As it stands today, companies around the world struggle to turn business data into business value.
This article covers seven tangible ways that an organization can ensure scalable and repeatable business value from its analytics and AI projects.
Start Explicitly Measuring and Tracking Value
AI adoption can help companies deliver strategic business value — the problem is there isn’t always a clear way to measure or track said value.
One way to change that is to assign an owner of value tracking. Often a business lead, this person should commit to measuring the financial impact of these pre-identified use cases every fiscal quarter. Their remit is to coach and evangelize the value of AI for the organization.
Have a Concrete Vision + Pipeline of Use Cases
This step is all about getting past the noise of experimentation and initial use case successes and onto short-, medium-, and long-term visions for AI at scale.
1. Generate More Ideas (and More Value):
Establish a process that allows both tactical and strategic projects to emerge from shared efforts. When this reuse becomes commonplace, organizations can uncover hidden use cases and capitalize on existing project work to spin up new ones.
Dataiku customer Orange, one of the largest operators of mobile and internet services in Europe and Africa, had a data science team that used to perform mostly ad-hoc analyses for the business and had limited ability to work on more complex machine learning (ML)-based projects. In order to generate demand and create a healthy use case pipeline, the company:
- Enabled 100+ analysts and business users with Dataiku in order to work on their own data analysis projects
- Freed up data scientists' time for more advanced work with potentially bigger impact
- Went from one request per year from the business for ML projects to several per month
2. Set Up Executive Workshops:
To ensure AI isn’t left in the hands of an elite few — such as IT exclusively, or those who lead the company’s analytics and AI initiatives (i.e., Chief Data Officer, Chief Data Scientist) — these leaders should set up workshops for the rest of the executive team to understand the role they will play in democratizing AI within the enterprise.
Dataiku customer GE Aviation offers a full-day training for executives that is focused on the value proposition of the company’s self-service data program and how/why teams and individuals partake.
3. Remember That Transformation Takes Time
Teams should aim to set up their AI capabilities the right way and in the shortest amount of time (without compromising steps along the way). In addition to evolving their operating model according to their AI maturity, organizations must be dedicated to implementing change management.
For example, Chris Kakkanatt at Pfizer has said that it took the company about a decade to move from theory to practice when it comes to achieving a truly human-centric, AI-driven transformation. By creating a culture of collaboration and co-creation around analytics, Kakkanatt was able to orchestrate the organizational and cultural changes necessary to connect technologies and people around the world at scale.
Adopt a Platform Mindset
How can teams pinpoint the right technologies and processes to enable the use of AI at scale?
An end-to-end platform (like Dataiku) brings cohesion across the steps of the analytics and AI project lifecycle and provides a consistent look, feel, and approach as teams move through those steps.
When building a modern AI platform strategy in the era of GenAI, it’s important to consider the value of an all-in-one platform for everything from data prep to monitoring ML models in production to scaling GenAI. Buying separate tools for each component, conversely, can be tremendously challenging as there are multiple pieces of the puzzle across different areas of the lifecycle.
Have a Global Understanding of the Costs
It’s impossible to measure value without having a handle on the costs of AI. A great step for organizations to take is periodically conducting a total cost of ownership (TCO) exercise for their existing AI tooling.
For example, Dataiku manages the entire analytics and AI pipeline for teams in a way that is flexible, collaborative, and governable. It minimizes the overall number of tools by offering one truly comprehensive platform, avoiding the need to cobble together multiple tools for ETL, model building, operationalization, GenAI, and so on. It is completely server-based, removing the cost of maintenance of desktop-based products and also provides a clear upgrade path with no need to transition platforms or migrate in the future.
Avoid Technical Debt at All Costs
This accumulation of messy code, aging systems, and temporary patches that will need to be fixed later is costly anywhere in an organization. Within an analytics and AI stack, though, it can be the difference between being able to generate an insight that drives millions of dollars of ROI and simply continuing with business as usual.
Being locked into a non future-proof solution means significant upgrade costs in the future and limited infrastructure options, both of which can hinder growth. Further, taking the build route for an AI platform not only piles on technical debt but is complex in nature as it involves cobbling together features outside of the core functionality of building a model which is not a smooth process.
To navigate this, analytics and IT leaders should focus on future-proof technology that is tech stack agnostic and open enough so that change remains an option.
Get Business People to Create Value (Not Just Data People)
Teams need to ensure that data and domain experts work together to co-build analytics and AI projects. Here are some initial steps to do so:
- Set up a bespoke upskilling program for different competency levels that involves both a technology perspective and applied perspective (i.e., how to actually use certain tools to solve real business problems)
- Establish AI governance processes to allow more and more autonomy from non-experts
- Provide the right tools that drive and support tight collaboration
- Create organizational structures and operating models for AI that support the upskilling of business people with data
- Carve out any job roles they need know and will need in the future, identifying any gaps and determining if they have anyone who can fill the roles internally
Build an Effective AI Governance Program
AI can very much be like Pandora’s box if organizations don’t know what they’re doing. It’s only with AI governance (in tandem with responsible AI and unified AIOps) that they will effectively be able to de-risk what comes out of that box.
A strong AI governance framework should:
- Centralize, prioritize, and standardize rules, requirements, and processes aligned to an organization’s priorities and values
- Inform operations teams, giving build and deployment teams the right parameters to deliver
- Grant oversight to business stakeholders so they can ensure that AI projects (with or without models) conform to requirements set out by the company
Dataiku allows people across an organization to access all data and work together on projects in a central location, fostering strong governance practices combined with deep collaboration.
Putting It All Together
While business value creation will ultimately end up looking different at every organization, at its core, it comes down to improving business outcomes to make key processes and functions better, faster, or more cost effective.
Whether an organization is just setting out to improve and accelerate its AI maturity (and therefore value generated) or is determining how to scale a specific element such as governance, following these seven steps will ensure proper coverage when it comes to measuring and tracking value from analytics and AI projects.