INSIGHTS
Why Predictive User Interfaces Beat Chatbots for Business Workflows
Thuy Lam
CEO and Head of R&D
Published 29 January, 2025
Chatbots have come a long way, but they’re not always the best fit for business processes, especially when tasks and workflows are predictable. While Chat UIs excel in open-ended conversations, they often add unnecessary complexity to tasks that could be handled more efficiently. Let’s explore why focusing on interfaces that anticipate user needs can lead to more efficient and effective business operations.
The Evolution of Chatbots: From ELIZA to Today
First, let’s start with ELIZA, the world’s first chatterbot (chatbot). ELIZA was designed for therapy, allowing it to converse with users, understand, and respond. However, ELIZA was never built to comprehend requests and execute actions.
Fast forward to today, with ChatGPT version XYZ. Yes, the ability to converse, understand, and respond in context has revolutionised the way society understands and appreciates AI. It has also somewhat masked the issues we faced 3–4 years ago when users grew despondent with ineffective banking chatbots.
However, is a Chat UI truly the best interface for human augmentation in processes? Specifically, as the interface to kick-start or augment agentic flows?
When Chat UIs Work and When They Don’t
Chat interfaces are generally great for question-and-answer interactions, where:
The customer doesn’t know exactly what they want, so they need to ask a series of questions to get to the actual point.
Your organisation aims to use a single channel to understand the customer’s requirement and then push the action down specific pathways.
Take Google as an example. A user wants to book a flight. They go to Google, type, “I want to book a flight from X to Z,” and hit enter. Google understands the user’s intent to book a flight and provides a series of links for them to proceed.
This is a good starting point for discussion. The Google scenario assumes you know very little about the user until they submit a request. However, as an organisation focused on streamlining internal and external business processes, I’m guessing you already know:
- The tasks an internal role is responsible for.
- The processes that need to be executed for those tasks.
- The services your customer requires.
So why would you kick-start or augment a process with a chat interface?
For example, payroll delivers specific services. Why not present a simple interface where users click a button, provide the necessary information, and submit their request? Why waste time typing, “I would like to query my pay amount”?
The Role of Chat UIs in Knowledge Retrieval
There’s another argument that chat interfaces make internal knowledge libraries more searchable, enabling employees to work more efficiently. Let’s explore this. Consider the role of a procurement analyst (PA). Their job is to analyse spending and contracts across an organisation to recommend cost-saving measures.
One could argue that a chat interface enables the PA to ask questions like, “What is the total spend on stationery, broken down by vendor?”
My counterargument is this: if we know such questions are common, why not present the relevant contracts needing renegotiation or examination directly? Add in a market analysis of expected costs, and you’ve streamlined the process without needing a chat interface.
The Case for Tailored Interfaces
In my opinion, if your organisation understands its processes, tasks, and services – and can anticipate and plan for common user needs – it becomes possible to design interfaces that simplify and accelerate task completion. By focusing on predicting user requirements and streamlining engagement, you can provide a more efficient and satisfying experience.
Presenting a user with a chat interface, in my view, is essentially saying:
- “We don’t know what you want.”
- “We haven’t taken the time to find out.”
- “Why don’t you start by telling us?”
While chat UIs have their place, they are not the ideal starting point for well-defined, predictable workflows. A UI/UX/CX specialist rarely begins a design with a chat window because the best designs are built around clarity, efficiency, and user
needs – not an open-ended conversation. By prioritising tailored, task-focused interfaces, your organisation can deliver faster, smarter, and more user-friendly solutions.
Agentic AI’s evolution will likely happen in two phases. Phase 1 focuses on overcoming foundational roadblocks, like legacy system integration and data quality, while implementing structures for AI agents to make semi-autonomous decisions. In this phase, organisations build insights and decision audit trails, preparing for the leap to Phase 2, where AI agents can make fully autonomous decisions and execute tasks end-to-end. Let’s explore what these phases look like and how they’ll shape the future of AI-driven decision-making and task execution.
The Evolution of Chatbots: From ELIZA to Today
First, let’s start with ELIZA, the world’s first chatterbot (chatbot). ELIZA was designed for therapy, allowing it to converse with users, understand, and respond. However, ELIZA was never built to comprehend requests and execute actions.
Fast forward to today, with ChatGPT version XYZ. Yes, the ability to converse, understand, and respond in context has revolutionized the way society understands and appreciates AI. It has also somewhat masked the issues we faced 3–4 years ago when users grew despondent with ineffective banking chatbots.
However, is a Chat UI truly the best interface for human augmentation in processes? Specifically, as the interface to kick-start or augment agentic flows?
When Chat UIs Work and When They Don’t
Chat interfaces are generally great for question-and-answer interactions, where:
The customer doesn’t know exactly what they want, so they need to ask a series of questions to get to the actual point.
Your organisation aims to use a single channel to understand the customer’s requirement and then push the action down specific pathways.
Take Google as an example. A user wants to book a flight. They go to Google, type, “I want to book a flight from X to Z,” and hit enter. Google understands the user’s intent to book a flight and provides a series of links for them to proceed.
This is a good starting point for discussion. The Google scenario assumes you know very little about the user until they submit a request. However, as an organisation focused on streamlining internal and external business processes, I’m guessing you already know:
- The tasks an internal role is responsible for.
- The processes that need to be executed for those tasks.
- The services your customer requires.
So why would you kick-start or augment a process with a chat interface? For example, payroll delivers specific services. Why not present a simple interface where users click a button, provide the necessary information, and submit their request? Why waste time typing, “I would like to query my pay amount”?
The Role of Chat UIs in Knowledge Retrieval
There’s another argument that chat interfaces make internal knowledge libraries more searchable, enabling employees to work more efficiently. Let’s explore this. Consider the role of a procurement analyst (PA). Their job is to analyse spending and contracts across an organisation to recommend cost-saving measures.
One could argue that a chat interface enables the PA to ask questions like, “What is the total spend on stationery, broken down by vendor?”
The Case for Tailored Interfaces
In my opinion, if your organisation understands its processes, tasks, and services – and can anticipate and plan for common user needs – it becomes possible to design interfaces that simplify and accelerate task completion. By focusing on predicting user requirements and streamlining engagement, you can provide a more efficient and satisfying experience.
Presenting a user with a chat interface, in my view, is essentially saying:
- “We don’t know what you want.”
- “We haven’t taken the time to find out.”
- “Why don’t you start by telling us?”
While chat UIs have their place, they are not the ideal starting point for well-defined, predictable workflows. A UI/UX/CX specialist rarely begins a design with a chat window because the best designs are built around clarity, efficiency, and user needs – not an open-ended conversation. By prioritising tailored, task-focused interfaces, your organization can deliver faster, smarter, and more user-friendly solutions.
Ready to move beyond chat?
If you’re looking for smarter, more efficient ways to streamline business processes, explore Action Fabric, the AI-powered platform designed for seamless, end-to-end task execution. Ditch the back-and-forth of chat UIs and embrace automation that truly delivers.
Visit www.actionfabric.ai to see how intelligent interfaces can transform your workflows.
