Hi @vinit, thanks for the welcome. I think your example is a great landing page, but I’m wondering what the optimum length for a chatbot should be. I tend to be impatient and don’t generally stay involved if I’ve got to do too much interaction. What’s your take on this? With your access to all the data, how big is the average chatbot?
2 Likes
Hi @scuba722,
By length of the chatbot you mean the number of gambits right?
To decide that, you have to think about the purpose of the bot.
For landing pages, it is usually about communicating about the product and business, and most important what value proposition you have for the landing user.
In such a case you should focus on your core messaging first and then you can ask the user if they need further clarifications. Depending on their response, you can either take them on a branch which continues the conversation and explain more OR it can close the conversation and give them a call to action. This is a really good way to dynamically adjust the conversation length, which is possible only in chatbot.
Check out our Product Explainer bot here:
https://tars.hellotars.com/conv/XLEzwk
Here in between the conversation the bot is asking the user if they understood the product, or do they need more clarification.
1 Like
Hi, everyone! My name is Antonio, I’m from Portugal and I mainly work with WordPress.
I’m having fun with bots (who isn’t, right? 
) and I hope I can learn and contribute with something in this community.
4 Likes
Hi @ajcalmeida,
Glad to see you here.
Feel free to share some of your fun bots here for us to try and get inspiration
Greetings Fellow Tar fans and Bot buddies.
My name is Dave and I own a marketing agency for many businesses and look for technologies that scale. A well crafted solution for internet marketing exists in combining semi-to-fully automated solutions that only requires limited assistance from working humans. A mission of mine is to build an assistant that’s so good for your business in digital communication and that nobody knows it’s actually a master crafted bot.
I also own WISP, an internet service provider that will scale nationwide in presence and I look forward to implementing a bot to take orders online and to evolve to handling much more in support, CX, localization and evolving dialogs for all parties that handle.
(FYI- I’ll be posting another thought and open question about the benefits of bots and overall benefit of this platform. #stayontopic).
Thanks for making this forum and the Tars bot-tool!
1 Like
I think where you may get into sketchy territory is trying to conceal that the software is automated. I believe it can be done, but the finesse level would have to be off the charts. Also, you risk your users finding out it is automated and then lose a significant amount of trust if they feel they were intentionally duped. To pull it off, you will need some sort of API that analyzes user text (offering only user text submission as the input means, so they believe they are chatting with a real person). The API would have to make a determination as to the tone and intent of the person’s writing and call up a logical response in turn. This would be an amazing project to work on. I’m sure those working in AI right now are attempting to build some kind of text recognition software that operates in a similar manner. Anything is possible!
4 Likes
Levi,
Good point on being transparent. I’d be more inclined to see Ben as programmed extension of myself (my business), not instead of, or a replacement for, a human. It’s obviously a bit out of scope for this Intro thread, so the blue sky discussion and specific feature requests I’ll provide elsewhere if they are desired.
2 Likes
Hi All,
Thank you @vinit and @ish for creating this great product, I am so excited to see the amazing bots that people have created using Tars and the helpful community.
I am a co-founder of a not for profit (public charity) which is a professional organization of Information Security Professionals https://isc2chapternj.org
We along with a few other InfoSec organizations collaborate and host a conference in New York City.
Every year sending out Certificates of Attendance and collecting feedback is a chore.
This year I have created a bot to see if both those tasks can be automated.
I was amazed how little time it took me to create my first bot 
I am revising it as little everyday (the conference is on 10/5/2017).
Could you please test my bot and provide feedback? https://convbot.hellotars.com/conv/SkYKnt
Thanks
Niloufer
1 Like
Hey @sponge2,
Great to have you here, and i feel really happy to see that you liked the product and are able to use it.
I checked your bot and i love it. Simple and Straightforward.
May be after the events you can add some pictures from the event in the Bot itself.
That’s a great suggestion @vinit I will add one picture each as I ask for feedback on the speakers, content and food/facility.
1 Like
Great suggestion @Christer by adding those words we can get approval, engagement from the user.
Anything else to look out for?
If I were you, I’d ask the user for their name up front either before they select Feedback or Attendance button, or just after. Then you can refer to the user by name through the rest of the interaction. I’ve found that on the simple and straightforward bots, this addition will bring the UI to another level with minimal effort on your part. Simply add this code {{ursp.xxxx}} to the bot speech bubble (Where “xxxx” is the name of the gambit the user submits their name in) whenever you would like the bot to refer to the user. So your bot could say something as simple as “Hi, John! So what did you think about the speakers at the conference?” Also you can branch your ratings, so if a user rates three stars or less, your bot could prompt them with a text submission area that asks if they have any suggestions or what their specific issues were, then send them on to the next question.
4 Likes
Thanks @Levi for your valuable feedback, will definitely move asking for name and email early on and ask for comments based on ratings.
Appreciate your input.
2 Likes
Hey guys,
My name is Kevin and I’m joining the TARS community from Geneva, Switzerland. I’ve been following TARS and reading most of the posts for nearly 2 months now so I figured it was about time that I introduce myself. My goal joining TARS is to be able to create bots that can help small businesses by enabling them to have a quick access to HR and legal related information so my bots act as HR managers for small companies.
2 Likes
Hey @Kevin
Great to have you here. Bots in the Legal and HR space is a really great idea, specailly because of the amount of information that you want to be able to communicate to the end user. Would love to check any Bots that you have made.
BTW, check this Bot in the legal space which was built by one of the Bot Makers on the platform
https://lawsuitfc.hellotars.com/conv/ByA7Zg/
2 Likes
Nice work ! I will definitely check it out more in depths ! Here’s the bot I shared also in the other discussion thread: https://convbot.hellotars.com/conv/SyBS1P/?_speak=en
2 Likes
Hey @Kevin,
Thanks for sharing this. This is a really interesting and information Bot. Specially for this space, which has so much information to dispense.
Great work 
1 Like
Hi,
I’m Navin Narayanan, partner at Centre of Gravity, a consulting firm based in Bangalore, India. We work on gaining deep understanding of human beings, their life and preferences. And using this human knowledge to help CEOs and leaders make decisions.
We use chatbots as a way of humanizing large scale market research and modifying them into human conversations. This yin-yang of face to face deep conversations and chatbots give us a large scale to human insight.
And as far as our experience in using TARS, we believe we have only begun the journey. Looking forward to listening and learning.
2 Likes
This is an interesting approach albeit an intuitive step with chatbots. I think TARS is right on track with your intent with the massive amount of user interactions you can capture and parse out. I would keep tabs on conditional jump logic and auto-suggestion lists/text input as this allows the user to put things into their own words while still offering the bot an ability to branch to different conversation flows based on keywords. I’ve developed a couple simple FAQs that do this.
The beauty is you capture the way people want to interact with the bots and can tweak the interactions to meet those needs. They are much more likely to type exactly want they want in these interactions as well (they feel like they are talking with a bot after all).
I would imagine with a lot of focused time, one could develop a deep conversation bot if you took the time to sift through the info, thought of intuitive ways to capture the tone and intent of the conversation, and move the conversation that way. I would look to an API to do most of the heavy lifting. Anyway, welcome!
1 Like