How to get from Proof of Concept to Generating value for your company with Generative AI
How to scale the pilots that you already have in place?
Working backwards from your business outcomes, working backwards from customers.
- What are our customer pain points?
- Where are our employees struggling?
- What are the blockers on the business?
- What specific outcomes are we aiming for?
Think of the why before the how.
While the accessibility of AI can spark quick action, working in silos can lead to partial or incomplete solutions. Rahul Pathak VP, GenAI & AI/ML, AWS
If you are siloed you need to figure out how to get information from one team to the other.
Creating a cross functional team you allow people who understand the business to work with people who understand technology.
Why companies don’t do it?
Office politics / policies and power struggles are generally the underlying motive for not allowing cross functional teams to exist.
Everything you do has to work back from the customer.
Decision making framework for AI
The most valueable asset: your data
You need to have good quality, reliable data to feed to AI applications. The old statement “garbage in garbage out” still applies also to AI.
Data strategy includes what data your need, where to store it, how to collect it.
It’s not just the amount of data, is also the quality.
Next step is integration
Successful AI strategy requires seamless system integration.
You want to build a flexible architecture that is quite easy to integrate and scale.
How data foundation begins
Data foundation begins by analysing deep institutional knowledge and getting it into knowledge bases. The knowlede base becomes the foundation for the transformation. It’s a change process.
Discipline equals freedom
Having a clear governance and well defined guidelines and policies creates freedom to act and improve on AI usage.
Good governance sets people free to innovate.
Se
Security is the enabler of innovation.
Test, test, test is the secret sauce of any innovation
AI requires buy-in from the top management
Amazon also uses AI to improve customer outcomes:
- Chatbot for customer service
- Audio Video Image generation in Ads
- Optimised inventory management
- Automised robots for fulfillment in logistics
- Shopping assistants to make it easier to find the right products
- Size and fitting recommendations
What’s not going to change in the future
Instead of thinking of what’s going to change in the future, focus on what is not going to change. Your mission, your vision and your customer satisfaction needs.
AI models will keep improving, they will become cheaper and more reliable, therefore any application you build should be designed and engineered to change over time.
Italian 
English