Read More in the Full Ebook
How the Right Operating Model Will Help Upskill Your Data & Analytics Workforce
Know the 5 Operating Models for AI Initiatives
CLICK ON EACH CARD TO DISCOVER MORE
Decentralized / Siloed
1
©2022 Dataiku. All rights reserved.
Legal Stuff
Privacy Policy
The top two blockers for scaling AI are hiring people with AI skills and identifying good business cases. To address both issues at once, build teams made up of both data and domain experts, plus evolve the operating model for AI initiatives over time. This ebook demonstrates how to execute on this winning combination.
See the chart below for an overview of operating models by maturity to know what model is best for your organization.
AI Maturity + Associated Operating Model
Individual teams within a company do their own independent experimentation with AI.
Little to no sharing of infrastructure, data, best practices, or talent.
Almost always a temporary operating model, where the goal is to determine if there’s enough AI value to further invest.
Decentralized / Siloed
CENTRALIZED CoE
hub & spoke
center for acceleration
embedded
1
Decentralized / Siloed
•
•
•
What are the entry and exit requirements?
None
Identify many AI use cases and practices worth investing in.
Desire to increase reuse, decrease duplicate work, or generate economies of scale.
centralized center of excellence
2
hub & spoke
3
Center for acceleration
4
embedded
5
What are the entry and exit requirements?
Executive sponsor, measurable goals, champions in business units and functions, and funding.
A large, valuable backlog. Business units and functions wish to go faster.
2
CENTRALIZED CENTER OF EXCELLENCE
Known as a CoE, this operating model is designed to jumpstart the adoption of AI within an organization.
A centralized team develops and maintains AI products for many business units and functions.
•
•
Decentralized / Siloed
CENTRALIZED CoE
hub & spoke
center for acceleration
embedded
What are the entry and exit requirements?
Product owners in each business unit and function who prioritize domain-level needs.
Business units and functions wish to go faster, develop more AI products, and upskill frontline domain experts.
3
hub and spoke
CoE functions get distributed around an organization: AI experts (advanced, graduate-level data scientists) are in the hub, business units and functions are in the spokes, and they collaborate on product development.
Like in a CoE, the hub is responsible for infrastructure, standards, and tracking industry innovation.
However, ownership of AI products shifts to the spokes.
•
•
•
Decentralized / Siloed
CENTRALIZED CoE
hub & spoke
center for acceleration
embedded
What are the entry and exit requirements?
Best practices, rules, and governance in each business unit and function.
A company might decrease embedding and increase centralization to reduce costs or consolidate vendors.
5
embedded
Very few central, shared resources and rules such as Responsible AI guidelines, infrastructure, and a few common, curated datasets.
The most decentralized, agile, and innovative structure since many business units and functions are involved, and they are loosely connected by rules and resources.
•
•
Decentralized / Siloed
CENTRALIZED CoE
hub & spoke
center for acceleration
embedded
KEY TASKS OF A CoE
Manage a Portfolio of AI Products and Prioritize a Backlog
1
2
Create a Scalable Data Architecture and Infrastructure
3
Keep Track of AI Industry Innovation
4
Develop Champions in Each Business Unit and Function
5
Capture and Evangelize AI Value Stories
What are the entry and exit requirements?
Leadership’s support to develop a data & AI strategy that empowers every domain to make its own AI investments.
Deep development skills in each business unit and function. Business units and functions wish to go faster still.
4
ceNTER for acceleration
Decentralized / Siloed
CENTRALIZED CoE
hub & spoke
center for acceleration
embedded
KEY BENEFITS OF A CENTER FOR ACCELERATION
3
Increased AI use case ROI since the business invests their own time and effort
2
Increased agility since more roles, business units, and business functions are involved
1
Increased innovation since domain experts directly participate in development
Designed for broad AI adoption among frontline domain experts.
•
•
•
A refinement of Hub and Spoke and Dataiku’s recommendation for organizations who already have a mature CoE.
Shifts the onus for AI product development out of the center and into business units and functions and aims to create unicorn teams in every spoke.
If each team within your organization is getting started by doing their own independent experimentation with AI, you’re probably organized as a decentralized model. To move to a CoE from here, you need to have a desire to increase reuse, decrease duplicate work, or generate economies of scale.
Decentralized / Siloed:
Decentralized
explore
business value
time
educating
getting started
scaling
transforming
If you’re already set up as a CoE model, in order to mature and move to a Hub and Spoke model, you need a large backlog of projects and business units and functions that wish to go faster.
Centralized Center of Excellence:
If you’re already set up as a Hub and Spoke model, in order to mature and move to a Center for Acceleration, you need business units and functions that wish to go faster, along with the ability to develop more AI products and upskill frontline domain experts.
Hub and Spoke:
If you’re already set up as a Center for Acceleration, in order to mature and move to an Embedded model, you need deep development skills in each business unit and function and those business units and functions wish to go faster still.
Center for Acceleration:
If you’re already set up as an Embedded model, you’re at the most decentralized, agile, and innovative structure since many business units and functions are involved, and they are loosely connected by rules and resources.
Over time, your company may decrease embedding and increase centralization to reduce costs or consolidate vendors.
Embedded:
explore
Decentralized
CLICK ON EACH DOT TO DISCOVER MORE
experiment
Decentralized
experiment
CENTRALIZED CoE
establish
Decentralized
establish
Hub and Spoke
expand
Decentralized
expand
Center for Acceleration
embed
Decentralized
embed
Embedded
everyday ai
Companies that have scaled AI are 3x more likely than average to use a hub and spoke organizational structure, according to the Harvard Business Review.
Enter
Exit
Enter
Exit
Enter
Exit
Enter
Exit
Enter
Exit
Get the Full "Upskilling: How to Win the Battle for Data + AI Talent" Ebook
AI in 2023: Get the Full Ebook
Enter
Exit
Executive sponsor, measurable goals, champions in business units and functions, and funding.
A large, valuable backlog. Business units and functions wish to go faster.
What are the entry and exit requirements?
2
CENTRALIZED CENTER OF EXCELLENCE
KEY TASKS OF A CoE
Keep Track of AI Industry Innovation
Create a Scalable Data Architecture and Infrastructure
Manage a Portfolio of AI Products and Prioritize a Backlog
3
2
1
Known as a CoE, this operating model is designed to jumpstart the adoption of AI within an organization.
A centralized team develops and maintains
AI products for many business units
and functions.
•
•
What are the entry and exit requirements?
Enter
Exit
None
Identify many
AI use cases and practices worth investing in.
Desire to increase reuse, decrease duplicate work,
or generate economies of scale.
1
Decentralized / Siloed
Individual teams within a company do their own independent experimentation with AI.
Little to no sharing of infrastructure, data, best practices, or talent.
Almost always a temporary operating model, where the goal is to determine if there’s enough AI value to further invest.
•
•
•
If each team within your organization is getting started by doing their own independent experimentation with AI, you’re probably organized as
a decentralized model. To move to
a CoE from here, you need to have
a desire to increase reuse, decrease duplicate work, or generate economies of scale.
Decentralized / Siloed:
Centralized Center of Excellence:
If you’re already set up as a CoE model, in order to mature and move to a Hub and Spoke model, you need a large backlog of projects and business units and functions that wish to go faster.
experiment
CENTRALIZED CoE
If you’re already set up as a Hub and Spoke model, in order to mature and move to a Center for Acceleration,
you need business units and functions that wish to go faster, along with the ability to develop more AI products and upskill frontline domain experts.
Hub and Spoke:
establish
Hub and Spoke
Center for Acceleration:
If you’re already set up as a Center for Acceleration, in order to mature and move to an Embedded model, you need deep development skills
in each business unit and function and those business units and functions wish to go faster still.
expand
Center for Acceleration
Embedded:
If you’re already set up as an Embedded model, you’re at the most decentralized, agile, and innovative structure since many business units and functions are involved, and they are loosely connected by rules
and resources.
Over time, your company may decrease embedding and increase centralization to reduce costs or consolidate vendors.
embed
Embedded
©2022 Dataiku. All rights reserved.
Privacy Policy
Legal Stuff
Back to top
Decentralized / Siloed
CENTRALIZED CoE
hub & spoke
center for acceleration
embedded
explore
experiment
establish
expand
embed
Decentralized / Siloed
CENTRALIZED CoE
hub & spoke
center for acceleration
embedded
Decentralized / Siloed
CENTRALIZED CoE
hub & spoke
center for acceleration
embedded
Decentralized / Siloed
CENTRALIZED CoE
hub & spoke
center for acceleration
embedded
Decentralized / Siloed
CENTRALIZED CoE
hub & spoke
center for acceleration
embedded
Decentralized / Siloed
CENTRALIZED CoE
hub & spoke
center for acceleration
embedded
Decentralized / Siloed
CENTRALIZED CoE
hub & spoke
center for acceleration
embedded
Decentralized / Siloed
CENTRALIZED CoE
hub & spoke
center for acceleration
embedded
Decentralized / Siloed
CENTRALIZED CoE
hub & spoke
center for acceleration
embedded