3.7x ROI on GenAI

Tell me this: Have you started using GenAI in production yet?

If not, I understand. Until last year, only 5% of enterprises used GenAI applications in production environments. However, Gartner predicts that in just another year or so, over 80% of you will do so!

This kind of a meteoric rise in adoption is unforeseen! With good reason.

Small starts: 3.7x ROI on GenAI

In a study commissioned by Microsoft, IDC estimates that “for every $1 a company invests in generative AI, the ROI is $3.7.”

A deeper look into the report shows that a vast majority of this ROI is coming from productivity-related use cases, such as summarizing chats, generating first drafts of documents, and automating other repetitive tasks.

If you’re just beginning your GenAI journey, this is the place to start. Talk to us about use cases that you can leverage.

If you’d like to read the IDC report yourself, here you go.

Moonshots: 10x ROI

The same IDC study shows that leading enterprises make as much as 10x ROI.

This demands more than just productivity automations. It needs a deep and meaningful look at creating business value with GenAI.

Moonshot use cases tend to be personalization, customer experience, developer experience, drug discovery, data management, etc.

Have you read our case study on how Tune AI used data indexing to reduce document processing time from several hours to just 7 minutes? That’s more than 10x ROI!

[Read case study]

From GenAI to AGI

Another big buzzword in our world is AGI, i.e., artificial general intelligence.

While the experts can’t agree on a precise definition of the term, it is safe to say that AGI will have all the capabilities of humans, including reasoning, imagination, and more. This is the stuff sci-fi films are made of.

Naturally, it has its risks. Here’s what Anthropic found to be the possibilities of sabotage. If you’re thinking AGI is years away, think again.

Open AI CEO, Sam Altman, believes that AGI could be achieved in 2025 with current hardware too. Anthropic CEO, Dario Amodei, expects it might take a year or so longer. Well, they have vested interest in making these huge claims, but they’re making progress in that direction.

Whenever it comes, AGI can be an exciting technological advancement. If I can lay my hands on AGI next year, I would be excited to experiment.

What would you like to experiment on?

Until then, the reality

Before I say bye for this week, I found a couple of interesting GenAI tools for you and your engineering teams.

Data transformation toolkit: If you’re thinking about using your unstructured data to fine-tune models, you might be interested in IBM’s new open-source toolkit, Docling.

ML crash course: Google’s AI team has reimagined their free machine learning crash course. If you have a team of AI beginners, this is a great foundational lesson. Check it out!

We’ve been working on something very exciting with a client in the Middle East. That story coming up next week.

Until then, stay tuned.

Best,
Anshuman Pandey

P.S. Llama 3.2 Vision Language model is now available on Tune Chat