AI Journeys: Your Path to Adoption Depends on Your Business Context
Different AI Journeys for Different Contexts
Journey #1: AI Adoption for Automation
The AI Journey for Automation
Goal: Reduce the cost of known business operations
Examples: Speech to text transcription by AI instead of scribe
Key Activities:
Evaluate and select AI tools
Update processes
Evaluate quality
Configure tools
Deal with regressions
Stakeholders Involved In this Journey:
Beneficiaries:
Investors
If you do this well, everyone! Except for the competition that don't evolve
Customers
Negative beneficiaries:
If we do it poorly, expensive, the people who compensate for bad tools
No one left to help customers. Digital 1st
Decision Makers:
Those with power
Upper mgmt
Investors
Those with indirect influence
Project Managers
One would hope employees :)
The people who give the company its revenue
Researchers
Directly Responsible:
Tool researchers
Mentor Coaches for people who have to use tools (early adopters)
Supervisors to report on performance
Managers to advocate for the resources and time necessary
Gatekeepers:
CEO
Government
CIO/CFO
Money people!
General change resistors at high levels
Board Members
Community members
Customer groups
Risks:
Regressions, things that worked but no longer work
Elimination of variety / diversity of responses. The ability to respond.
Loss of control, inability to express proprietary & custom business logic
Loss of functionality / no backups
Security risks from pass data around
Rewards:
Save time, save money
Can help reduce human error
Repetitive work is offloaded
Typical Outcomes:
Good → Everyone is unburdened from tedium, save time and money for vaca
Bad → Spend a lot of money, fire a lot of people, realize its wrong, backtrack
Ugly → Bankruptcy, legal action
Journey #2: AI Adoption Journey for Insights
The AI Journey for Insights
Goal: Increase the frequency and quality of analyses when new decisions are needed
Examples: Search engines, BI, interactive predictive models
Key Activities:
Structure and organize data
Create databases
Develop visualizations, APIs, and query, UI/UX
Integrate new data and develop new UI/UX to give people ACCESS
Stakeholders Involved In this Journey
Beneficiaries:
People making big decisions especially strategic ones
Better alignment across departments. See what the ground truth is and the current state
Decision Makers:
Architects (they have to build big things, and care about design and insights before)
C suite
Directly Responsible:
Business Analysts
Developers
Data engineers
Data Scientists
Designers
Middle Managers
Gatekeepers:
Money people!
Finance
Middle Managers
Risks:
Room for error, input errors, black boxes. hard to do QA
Black swan events! How quickly can we update our models?
Multiple possible vendors of "truth"
Insufficient skill in validating models
Rewards:
Deliver features, functionality, and services that would not be viable
Increase alignment and clarity around how decisions are made
Standardize decision making processes
Increase agility for making new decisions. Better decision making
Done well, better algorithm mastery
Typical Outcomes:
Good → On demand insights, fresh and relevant and high quality HI + AI
Bad → Everything seems to work, but is wrong. Societal harm
Ugly → Modeling is premature, need to fix data first
Journey #3: AI Adoption Journey for Architecture
The AI Journey for Architecture
Goal: Create proprietary systems that support the company’s people, processes, and growth
Examples: A proprietary model trained on company data that creates outputs for customers
Key Activities:
Differentiate proprietary and commodity needs
Design and develop proprietary systems (find some architects)
Onboard and support users
Maintain systems. Refine features
Prepare for future states
Stakeholders Involved In this Journey
Beneficiaries:
The company, it has an ASSET!
Stakeholders:
Investors the value of your company grew
Customers, often the solution was created for their unique needs
Decision Makers:
Board
Champions on the C suite
Engineers, see what is possible. prototypers
CEO, CPO
Standards
Doers and the Customers the Users
Directly Responsible:
Prototypers
The business people
Architects and their teams of engineers and designers
End Users
Product M/O
DevOps
Gatekeepers:
Standards Legals
Finance
Risks:
Bad Architects took all my money
Overengineering. Poor resourcefulness
Solution doesn't match staff skills
Extensibility
Bad Prod Mgmt: users are mad
Failure to differentiate needs that common vs proprietary, or unique
Accidental multitracking to the end
Rewards:
Product(ivity). Customization. Incorporate everything you know
Typical Outcomes:
Good → Making happy the users -> $
Bad → Bad design by bad architects
Ugly → Accidental multitracking to the end
Journey #4: AI Adoption Journey for New Products
The AI Journey for New Products
Goal: Create and sell new features, products, and services that were not feasible
Examples: Google Maps, Google Photos automatic album generation
Key Activities:
Identify unmet user needs
Design AI that can receive delegated tasks
Develop UI/UX that incorporate AI seamlessly
Market value of new features
Get users to adopt features
Update pricing, partnerships, terms, etc...
Stakeholders Involved In this Journey
Beneficiaries:
The company's bottom line! New revenue lines
Stake and stock holders
End user!
Decision Makers:
C Suite, often CEO
CPO
The Board
Marketing Research, The Competition
Maybe even anyone... the people closest to the customers
Directly Responsible:
Lots of people
Engineers
Designers
SMEs people who know how to make stuff
End users collaborators
Everyone
Gatekeepers:
Finance
Legal
Certain C Suite
The Law
External Politics
The Board
Risks:
Too slow, paid to build, no adoption
It's bad, no one likes it
Insufficient change mgmt
Can't support a great idea
Static mindset
Run out of resources
Rewards:
Your company grows (new markets, new revenues)
New products proliferate! Discovery
Pivoting and learning to success!
Typical Outcomes:
Good→ Create a new market for your company, compete and win!
Bad→ Can't keep up, can't identify valuable needs to meet viably with products and services
Ugly→ Innovate new products and things don't go according to plan. need to pivot to something else more valuable
Driving These Paths to Adoption Depends on Your Business Context
You as an individual can get started on adopting AI but you will have to work with one or many people to see it to fruition.
The more complicated the AI solution, the more you will have to work with the other contexts.
The 4 Types of AI R&D Journeys