Ground-Truth and Gut-Check
I wrote a series of posts to emphasize that the most underestimated work-stream of AI strategy involved governance, quality, and stewardship of data assets:
When it comes to AI, to avoid poisoning the goodwill of customers, employees, and investors, Product leaders should be introspective and find out how wicked are the strategy and culture problems related to data governance.
“Knowing is half the battle.” - G.I. Joe
New Hotness
This week, thought leaders published commentary shedding light on a long-standing ground-truth: data-related problems might be more than just a speed-bump on the road to manifesting utopia.
SAAVI AI bills itself as "Agile AI for the modern product team."
Its CEO wrote:
"After years of hoarding any and all data, AI projects are revealing that not all data is created equal. Whether it's causality data for machine learning or content data to fine-tune for GenAI, data is still necessary, but the data lake may have run dry." - Maya Mikhailov
Then, Data Guy Eric Gonzalez commented:
You mean companies have to have clean, refined, and solid back-end data development to build successful AI products??
Indeed.
Generative Buzz helps organizations adopt AI, while emphasizing "ethical use cases and strong human-guided governance."
In a podcast, the founder mentioned a series of challenges, calling out "data problems", saying:
"organizations struggle with lead and customer data quality, quantity, access, etc. and often balk at the cost of getting their data house in order." - Geoff Livingston
Good question.
Culture Checklist
AI implementation expert Isar Meitis wrote about a checklist to get started.
I suggest adding Step 0 to the list.
Only idiots escape entirely from the world that the past bequeaths.
It is a gut-check to sense-make the evolution of strategy and culture around information operations which helped lead to status quo:
have we asked the 5 W's and 1 H for data sources?
do our people have the necessary skills?
do we trust our proprietary information which will be used as training data?
have we addressed prompt injection and data poisoning?
Big Democratized Data Science
For Product teams, the "data problems" above are holdovers from pre-existing (non)investments, failed strategies, and the technological promise of Big Data and data science.
The AI hype is just the latest, most-democratized information technology product.
An honest retrospective on a leadership culture’s stance towards IT, investment in up-skilling, and managerial perception of enterprise data management will reveal wicked problems.
Pervasive and unresolved pain-points can result in wicked organizational challenges.
Culture eats strategy for breakfast.
Some leaders use data / statistics like a drunk uses lamp-posts — for support rather than for illumination.
Strategy should be about aligning on what the culture won’t permit or tolerate to happen.
Advances in tech fostered the ability of entire industries to capitalize on information arbitrage.
The tech enabled more data Volume, Velocity, and Variety. New cultural norms developed which permitted and tolerated misuse, waste, or even fraud.
Two forgotten V's: Veracity and Value
Rather than challenge the status quo, data scientists could p-hack models to align with the sunk cost, ego, and motivated reasoning of whoever is paying the invoice.
"It is difficult to get a man to understand something, when his salary depends on his not understanding it." - Upton Sinclair
So, for years, enterprise data strategies aligned, because why rock the expensive cultural boat?
AI implementation will quickly reveal a lack of enterprise data governance, quality, and stewardship.
To mix metaphors, without data governance, culture - and GIGO - will eat your AI strategy for breakfast, poke holes in the boat, and poison goodwill of customers, employees, and investors.