Nearly half the CIO confirm that they are planning to deploy artificial intelligence (AI) in next 12 months. Hence as we enter 2021, This year I am combining 7 habit of highly effective people, trends in AI with AI in action stories.
AI Organization in 2021 can be put three stages as
Dependence: Here the organization depend on others to provide AI insights, or sometimes offload complete process to some external company. They say you take care of finding insights in our operation and deliver key results or if you do not deliver, we blame you.
Independence: Some teams in these organization are free from external influence they can built ai models and get ai insights, this is the goal of many organizations and preached as target for many, but it not the only goal in effective AI organization. There are far more mature and advanced levels.
Third and highest is interdependence: AI people in these organization include top executives in strategy and funding and diverse team of project managers, strategists, and application designers to provide diversity for thoughts, understand AI ethics and value of AI to customers for biggest measurable impact.
Below First 3 habits are intended to achieve independence, the next three are intended to help achieve interdependence and final one is intended to help maintain these achievements.
1. Be Proactive: Look through your organization and use your circle of influence to respond positively to deliver future AI experiences by
a. Hyper Automation: Here the idea is that anything that can be automated should be automated. Eliminate patchwork of technologies and focus on building services that are lean, optimized, connected and explicit, this will allow you to enhance speed and democratize AI
for example customer in hospitality are using app that provide user to check in on app, order, send recommendation this will generate relevant data to build tailored experiences.
2. Begin with the end in mind: Think through and know what you want in the future. Understand how decisions are made today internally and by customers/users and how can AI impact them, you can do this by
a. Internet of behavior (IoB): AI can be used to understand behaviors, you can combine data from customers data, citizens data from public sector, social media and location tracking and get deep insights into customers behaviors with small data to explain causation.
b. Total Experience: You can combine experience of User, Customer and Employees to provide a more cohesive offering to build competitive advantages,
for example in store curbside pickup and ordering app can allow screen sharing with store employees to allow enhanced experience to customers to find relevant items, and for employees to improve user experience and bugs discovery. With above data using generative AI organization can create and test on generated user persona`s in future rather than on real customers.
3. First thing first: Once you know what you want you need to prioritize here is a good sample to get started in year 2021
4. Think Win-Win: Efficient AI organization focus on making others successful by sharing data, insights, and best practices by enabling,
a. Anywhere operation: “digital first — remote first” deliver digital first experiences to support customers, employees, and business partners in physically remote environments. Building these will need your organization to put all your structured(tables)and unstructured data (picture, logs, audio, videos) at one place called data lakes/warehouses.
For example, an enterprise communication app allows users to chat in groups, have shared location to share files and allow video calls created a data lake where all structured (user profiles) as well as non-structured data (logs, images, videos). From data lake they are empowering everyone in organization to get data for training AI.
b. Distributed Cloud: Centralization of data in public cloud is facing challenges to deliver low cost, required latency and meet local laws for data governance. You need to ensure that your cloud provider can allow runing services closer to data with same management experience as public cloud.
For example, one healthcare company is using all the patient data using on premise compute and storage so their data never leaves their physical boundary to meet local laws and customer privacy, while their engineers can use same application and services that they are using in public cloud to extract, clean, train and deploy AI applications.
5. Seek First to be understand and then to be understood: You need to use empathetic listening and genuinely understand your customers with open mind to build impactful AI as
a. Intelligent composable business: Are the businesses that can adapt and fundamentally rearrange based on current situation. To deliver autonomy and democratization across the organization, ensure data is centrally managed and accessed from data lakes, warehouses, to stay agile and repurpose ensure AI workloads are containerized and orchestrated as micro services and not monolithic app.
For example, a CRM company contact tracing AI app used to develop sales leads was deployed to help government do contact tracing for COVID-19 bringing benefits to citizens and healthcare workers. This also made AI app train on Scale and established them as an AI leader.
6. Synergize: Efficient AI organization combine strength of data and ml engineer with data scientist and involve product and project managers to achieve AI goals that no one can achieve alone by allowing
a. Privacy Enhancing Computation: AI efficient organization need a trusted environment to process and analyze data, support decentralized processing along with encryption capability before processing. This will enable organization to collaborate securely across regions and with competitors without sacrificing confidentiality
for example with Open data initiative leading enterprise app provider are following a common data model to ingest data from SAP, Adobe, and Dynamics 365 to unlock AI efficiency that cannot be imagined working alone.
Sharpen the Saw; Growth
For this one AI organization need to continuously renew process, data sources and create sustainable long term effective pipelines by deploying.
7. AI Engineering: Here AI need to be part of mainstream DevOps Process to bring together disciplines of data, Machine Learning, Software Engineering. Consider responsible AI practices to address trust, transparency, ethics, fairness, interpretability, and compliance.
For example, once limited and expensive on premise data warehouses are changed by cheaper cloud scalable warehouses do you really need data lakes anymore? or If you can build face recognition model is it ethical to deploy it to watch all citizens?