1. What are agentic experiences, and what sets it apart from traditional chatbot experiences?
Agentic Experiences is a relatively new design field where people create AI Agents to act as an autonomous “teammates”, this could be costumer service etc, to help achieve a goal directly or in-directly. This is different from a traditional chatbot as Agentic Experiences are more hands on: they can browse the internet or data given directly to make its own educated choices. These educated choices is a huge difference as it adapts and overcomes and hurtle or obstacle that is in the way of the goal as long as that issue is goal oriented. Another key difference is the over all flow of conversation: traditional chatbots such as a chatbot that Walgreens uses has a set script: its a kin to: 1.) Pick up medication or Check prescription, so on and so forth. It presents its only options and anything off the script will be met with an error message or a Sorry message as it repeats back the script it was given. An Agentic Experience can adapt to the users needs/information to help provide a smoother experience overall.
2. Describe the agent did you designed using Voiceflow.
3. Explain the underlying design of your agent by using the table below. Add in as many rows as you need to explain your agentic design framework. You can rewrite these into paragraphs if needed:
| What tasks/steps did the user need to do prior to this agent experience? | What value does the agent provide? What tasks can it complete? | How does the user interact with the agent? | What information is required to enable this functionality? |
| Research computer parts | Gives a custom computer parts with what the user needs | Describes what performance and use of the computer.laptop | Purpose for the laptop/tower, performance, tower/laptop. |
| Research Pre made computers or laptops | Can give custom and premade computers as suggestions | Answers questions asked by the AI agent about specs and models. | Initial questions answered early to build more clarifying questions to funnel down the options |
| Make a choice, compile parts for a computer | Finally narrows down options to a computer/laptop the user would be most suited for. Also knows what parts are compatible for eachother etc.. | Asks to change some parts of the narrowed down computer/tower. If a laptop then asks for a bigger size or change specs/price | Has to know what parts the user wants such as GPU,CPU,Size,Etc, and fills in the blanks with parts that fit the budget and are compatible. |
4. Record a narrated demonstration video of your agentic UX.
video link: https://youtu.be/nVSq0PnU0Sw
5. What are the current limitations of your agentic experience? What opportunities are there to improve them?
Some of the limits that I thought of are credits? a lot of AI models have credits, I have never encountered reaching max credits while using VoiceFlow, but for more niche AI Assistants I can imagine how it would be frustration not being able to edit the model. Something that the user may encounter is the limitation of memory: usually VoiceFlow and ai agents are capped at the one conversation: meaning if you do not remember what the AI agent said, you have to retrace your steps to get to the same spot you left off last time you spoke to the agent. Another limitation is updates or change in goals: the AI assistant will not update unless given the information to update through the “back end” side if development. You cannot directly ask the AI Agentic robot to update in the chat which can lead to a messy work flow of going back and forth and tweaking information.
Leave a Reply