### The Dual Nature of AI: Hype and Complexity in Enterprise

Maryam Ashoori emphasizes that leveraging proprietary data is key for enterprises to achieve competitive advantage in AI.

### Navigating the AI Landscape: Balancing Hype and Practicality
* 🤖 Maryam Ashoori highlights a paradox in AI, noting that it is both "**overhyped and underhyped**," with potential glorified while complexities are often overlooked.
* 📊 She emphasizes that the effectiveness of AI is heavily reliant on the **quality of the data** it utilizes, particularly when exploring generative AI for **impactful outcomes**.
* 🔍 As companies transition from experimentation to production, they face challenges in leveraging **proprietary data to create differentiated solutions** that meet **specific use cases efficiently**.
* 🏆 Many organizations mistakenly believe that deploying a model marks the end of their efforts, but **ongoing observability is crucial** to prevent model drift and ensure **consistent performance**.
* 📈 Ashoori stresses the importance of **customizing AI models with unique enterprise data** to gain a **competitive edge in the market**.
* 🔄 She advocates for a mindset of "**progress, not perfection**," where rapid iteration allows businesses to **test and learn**, adapting quickly to **technological advancements**.
* ⚡ Agility in processes is deemed essential, as it enables companies to **capitalize on evolving technology** and **maximize their return on AI investments**.

Maryam Ashoori: We are observing an interesting contradiction unfolding about AI, it's both
overhyped and underhyped. We are glorifying the
potential of AI for enterprise, but at the same time, we are
overlooking the complexity of

what it takes to deliver
on that potential. (upbeat music) Ashoori: AI is as good
as the data that it uses. When you're exploring generative AI, you're looking for a wow factor, you're
looking for an aha moment.

That's where very large
general purpose models shine. But as you move toward production and scale, you soon realize
the path to success is not that straightforward. And one of the challenges that we keep hearing from
our customers dealing with when it comes to differentiation,

competitive advantage, is how
to take advantage of the data that they have, the proprietary data, the data about their users
in that domain, the specific data to create something differentiated that not only differentiates
them from the rest of the market, but also
delivers the performance

that they need for the
specific target use case that they have for a fraction of the cost. While some of the firms, they feel like when the
model is deployed, that's it. Their job is done, but it's not.

Observability is extremely
important to make sure that the model doesn't drift. We are constantly checking
for the biases in the content that is generated in the output, but the enterprise data
is the proprietary data, special data about the customers,

about the domain-specific
data that has not yet found its way to
customize those models. And that's the place for
enterprises to take advantage of, to build something differentiated that is not only customized and personalized to their own needs, but also gives them a competitive
advantage to stand out

and win the game in the market. You know, progress, not perfection. The beauty of iteration
is that it gives you agility to test out something quickly. If it works, it works. If it doesn't work,
there are lessons for us

to learn and move on. And this agility is extremely
critical in this market because the technology is changing rapidly and whoever has the
agility in their processes to take advantage of
technology advancements that are rapidly evolving are the ones that can truly maximize
the ROI on the investment

Data: The Potential Diﬀerence Maker in AI Success | Business Insider

What you want to know about business. 

Visit our homepage for the top stories of the day: https://www.businessinsider.com 
Business Insider on Facebook: https://www.facebook.com/businessinsider
Business Insider on Instagram: https://www.instagram.com/insiderbusiness
Business Insider on Twitter: https://www.twitter.com/businessinsider
Business Insider on Snapchat: https://www.snapchat.com/discover/Business_Insider/5319643143
Business Insider on TikTok: https://www.tiktok.com/@businessinsider

If you have a permissions or licensing request, please connect with our partners at PARS International who will be able to provide you with an accurate quote for the overall usage rights that you’re interested in. A representative from PARS will contact you shortly after filling out this request form: https://www.parsintl.com/publications/insider/


Business Insider

Contradictions and Challenges in AI Adoption

AI is both overhyped and underhyped; its potential is glorified, but the complexity of implementation is overlooked.

AI's effectiveness depends heavily on the quality of data used; generative AI seeks 'wow' factors and 'aha' moments.

Large models excel, but production and scaling present significant challenges.

A key challenge for businesses is leveraging proprietary data for differentiation and competitive advantage.

Importance of Data and Model Observability

Deployment is not the end; ongoing model monitoring is crucial to prevent drift.

Bias detection in generated content is important, but utilizing proprietary data for customization is key.

Enterprises should leverage proprietary data to create customized, differentiated solutions for competitive advantage.

Prioritize progress over perfection; iterative approaches allow for quick testing and learning.

Agility is crucial due to rapid technological advancements; adaptable processes maximize ROI.