically, Evan defined the use case, business justification, solution, journey roadmap and implementation for Watson Cognitive API’s in an Insurance contact center.  He is actively expanding cognitive initiatives in Claims, Underwriting, London Markets, Banking Core Modernization, Travel and Transportation Irregular Operations and has been actively involved in the industrialization of an AI Chatbot to support Virtual Helpdesk

Andrew SlappAndrew Slapp is the Managing Partner for Emerging Solutions within DXC Technology ANZ. Andrew has a 15-year career in new technology innovation and commercialisation and is responsible for a range of solutions within DXC Technology, including Smart Spaces, Robotic Process Automation and AI.

RELATED LINKS

The evolution of AI: Why it’s finally time for business to get excited

The evolution of AI: Are chatbots driving a new customer experience?

Comments

  1. Criag Raj says:

    How are these lessons different from other IT implementations?

    Like

  2. Evan Salop says:

    Craig – great question. At the broadest level, one might infer that some of the bolded words can be applied to many Transformation projects. The description details lessons we have learned from specifically applying cognitive capabilities to achieve business outcomes. We have also learned that skipping/shortchanging some of these lessons can negatively impact desired outcomes. One other nuance – cognitive initiatives need to be driven by the Business/supported by IT, so while some lessons from IT implementations are relevant, we highlighted the specific ones that we learned have the most impact. Glad to discuss further with you and thanks for your question

    Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: