Media Query Source: Part 54 - Reworked (US digital magazine); Can AI add value before digital transformation is complete?

• Reworked (US digital magazine)
• Can AI add value before digital transformation is complete?
• It's a good idea to define "AI" and "digital transformation" first
• An entire business doesn't need to be digitally transformed before AI work starts

The query responses I provided to Reworked on June 24, 2025:

Reworked: As the corporate world races to adopt artificial intelligence, a critical question remains largely unanswered: is there any real value in introducing AI into a workplace that hasn’t completed its digital transformation? While AI promises speed, efficiency, and intelligence at scale, many organisations still struggle with legacy systems, fragmented data, and outdated operating models. Without a stable digital foundation, can AI genuinely deliver on its promises—or does it risk automating dysfunction and amplifying existing flaws? As businesses invest in intelligent automation, the timing and context of AI deployment could determine whether it becomes a transformative force or an expensive distraction.

Does a company need to finish digital transformation before using AI?

[omitted]

Should AI come before or after digital transformation?

[omitted]

Will AI help lead digital transformation in the future—or still depend on it?

Gfesser: It might be worth noting at the outset that "digital transformation" and "AI" need to be defined first before asking questions like these.

Generally speaking, "digital transformation" essentially means that software and data are used to change business operations in a fundamental way. "AI" also makes use of software and data, but is more specific to learning from data to drive targeted aspects of a given business, either business operations like digital transformation, or products offered by a given business.

No, a given business doesn't necessarily need to complete digital transformation before making use of AI, because use of AI can target a specific area of the business. In other words, a given business doesn't need to be completely transformed before AI can be used for a specific business area.

That said, the business area being carved out, or the product being considered, needs to be ready for AI to be applied. This can mean many different things depending on the context. In the case of a large language models (LLM), for example, this can mean that the body of knowledge needed for a product has been made available in an accessible data store. The entire business doesn't need to be digitally transformed in order for a given use case to be implemented.

Digital transformation can be a prolonged undertaking for large organizations, and as such, waiting until this has been completed before pursuing AI might not make sense, but the greater the scope of the AI being pursued, the greater the scope of the business that needs to be taken into account.

As with development of traditional software and data products, it often makes sense to pursue AI in an iterative manner so that the business can understand the work involved and get some early wins to pave the path for more extensive subsequent AI efforts.

And as with traditional software and data product development, it's important that the return on investment is worthwhile. So it's important that stakeholders will value what is being built, and that the appropriate champions are involved so that this work is prioritized.

AI will continue to depend on any underlying groundwork on which it depends, be this the data on which it has been trained, or the application programming interfaces (APIs) with which it needs to integrate. But one of the key takeaways here is that an entire business doesn't need to be digitally transformed before this work can start.