<img decoding="async" alt="The Ultimate Playbook for ai” width=”100%” src=”https://technicalterrence.com/wp-content/uploads/2024/05/The-Ultimate-Playbook-for-AI.png”/><img decoding="async" src="https://technicalterrence.com/wp-content/uploads/2024/05/The-Ultimate-Playbook-for-AI.png" alt="The Ultimate Playbook for ai” width=”100%”/>Image by author
What would happen if ai did not exist? In some ways, no such technological craze has swept the entire industry.
For a business leader whose only primary goal is to drive business growth by leveraging technology, the first thought is the customer: who do we serve? Who is our audience? What do they want/expect from us?
And the second immediate thought is: its weak points. What do they need that no one, not even competitors, can serve them?
Customer-centric business strategy
And there begins the series of questions that, once addressed, will guide the business towards success.
- What makes customers' lives easier?
- What makes an experience perfect for them?
- What are your unmet needs?
And so begins the path to discovering the means to an end, also known as technology.
In particular, we haven't talked about ai yet. Listing the business strategy, levers and prerogatives is the most crucial and important step in deciding “what to solve” and “who to solve for”.
After that. The question arises “how to solve it”. Is ai a good solution to solve this business problem?
Right now, companies need a framework to decide which use cases ai is suitable for. This is what I suggest: the “PRS” framework. What represents “Patterns that Repeat in Scale”.
<img decoding="async" alt="The Ultimate Playbook for ai” width=”100%” src=”https://technicalterrence.com/wp-content/uploads/2024/05/1715544582_161_The-Ultimate-Playbook-for-AI.png”/><img decoding="async" src="https://technicalterrence.com/wp-content/uploads/2024/05/1715544582_161_The-Ultimate-Playbook-for-AI.png" alt="The Ultimate Playbook for ai” width=”100%”/>Image by author
Pattern
Let's take an example to internalize this framework:
For example, taxi service providers guarantee the availability of taxi drivers at a profitable price, which considers several factors:
- Proximity of the group of available drivers to the taxi requester
- Distance to destination
- Peak demand causes price increase because there are more taxi seekers compared to taxi drivers
- Reportedly, lowest battery status of the taxi applicant's phone probably suggests a higher fare price. This gives the taxi service provider the signal that a low cell phone battery may increase the taxi requester's desire to pay more for the same trip due to a sense of urgency.
- Taxis availability and prices also vary depending on factors such as regular versus premium taxi service, time of day, or unfavorable weather conditions.
<img decoding="async" alt="The Ultimate Playbook for ai” width=”100%” src=”https://technicalterrence.com/wp-content/uploads/2024/05/1715544582_193_The-Ultimate-Playbook-for-AI.png”/><img decoding="async" src="https://technicalterrence.com/wp-content/uploads/2024/05/1715544582_193_The-Ultimate-Playbook-for-AI.png" alt="The Ultimate Playbook for ai” width=”100%”/>
Image by author
All this, while ensuring that taxi drivers are sufficiently incentivized to continue enriching the customer experience.
So, we understand data patterns.
Repetition at scale
Next comes repeatability: all of these data attributes are repeated for every taxi requester and every trip across geographies, which inevitably leads to our last point: scale.
Think about how unattainable this problem would be if there had been a manual or non-ai workflow to solve this compute-intensive business case.
Data strategy
Once we have built the business mindset, after which we have identified the problems that are good arguments to solve through ai, let's put all our attention on the data. After all, data is the core engine that drives the success of all ai algorithms.
I also have a framework for this: AAA, which stands for Availability, Accessibility and Authorization.
Consider this:
Do I have the data?
vs.
Do I have exhaustive data?
There is a minor but crucial difference between these two statements.
It is not enough to have data. All the data necessary to model the phenomenon is needed to ensure that the model sees all those attributes that a human expert can also see. Therefore, data availability is key.
Next is data accessibility. Having data available is one thing, but being able to access it easily is another. It is important to create data pipelines to ensure seamless access to data.
So far, we have covered a lot of ground in getting the data into shape, but what if we are not allowed to use the data for model training or analytical purposes?
This is where most organizations make a mistake. Make sure you obtain the necessary authorizations or, better yet, only use the data for which you have the necessary permissions.
With the 3As of data strategy, there is still one unanswered question: what is the sequence or order between business strategy, data and ai?
So many strategies!!!
To a large extent, ai strategy is always a function of business strategy and is aligned with data strategy. It is prudent to continue working on ai use cases in addition to keeping the 3As of data in progress.
Similar to the iterative nature of ai projects, the ai roadmap needs continuous refinement as data infrastructure is prepared and improved to maximize the potential of ai technologies within an organization.
Continue to analyze and track key performance indicators (KPIs) such as accuracy, efficiency, and ROI to periodically evaluate the status of ai initiatives to measure their effectiveness and identify areas for improvement.
Additional tip
Most ai projects and strategies are affected by a lack of timely communication. It is essential to conduct milestone checks and actively solicit feedback from stakeholders, including end users and business leaders. All successful ai projects go through several cycles of iterations through feedback that informs adjustments and improvements to existing models or the development of new use cases.
Furthermore, models are not developed just once and never re-examined. It is quite possible that business priorities have changed over time, which should be reflected in the ai strategy as well as its implementation.
Vidhi Chugh is an ai strategist and digital transformation leader working at the intersection of product, science, and engineering to build scalable machine learning systems. She is an award-winning innovation leader, author and international speaker. Her mission is to democratize machine learning and break down the jargon so everyone can be a part of this transformation.
