How to BUILD AI Callers with NO CODE | Bland AI Conversational Pathways

in this video I'm going to show you how you can create an AI collar with no code at all using the Bland AI platform and Bland AI has a feature called conversational Pathways which allows you to essentially create AI colors with no code just using a simple and easy to ous drag and drop interface so I'm going to go through that uh feature of the pathways uh entirely and then also go through a live example to show you how you can create your own AI caller using conversational Pathways in the matter of minutes so once you create your Bland account you'll end up in the homepage and you can get started with conversational Pathways right over here and once you land in the conversational Pathways uh page you can see that you're able to either duplicate a template that land has already created for us we can also generate a pathway which is a new limited access feature right now and makes it a lot more easier to you know create Ai colar and I'll be going over that in another video but uh for this video we'll be creating our Pathways from scratch and going over um them in in that fashion so you can go ahead and hit create pathway you can give it a name so AI caller one something like that create pathway and then then you should see yourself end up in this drag and drop interface so with conversational Pathways you have two main components to them and they are the nodes which you see over here this is an example node this one over here and this one over here you have different types of nodes by the way and then also you have connections kind of licking uh the nodes together um depending on what you're trying to do right so with Pathways you you can uh use nodes to represent different points in the call and for example you have like this greeting node right over here and then if you're trying to create let's say like an inbound callar for a restaurant which is the example that we'll be going over uh you can have a node to to greet the customer and then depending on what they say um you can give this uh link or connection a description so you can say if they use user wants to book a reservation after greeting then you can uh create a different path for uh booking a reservation if they have a question about the menu or about something about the restaurant you can have an entirely different path for uh that sort of scenario so you can kind of create different scenarios for your call and um you know have those be represented by different nodes or collections of nodes and essentially what you're doing is is creating a call tree or a call structure if you're familiar with those so I'm starting with a single node right over here and if you double click on it you should get the setup or configuration for the node and the first thing that you'll see is the node type and with node type you can specify what type of node you want this to be uh we'll go over each and every one of these different note types but obviously you can see you have knowledge base you have web hooks you have you know transferring SMS pressing a button um but the default uh node is probably the the regular or you know the standard node that you'll be using a lot whenever you just want to generate a response like uh for here if we're just greeting the customer we just want a default no we don't want to do any sort of special ability or like call a knowledge base or a web hook right we just want to greet the customer and so we can give this uh note a name so something like greeting and then also you can specify whether you want it to be a static text or you want to give a prompt uh so what this means is that let's say you want the AI caller to have a fixed response uh every time it comes to this node so if this is a greeting node you want it to have the same exact greeting every time so say hi hello you know my name is John from John's Barbecue right you want it to say that exact text every single time then you can say uh turn static text on and then also put the text right over here but if you want it to to kind of dynamically uh generate a response you don't want it to be fixed every single time you can just turn off Static text and then pretty much prompt it um to to create the generator response um which should be pretty easy to do you usually want to create these uh prompts to be concise and you know as brief as you can be you don't want to make them too long just to reduce the latency and response timing um when it comes to you know doing these conversations and then also um we have a lot of different other settings that we can mess with so we also have condition options so condition options are essentially uh a way to make sure that the AI caller doesn't leave a node until you know X condition is received so say you need some piece of information like you're trying to book a reservation you need to make sure that you have the number of guests correct um so before moving on to the book uh reservation node where you actually book The the event in the calendar or the book system you want to make sure that you don't leave that note until you get that piece of information so you can do that right over here and say um follow this typ type of example and make sure that you know hey you don't leave this note until you get the number of guests that um the customer wants to book for right so that's how you use condition options you can turn that off if you have no conditions at all and then you also have Global nodes and Global nodes are just a way to create a node that can be accessed at any time um so I can give an example so let's say you we have the start node then we have a book appointment node we have another node for answering questions about the menu and let's say for example the customer takes the book reservation node and while they're trying to book a reservation um they're trying to decide what time to come at they're they suddenly ask hey uh what time are you guys open What's the working hours right and that might not be built into our path right over here for the booking the reservation because we're solely F focused on booking the reservation so we might want to have like a GL Global node that can be accessed at any time so that whenever a customer asks about the the restaurant any information basic information about the restaurant like the working hours the address things like that that can be you know and you can pretty much go to that note at any time and then quickly return back to where you were um so that's what Global nodes are um you know created to do is to to be accessed at any time um it's really good for like knowledge bases and um any basic information you want to have available at any point in the call you you probably want to create a global node maybe like a knowledge based uh type of node so it can be Global and knowledge based at the same time um but we'll be going into knowledge bases and uh that type of stuff a little bit later on and then you also have uh if you continue with the global node you can give it a a pathway so like if we were talking about that knowledge based example the user has a question about the restaurant and then you can also give him more context um so when it when it should choose this pathway so if you want to get some more information and then then also once this uh you know node is finished if it's a global node and you finish executing on This Global node like the customer's done asking questions about the restaurant then we can either go back to the previous node that it was at so if it was just booking a reservation can quickly go back to a reservation or you can re redirect to any other node that you want it to to go to um so that's Global nodes and then we also have advanced options where you can pretty much configure like the model and um this is kind of like how you generate the responses so you can pick the model type that you want uh to be generating the responses so if you want a higher intelligence model you want more speed um you want them more balanc or you want a little bit more higher speed but not as bad intelligence um this allows you to kind of you know pick between what you want to trade off because obviously you have to make some sort of tradeoff um do you want the most speed you can get do you want the the most intelligence um so it's all dependent on your use case how much you value response time versus actually maybe making the correct decisions during the call um and then also we have the temperature um and the temperature is essentially controlling the creativity of the responses if you want it to be uh more reliable more predictable maybe say some of the uh same responses more often then you want it to be a lower number and obviously this can be only between 0.0 and 1.0 but um you can usually pick something like 0.1 0.2 is good for more predictable um less random responses and if you want creative responses for whatever reason then you can also um make this a higher value up until like something like 0.8 0.9 um and then you also have some other smaller settings like the interruption threshold how often um do you want it to to kind of reduce interruptions at specific points in the in the conversation as it says and then also skipping user responses so if you want to you know generate a response at this node and then quickly go to the node without waiting for the user to say something and then also you can turn on blocking interruptions which is just um ignoring any interruptions that the user uh or the customer is um making so if you're if the AI is talking the and the the user or the customer tries to interrupt it'll just kind of ignore their Interruption if this is turned on then we also have EXT ract call info into variables which if you turn that on um if we go back to the example when we were mentioning the conditions let's say we want to get the number of guests for a reservation we also have to extract that variable as well right to make the reservation let's say we have a uh reservation tool or a web hook um in order to to make that reservation we're going going to have to send data to that web hook we'll have to send like the number of guests right um the date the time um those sorts of data points and and variables and the way that we can capture that information specifically is um through this block right over here so we can give it a name so something like the number of guest you can just say guess and then also give a a type so integer for numbers and then the description of the variable to help the AI kind of capture that variable so it knows what it's exactly looking for um and then you can add more variables uh right over here if you have multi multiple variables that you want to extract and then you can also turn this off and then also we have fine-tuning examples so fine-tuning examples is uh good for if you want to improve the performance of the AI on a specific node so if you see that it's not generating good responses you might want to add some examples to the dialogue example say give it some context or like a specific scenario so if the user says something like this then I want you to respond uh in this way right so you can give it examples of different uh scenarios and and what to expect so that way it performs better at this node you can also add pathway example so it's pretty much the same thing as dialogue except with path way so if it's having trouble picking the correct path So like um you know you see that the AI caller is not consistently going to the booking appointments path when it should be going to the booking appointments path or booking reservations um maybe you want to add some examples so if the user says something like this then make sure you go on the book uh reservations path and you'll see different Pathways kind of show up right over here and you also have condition example so if it's having example uh uh if it's having trouble kind of working with the conditions like the condition options um like it it's having trouble deciphering when the conditions um are achieved or not achieved then you you can also add examples right over here saying if the uh customer says something like this like they don't specify the number of guests or they say like a negative number of guests for some reason then you want to say hey the expected condition uh achieved is like false or something right um so you can give example for the conditions as well but that's how that works is the that's the fine-tuning examples and that's also pretty much the all the settings that you have to worry about for the default node now if we go through the other node types um you'll see that a lot of them also have the similar options or less options so we go over to knowledge base base or right over here you'll see that you have the added text box for knowledge base so you can just directly paste in your information right over here and um you know you have all the other options that we just went over and then also let's say you actually have uh a very big knowledge base right because this is um also has a limited character set um limited amount of characters that you can put in there you also have another Noe type called Vector database knowledge base which allows you to create a much bigger knowledge base for those that have a lot of information that they want to kind of incorporate into the AI Coler or the the actual node and to create the vector database you actually have to um go over to the API and if you go to the API documentation and should see create a vector store you can create a vector store by uh or this is same thing as a database Vector store Vector database same thing um you can send a post request um put in your API key and then a name for the vector database a description of the vector database and then also the actual you know contents of the database or the the document that you want to you know vectorize and use as a as a knowledge base right so that's how you would kind of use a a bigger knowledge base is through the vector databases and then if we go back you also see that we can send uh Speech during the uh Vector database knowledge based query so what the agent should say while it's kind of running the query to to see what um you know relevant knowledge is inside the actual knowledge base since this is probably going to take a longer time if you have a bigger knowledge base right so you might want to say something during that time while it's running all that stuff in the background um so that's that and then we also have a couple other nodes so we have the transfer call node which allows you to pretty much transfer the call from the AI to a different number which could be controlled by a human and then you can also say some last words right over here so you can make this something like static text and say I'm about to transfer the call and then you can put the transfer number right over here and we also have the end call uh node which is pretty self-explanatory you want to end a call you can say a couple last words right over here and then have it end the call automatically we also have web hook so if you want to do any sort of actions um like you want to interact with the API you want to um interact with your CRM calendar you can use something like make for for this to to easily integrate with a lot of different apis and platforms and um that's kind of what this is for so you also have um the web hook block where you can specify if this is like a post post request for example and then the actual web hook URL can spe specify any headers uh for the request the actual um data that you want to send in your web hook request so example like um your booking reservation again um you can send the number of guests the uh date and time of the reservation and then also if you're going to be extracting any sort of information like the web Hook is sending a response you want to extract any data from that response you can do that right over here you can specify the name of the variable which you want to capture that data into you can specify how to you know extract that data if it's like the first field in that Json response and then also um you you have an example right over here of how to use a variable if you're asked here is a weather you're essentially telling the AI what what's being returned um kind of like the um if the reservation is successful you can say something like if you are asked or something like that the reservation it was successful and or you can just say the reservation is successful done and then you also have a timeout uh which you can specify if the webbook takes too long you can just have it time out and then you can also specify the pathway that you want the request to take uh so you can have it by default just once it completes just go to this pathway so it will have all the other nodes right over here that you've created and you can specify what node to go to after this API request completes or let's say if we have have a a reservation success uh variable we can say if the reservation success uh is equal so you can select something like equal in this box and then say is equal to true then you we want to go to a certain node if it's equal to false then we also want to go to a different node right we want to handle those scenarios a little bit differently then you can also send some speech while the web Hook is executing and that is all for the web hook node and then we also have a wait for response uh sort of node which is pretty much the same thing as the default node but it allows the AI to kind of wait for the user if they need to grab some sort of information or they need to look up something um like the the AI is waiting on a specific piece of information from the the user to uh kind of respond with so you can have it wait for an extended amount of time using the wait for response node and then we also have the uh Custom Tool or the transfer pathway so we'll go over transfer pathway which is um also kind of self-explanatory so we have uh the name of the node and then also if we want to transfer to an entirely different conversational pathway we can select right over here the different pathways that we want to transfer to um so it's kind of like similar to vapy squads if you're familiar with that if you want to uh transfer to a different department or like a different sort of AI agent or caller you can do that with transfer Pathways and we also have custom tools which is very Sim uh similar to web hooks um but the thing with custom tools is they're a little bit more repeatable so you can kind of create the the tool one and then just reuse it um but very similar to you know web Hooks and you can create this on the blend dashboard they have a a sort of playground like right over here on the left hand side um and you can create the tool and specify all the configuration for that and we also have the SMS which is also right now limited to Enterprise users but essentially it just allows you to to send text messages and SMS and then lastly we have the press button node type which allows you if you're on like an ivr call so for example sometimes you call um like customer support um and they're trying to direct you to the correct agent um you might want to press one for a certain department and then another number for another department so that's kind of what this note allows you to do is allow you to to actually press those buttons allow your AI to take those actions to press the different Keys uh depending on what you want it to do um so that's what allows you to to kind of uh do is you know you can specify sequence of digits um and then you can also dynamically do it through like a agent decision um but that is all for the different node types and then a couple of smaller things also on the uh Pathways feature um so you can also add a global prompt to all of the nodes and this is just a prompt that's kind of appended to all of the the nodes at the start so this can be something like a a system prompt that you would normally do um for just like the entire call and it gives the AI context on like how it's operating like the tone that you want it to operate in um and then also maybe you want to give it a name or something like that um you also want to say um something like whenever it deals with times deal with it in a military format so instead of saying like 2: p.m. it should be something like 14 right um so things like that any something anything that's kind of general that you want it to apply all throughout the uh call you don't want to to kind of do it for every single node you can do it really easily through here using the global prompt um so it just serves as like a a general outline for the call and just improves the responses and helps you out as well when you're building out the caller and then also we have um the different options so we can easily create a new node using that button right over here we can also um duplicate a node so if you click that button right over there you can duplicate you can also copy the node so that will copy it you can delete it with this trash button or you can also um just backspace um so if we click on that backspace it should delete the node then we can also again connect the nodes you can connect like this and then also you can connect it to itself so you can go from here to the start and then it will connect to itself and you can click on the actual connection delete it and then also you want to to give these connections uh descriptions um so when should the uh like the start node go to this node um like let's say maybe this is the booking reservation node um we can say something like when the user wants to book a reservation and that kind of allows us to to tell the AI when to go to this node um so that's connections and then we also have some stuff on the side so we can zoom in zoom out we also have like a full screen option we can lock the pathway and then this is also like a undo button um so that's that we also have the redo and then also we can test the pathway so um this is also a really nice feature up lend which allows you to test your pathway instead of just having to call it every single time which can be a little bit you know uh repetitive and a little bit annoying sometimes you can just chat with it so you can tell which node to start at if you want it to give uh any sort of information of variables beforehand so you want to give it like the first name of the customer you can give the the actual name of the variable and then the value say something like John and then you can actually start the test but if you're going to use a variable like this the way that you want to actually use it within your pathway is say something like um go to default and then say hey first name you want to put it in double curly braces and then that way um when it actually starts the call it will look at the first name variable and say hey John how are you doing today um so that's kind of how that works so now we'll get started with a actual walkth through uh for a pathway and creating it from scratch and we'll be creating a pathway for a inbound callar for a restaurant um very similar to kind of like Bland template and we'll be creating from from scratch this time instead of just kind of duplicating it and we can start from the global prompt and you can see right over here I've already added a global prompt but I just wanted to give it some context hey this is your name um you're working at a restaurant um be aware that it's kind of like a phone call and um you want it to act appropriately then also some just um Miss aneous sort of information um to use uh or convert all the times to 24 hour timings and this will help when we actually book the reservation since we're going to be using Google Calendar to book uh the reservations um and that will help when we're actually picking the the start time and make and and all that sort of stuff so you can go ahead and add that now after the global prompt we can go to the Sterling node and keep the same name of the node keep it as a default node since all we're going to be doing is greeting the customer and then we also just want to have the same greeting every single time so we'll do static text and then for this uh sort of greeting I have a pretty simple one just saying the name and then how can I help you today and then for the other options we really don't uh want to mess with anything here we can specify any sort of model type uh that we want to use as well but I'll just go with the default one and the default settings can go ahead and hit save that's a pretty simple node and then add the next node which will be the booking reservation so we'll say something like keep it as a default node and then also call it um get the booking information since before we can actually make a reservation we want to make sure we have the appropriate information like the number of guests the date the time that sort of stuff then we'll also keep it as a prompt and then make sure that the AI um gets the appropriate information ask the the questions one at a time as well and we also want to turn on the condition just because we don't want to move away from this node until we get every single piece of information that we need so an example condition can be um you must get the date time number of guests for this reservation before you know actually moving on to the next node can also hit save since we don't really want to um mess with any of this information we we won't um extract the the actual variables just yet um we'll do that in the the next node so once we're sure that we've actually gotten all the information from the customer like the customer said all the information then we can extract in the next node so we'll do it there we can also connect these nodes so if you zoom in can connect those two nodes and then Al specify the um when to take this connection so we can say the customer wants to book a reservation and then save that now the next node will be the actual webbook to book the reservation once we have all the data that we need those three variables so what we'll do is actually make the connection right now and say something like um we also want to have two different scenarios so depending on the number of guests if the number of guests is kind of low we can book the reservation ourselves but for like larger parties and larger number of guests we want to direct the call to like a live Human agent since it's a more higher priority call um so we'll book the reservation for let's say um number of guests is 12 or less then save that and then double click on on that node and we can hit web hook and then say book reservation we can also extract the call info and specify the actual variables that we want to extract so we're going to be extracting three different variables so we're going to extract the date we're going to also extract the time and um the number of guests so we can say string this is going to be an actual string string and then this is going to be an integer and then we can also say something like the desired date of reservation and we can also specify the format so we can say year month and then the day format and then we can do something similar for this one as well so desire desired time of reservation and we can also specify even though we've set in the global prompt just for extra emphasis say 24hour format and then EG 13 and then also for the guest can just say something like number guest for this reservation and now comes the actual like web hook and setting it up so what we're going to do is this is going to be a post request since we're going to be sending data this will be the URL of our web hook which we'll set up and make in a little bit so we'll leave this blank for now headers um don't need to mess with that just yet the body and the response data for our example since we're trying to keep it simple um we won't really use the response data that we'll send back we'll be just sending a true every single time um so we really don't need to to mess with that didn't want to show it and then also the headers um we're not going to be setting uh sending any sort of headers in our web hook request and the body we do want to modify with our data so we'll say something like add some spaces the date variable and then actually specify it since this is a string we'll put quotes and then um insert the date variable right over there put a comma and then kind of replicate that for time as well time variable that is going to be the actual time so what this will look like is this will be the date field and then Bland will replace this part with the actual variable so if this is something like um 2024 um 07- um 05 then it will put that actual value in here um so that's what we're doing and then the last field is going to be the guest um so we'll say the guess field um is going to be the guess variable and that should be all for the body and then for the pathway we we haven't set this up yet we haven't set the next node but we would put once the API request complete since we're just going to assume for this um use case or this AI Coler that it'll always be successful um we'll make sure to just go to the next node um which we're about to create so we'll come back to this and then we'll also say while it's executing the web hook just say something like all right one moment please uh while I triy to book your information book your reservation and that is it for the web hook node we'll of course modify this again once we create the next node but just select any sort of nodes I just selected greeting and we'll change this afterwards just because you can't save until you select another node for this one so we'll hit save and then um from here we can add the kind kind of like the end call so once we book the reservation we can end the call so we can say add new node then we can also specify go back and pathway after API requests specify the new node we just created and now every time this web hook will finish it will go to like the reservation success node which we're about to create so we'll go back here and then we'll specify this to be an end call uh notes since we just want to end the call after we you know book the reservation obviously you could add some other stuff like um keep on answering the the rest of the questions but um we'll just end the call from here and then we'll say reservation success and then we can also specify a prompt so we'll just say something like this to tell the AI to make sure that um you let the customer know that you've confirmed the reservation and then um just say goodbye so we can um you know make sure everything is good um don't want to mess with the other options again don't really have to put any sort of conditions or make this a global note or extract any sort of variables so we can hit save and then we can fill out some of the other scenarios so we also mentioned right over here that for the smaller parties we want to hand hand uh those sort of reservations ourselves which we are but then for the bigger ones where we have more than 12 guests for example we want to transfer to a live Human agent so we can do that by adding a new node and um we can say transfer to agent you can just give it any name for the transfer node and we will say transfer call for the node type and then we can also say like some sort of static text so um just to make sure it says the same thing every time it's trying to transfer you can say one moment please while I transfer you to one of our agents and then we can put in a transfer number the actual number we want to transfer to um I'll just put some dubby number um so you can put that in and then for the rest of the node don't really have to modify it again you can just hit save and then make sure you also connect these nodes so you can select the bottom of the get the booking information connect it to the transfer and then also specify the connection so the number of guests is more than 12 can keep it pretty simple simple like that and that's pretty much most of our pathway we also do want to handle the scenario when we have questions about the restaurant in general so we can um create a knowledge based Global node for that uh sort of case like we specified like when when we were discussing the knowledge based node uh we mentioned how you know those you fit pretty well for Global nodes so we'll do exactly that and create a new node for the knowledge Bas so we can create a new node off to the side for this knowledge base and double click that make sure you select knowledge base we can call it something like since this is going to be for any sort of information about the restaurant we can say restaurant information questions and we'll keep it as a prompt just because we don't want to to kind of say the same thing every single time we want it to to kind of select the most relevant information from the knowledge base and then um say say that appropriate information to the customer for the prompt we can put something like answer any questions about the restaurant that they they might have by referring to the knowledge base and then keep the the responses pretty short like one line and then the actual knowledge base we'll put something like the address and then a short description of the type of food um we can also allude to the actual menu being available on the website and then some other information like there's no catering or something like that work hours um so that's that's pretty much the knowledge base that we're going to be using uh for this sort of scenario and then we can also enable it to be a global node say the pathway label can be something like customer has questions about the restaurant and or their menu and then we can also make sure that it goes back to the previous node so if it was just booking reservation just go back to the reservation node as soon as you're done answering any questions and that should be it for this node so you can go ahead and hit save now the last thing that we want to do is also make sure we finish up with the web hook node since we haven't actually set up the uh make scenario or um web hook that we're going to be using so if we go back here you'll see web hook and we left this URL blank so we want to fill this out so we can create something like this and uh make where we actually booked the reservation and if we head over to make and I already have the scenario set up it's a pretty simple one we have the web hook and then we book the reservation um but what you want to do is obviously start by creating this web hook Mont mod and you can add a new uh sort of web hook and copy this address and you can use Postman to to kind of simulate um an example uh request to determine the data structure and if you remember we have three variables in uh the data structure that we're going to be sending we're going to be sending the date the time and the number of guests so you can use something like Postman to to kind of simulate this and if you go over to postman all you need to do in post man is just create a new request set it as post and then also copy that web hook URL from M paste it in over here and then also for the body you want to select raw Json and then put some example data so we have the date we have the actual value for the date same thing for the time and the guest and then you can send that and then it should kind of go through it to to the make scenario and then from there it should determine the data structure of the web hook you can hit okay and then select the next module to be tools and set the variable so what we're going to do is just kind of concatenate the or combine the date in the time into one sort of value or variable um so we're going to take the parse date function and make and we're going to take the date and the time combine those two in in that single value and then format it or like parse it into this format um so we already have the date in this format we already have the time in this format in the military time format we're just all we're doing is combining combining them into a single value separated by the space and then I'm also on the East Coast New York time zone so I'll specify that and then we should get um just to give you an example of what it could look like it'll be something like this where we have like 2024 and um the the month the actual day and the actual time in military and that's exactly what we're doing um just to give you an idea of what it could look like and then we also want to book the reservation once we have everything ready so I select my actual calendar um the actual name of the uh reservation in my calendar so I just put something um pretty simple so I just said reservation of the number of the guest so reservation of five um and then you can also add in the start date uh so this is going to be the actual date time so it will uh have the start date time or start date to be this date and time and then we don't really have to specify an end date we can just specify the duration so we'll make sure that every reservation is uh an hour and a half and of course you could kind of add to this you can you know make the event name a little bit more detailed you could ask the customer for their name and then add in that name in this sort of event or reservation that you're creating just make it a little bit more uh detailed and um easier to kind of understand going to hit okay on that and then we have the final web hook response where where we can return any sort of information that we want to return again like I've said it before this is always um we're going to assume that this is always um successful so we can just return anything um I just put something like true but I'm not going to really uh use this response um it's just going to be there for Bland to realize when the webbook is is kind of done um so you can hit okay on that make sure you save it and then also you can schedule it to to run uh immediately um that way whenever you do your test calls it will be ready to go um so we want to make sure we also copy this address and now go over to Bland and now in Bland we can actually paste that in there and save our pathway and that should be the entire pathway and we can actually get on to testing it out and seeing how it works so before we send the call again we we want to kind of test it with the chat mode so we can hit test pathway we want to start at the greeting and we don't really need to send any variables um we're not using any sort of um preall variables throughout the the call we we extractor our own variables but we don't really need any variables um before the call that we're going to be using so we can hit test pathway and then go from here so you can say um I'd like to know where you're located so we can test out the knowledge base node first okay looks good yes um can I book a reservation um and then we can also specify the date so we can say July um 15th or put 12th 2024 and we can sa for the time uh we can also test this military time conversion make sure it's doing it correctly so we can say something like 600 p.m. and for the number of guests um we can say something like 9 let's see how it does on the actual reservation so it says okay one moment I try to book your reservation and it should have booked the reservation so if I check my calendar and I do see the the actual reservation on July 12th um from 6:00 to 7:30 and it says reservation of 9 and it looks good like it's working fine so we can actually now test it on the call and see um that you know it's working fine on the call as well not just the chat now to actually send that call we can hit send call now and put in our phone number so if we're going to be calling ourself we can put our phone number right over here select The Voice voice um so I've selected floran and can also add in that request data if we're going to be any using any sort of pre-call data throughout our call so if we already know the first name of the person calling just by looking at their phone number we could add in something like that again we didn't do that for the chat so we we won't do it here and everything else we don't really have to add we can um make this English and Max duration is fine so we can go ahead and send a call once you add that phone number hi this is David from John's festar Barbecue how can I help you today hey I was wondering what uh address you guys are located at we're located at 13884 Main Street Chicago do you have any more questions yeah I wanted to book a reservation what date were you looking to make a reservation for uh July 12th 2024 what time were you thinking of coming in on July 12 2024 um I was thinking about uh 400 p.m. how many guests will be in your party uh just three all right one moment please while I try to book your reservation I've made a reservation for you on July 12th 2024 at 4 P.M for three guests we look forward to seeing you then goodbye all right thank you bye now for look at the actual calendar we should see the reservation is booked successfully so looks like the actual um AI color is working just fine so that is conversational Pathways in Bland AI if any of you would like to see detailed videos on things like custom tools or some of the other features in Bland uh feel free to to let me know and I do offer a free strategy session for any of you who have questions about AI calling and and um you know want to implement it in your business um feel free to book a call using the link in the description and yeah um thanks for watching

Channel: Ibrahim Munir

Views: 271

Keywords: bland ai conversational pathways, bland conversational pathways, ai phone caller, ai voice bot, ai voice agents, bland ai, conversational pathways

Id: d2gmD9MgesY

Length: 50min 38sec (3038 seconds)

Published: Thu Jul 11 2024