Forget chatbots: The real key to enterprise transformation is an AI platform

Forget chatbots: The real key to enterprise transformation is an AI platform

AI chatbots can accelerate enterprise modernization, but their limitations become apparent when businesses attempt to scale complex transformation projects. Eamonn O’Neill, Co-founder and CTO of Lemongrass, explains why complete AI platforms offer the consistency, control and knowledge management capabilities enterprises need.

If, like me, your job focuses on working with complex enterprise software systems, there’s a lot to love about AI chatbots. They help to do things (like generate custom code or troubleshoot failed software tests) that would require a lot more time and tedium on the part of engineers working without the assistance of AI.

That said, if you work with chatbots to help accelerate enterprise transformation long enough, you’ll start to hit walls. The reality is that chatbots alone aren’t enough to optimize transformation projects. What businesses really need is a complete AI platform.

The advantages and limitations of chatbots for enterprise transformation

To be sure, chatbots have much to offer as assistants in the transformation process, and it’s no surprise that many of the clients I work with have now connected AI chatbots to platforms like their SAP environments to help enable transformation and modernization efforts.

With chatbots, engineers can ask questions about enterprise environments and receive instant responses. They can also prompt the bots to initiate changes in systems automatically.

But when it comes to tasks that require sophisticated knowledge management and tight control over deliverables, chatbots often fall short due to limitations like the following:

  • Limited context awareness: Even when chatbots have access to a business’s internal data sources through techniques like RAG, they lack the ability to understand how that information is structured, or how best to leverage it when responding to prompts.
  • Non-deterministic behavior: By default, chatbots can behave in somewhat unpredictable ways. The same request won’t always yield the same output – and in the context of transformation projects where there is little room for error, this unpredictability can be unacceptable.
  • Limited memory: Most chatbots store prompt or conversation histories locally. This makes it challenging for them to learn from historical prompts or other users’ requests when generating new outputs.

Put simply, chatbots can help with relatively basic transformation tasks that don’t involve a lot of variables or require access to specific data sources. But they start falling short when tasked with highly sophisticated requests, or ones where extreme accuracy is critical.

Upleveling transformation with AI platforms

Leveraging AI to maximum effect in the context of enterprise transformation requires businesses to upgrade from basic chatbots to complete AI platforms.

By AI platform, I’m referring to generative and agentic AI systems that are distinguished by the following key features:

  • The ability to leverage different types of knowledge fabric, which don’t simply expose custom business data but also define relationships within it. Knowledge graphs help AI tools interpret ‘tribal knowledge’ with business data more effectively and reliably.
  • Centralized storage of conversation histories and prompts, and the holistic leveraging of that data when responding to new prompts.
  • Support for creating standardized, deterministic outputs when necessary. This becomes possible when businesses implement tighter control around AI systems’ knowledge fabrics. Methods like prompt markup (which feeds concrete instructions to AI tools in ways that natural language can’t) can also help.
  • Comprehensive management of the tools available for AI models and agents to use, including rigid control over which tools they should employ for which tasks. This is another area where prompt markup is valuable as a way of defining specific tools for different roles and workflows.

With an AI platform in place, the role of AI in transformation projects starts to extend far beyond simply answering questions or generating boilerplate code.

Instead, engineering teams can begin doing things like writing templates that allow them to execute thousands of prompts (and, by extension, initiate thousands of transformation tasks) in a structured, well-managed way. They can also retain tight control over deliverables (meaning the actual outputs or outcomes of AI workflows).

This is what it truly takes to leverage AI at scale in the context of enterprise transformation. Simple chatbots are great for completing one-off tasks, but they lack the repeatability and consistency necessary to automate large-scale projects.

Hence why the implementation of AI platforms will, I believe, become central to the way smart businesses approach transformation going forward.

Browse our latest issue

Intelligent CIO North America

View Magazine Archive