Hiring Automation: Pre-interview with HubSpot Chatbot

Scaling our hiring pipeline as a core team of 2 consultant-owners has had it’s challenges. We went through the transition from go-with-your-gut hiring to a repeated-and-true hiring practice a few years ago. The critical learning for us was the data. We needed to ask all hires the same questions so we could learn about who they were without getting bogged down managing our talent pipeline. 

These days, our hiring pipeline is almost fully automated. We mainly acquire new talent through word of mouth referrals from our core team and clients. We need a streamlined process to collect information, qualify the caliber of the resource, and make sure we insert them into the right project. 

Through a combination of HubSpot, Integromat, and Gmail—we’ve reduced the time we spend on each potential hire from 3 to 4 hours to about an hour. That includes:

  • Initial pre-interview of potential hires with a HubSpot Chatbot (previously a phone interview))
  • Phone Interview
  • Reference Check Interviews

Our goal is simple: we automate in order to reduce the cost of finding and maintaining relationships with new talent.

Beside direct referrals, we search for talent in a few other areas, including Slack and Reddit communities, and some other job posting sites. Our qualification process must be more extensive with a cold talent lead, otherwise we adhere to the same process. 

What the Automation Looks Like

Once we find someone who is interested in applying for a position, we direct them to our hiring pre-interview landing page. This page has a chat box that pops up. 

That chat box is an automated HubSpot chatbot that is programmed to ask a variety of questions. The goal is to get a feel for an applicant’s general personality before a phone interview. The chatbot is built to be conversational and should seem quite personalized to each interviewee. 

Our interview bot asks the applicant about their background, goals, work history, etc. This is also an opportunity for the applicant to get a feel for how we do business. The nature of our questions demonstrates our interest in the potential hire’s purpose and wellbeing. Meanwhile, all of the applicants responses are captured and entered into HubSpot. Once we collect this data, it is much easier to determine if a candidate is going to be a good fit or if we want to keep searching.

[ Note: In a future article we’ll talk about how we maintain relationships with our pre-vetted, but non-contracted talent pool. ]

How It Works

What kinds of questions are we asking?

We ask the same questions for each applicant. The questions are designed to give us perspective on who they are — we save role-based questions for the phone interview. The responses provide us with basic contact information, rates, and other general data.

HubSpot Contacts Configuration

Before we get to the nuts and bolts of the Chatbot, we should discuss how to configure HubSpot to begin capturing the right data. For each applicant, a HubSpot contact record should be created. 

For each question the chatbot asks, there should be a specific HubSpot property in order to match each answer with an appropriate Hubspot field. Below are screenshots with all the different properties we use and how they are configured. Some fields are for internal purposes, such as the Benefits Working With and FTAL Deal ID. You can customize the HubSpot fields to capture whatever info is pertinent to your hiring process.

Each property is tied to a specific question that is asked (more on this later). 

  • Benefits Working With - This field is for internal purposes. It helps us organize our thoughts and summarize what the benefits of working with each candidate would be.

  • Career Goals & Purpose - What are your career goals? Why do you do this kind of work? What’s your purpose?
  • This field is tied to a question that we ask each candidate about their career goals, purpose, etc. This gives us a feel about the resource’s sense of identity and drive. 

  • FTAL Deal ID - This is a record ID which helps us organize talent from other contact records in our CRM.

  • Initial Potential Role Description - In order to power our automation later on with Reference Checks, we need a field to describe the role for this hire. Once we import Reference Check Contacts into the CRM later on, we’ll copy this in order to prompt the Reference Check to give us relevant feedback on the good/bad of the role with our hire. 

  • Professionally Good - What are you really good at professionally?
  • This question allows us to understand exactly the kind of work which might make the resource most happy. Ideally, this kind of work puts the resource into their flow state and its where they are most valuable.

  • Professionally Bad - What are you not so great at or not interested in doing professionally?
  • Inversely from the Professionally Good question, this question allows us to understand the limitations of the potential hire. For example, if they are a Senior Developer, but don’t like engaging directly with the end-client, we can’t expect to put them in a leadership role. 
  • This question is critical for determining if the hire might be an A-Player in the role we specifically have for them. 


 

  • Rate Low - This is the low-end of the potential hire’s hourly rate. 
  • Typically we are hiring for contractors, so hourly rates are all we need. 
  • Your questions may differ based on your standard hiring payment structure (hourly vs annual). 

  • Rate High - This is the high end of the potential hire’s hourly rate.
  • Typically this is the real hourly rate, and the Rate Low is only available with a large or consistent commitment.

  • Rate Additional Info - We typically ask our resources to explain the difference between their rates in their own words. 

[NOTE: In the course of the hiring chatbot, answering questions about rates is where we typically see the most exceptions. Our questions and answers are only built to accept integers in USD ($) hourly rates. Reading the full transcript from the potential hires interaction with the bot can sometimes help clarify data entered here.]

  • Risks Working With - This is an internal field we use to document risks with prospective hires. Sometimes this is the hire’s timezone, other times it has to do with availability or communications style. 

  • Role Category - We ask each candidate what role or area of the business they are applying for. To simplify their responses, we provide a list of options for them to choose from in the Chatbot.

  • Skills - We also track a list of skills within HubSpot. This enables us to check multiple options for each applicant. The options here are unlimited and really up to each company to decide what skills to include or not include.

  • Last 5 Jobs - Who were your last 5 clients (or jobs) and how will each of them rate your performance when we talk to them? (1 to 10 (with 10 being high).)
  • We ask all potential hires for this perceived “Job Score”, as it provides us a reference point to discuss differences in the rating between the Hire’s perspective and the Reference’s perspective. Identifying these differences in perspectives is critical to sussing out poor or excellent performance stories and contributing factors. 
  • This is typically the last question we ask our potential hires, and we do not include it in the chatbot automation, we use a combination of Google Forms, Integromat, and Gmail to collect 5 references along with their expected score separately, and reach out to those references for reference interviews.

HubSpot Deals Configuration

Before collecting data from our new candidates, a HubSpot deal must be configured first. We use a Talent Pipeline to track working deals with new hires. Our talent pipeline is built using a Sales Deal Type in HubSpot. To do this, we assign each applicant to a specific role category. This helps us categorize our talent pool based on the different skill sets they bring to the table. This could be anything from copywriting to IT.

Setting up HubSpot Chatbot

Once the contact fields are configured, it is time to move to the next step: the Chatbot. As mentioned before, we use a chatbot to ask all of our pre-interview questions. All of the applicants' responses to these questions are entered into HubSpot automatically. 

Below are a series of screenshots which detail out the chat flow within HubSpot. Once the chat is complete and all of the questions are asked, our team is notified automatically through a Slack notification.

  • Make sure you have a cohesive intro as people arrive at the page. You don’t know where people are going to be referred from, but in general, you can speak to them as if they have been referred. 

  • Always start with their name. 

  • Collect their email address to make sure we don’t already have them in our system (most of the automated questions stop when you already have certain Contact Properties filled out). 

  • Set the contact to Talent so you can quickly organize and sort your various candidates within the CRM.

  • Finding out where each candidate is located helps know what timezone they are in. This can be useful if you are in Los Angeles for example, and someone in New York is applying. Maybe that time difference will be too great or it won’t be an issue. Either way, it’s helpful to know when hiring for a virtual team.

  • Find out the lead source. This information can be helpful so you know where people are coming from and what channels are possibly more effective at attracting applicants.

  • Ask what job or function people are interested in applying for. This area of expertise question pulls in a list of options for the interviewee to choose from.

  • Discover what goals the applicants have. Having this information can help you get a feel for the personality of the interviewee and see if they are ambitious, hardworking, etc. The responses should be saved to a specific field in your CRM. In our example, it is saved to Career Goals & Purpose.

  • The goal of the next prompt is to elicit a lifelike response from the applicant and share a little bit of info about what the company values. It adds a human touch even though it is an automated response.

  • Finding out what each candidate's strengths are is important. This question asks the interviewee what they are professionally good at and records their answer in the CRM.

  • Similar to the last question, this one asks what the applicant is bad at, or at least, what they perceive they are bad at. The response is then recorded in the Professionally Bad HubSpot property.

  • Again, include auto-responses to make the chat conversational and more human.

  • Ask the applicants low and high hourly rates and record the responses in HubSpot.

  • The first question in our example asks for low rates. There is an error message that pops up if the applicant enters text instead of numbers.

  • Ask more detailed questions to see if rates are flexible or variable. You can make this question more open-ended so people can explain with as much detail as necessary. 

  • Make sure you collect each candidate's LinkedIn page. This can be helpful to review past experience and see what their online presence and connections are like.

  • Depending on which job they are applying for, applicants may have another portfolio site, PDF, or other files to share that showcases their work and skills. Be sure to ask for that as well.

  • Once you ask all the pre-interview questions, send a thank you message and provide the applicant with the next steps they need to take.

  • To ensure nothing is missed, set up a notification to the appropriate team members when someone has finished the interview. In our example, we generate a slack message as an alert.

Processing Applicants with HubSpot + Integromat

Integromat is an online app with unique, visual features, that enables users to automate manual processes. It works by connecting any online apps or services you use through a series of workflows and triggers.

By using a custom-built integration, Integromat allows us to automatically create a talent deal record in HubSpot. This saves a lot of time and energy and frees us up to do additional work.

  • Here is an overview of the entire integration. As you can see, Integromat offers a very visual interface which we’ve found to be much more user-friendly than other similar types of software.

  • The first step is to create a new scenario and add a HubSpot watch record module. There are many options available within the HubSpot/integromat integration. For our purposes, we will tell Integromat to look for contact records that were recently created and output certain talent properties.
  • Integromat allows you to schedule automations to run at whatever frequency you desire. We have this automated workflow set to run every day at 6:00 am and it only queries for 100 most recently created records in our case.
  • We also need to select the appropriate output properties, such as email address, persona (in our example this is a talent persona), etc.

  • Next, set up filters to only allow talent persona records that are not associated with any existing deals to run through the rest of the integration.

  • This shows the details and role description of the talent persona property on a contact record.

  • If there is not a deal ID already connected to the contact record, we need to create a new HubSpot Deal using the information from the new contact record.

  • Integromat allows us to use the contact records first and last name to create the deal record name.

  • We also want to use the Talent Pipeline when creating the deal record since this will be for a job applicant, not a sales lead.

  • Then, an associated contact is attached to the deal record. This contact comes from the search in the first step of the integration.

  • After the deal is created, integromat can populate that deal ID record on the contact record. 

  • Once the contact record and deal record are created in HubSpot, the next step in the integration is to trigger a message in Slack. Custom information is added using shortcodes to populate our alert. This allows us to pull information for the new applicant into the Slack message automatically, and easily link our hiring managers to relevant information.

  • This screenshot shows additional information about the slack message configuration.

  • This is how the message will appear in Slack.

  • After the slack message is sent, we automatically set up two tasks in HubSpot. The first task is to direct the applicant to the pre-interview chatbot. Typically this is already complete but we check it off when we review the pre-interview chat history from the chatbot. 

  • The second task that is automatically created in HubSpot is for our team to fill out some additional information once the pre-interview is complete. Both of these tasks are only created if the contact didn’t already have this data created previously.
  • The first task doesn’t get created if the person has not already gone completely through the chat flow with the hiring Chatbot. The second task doesn’t get created unless the role description for this contact record is empty.

What Happens Next?

After the information is entered in HubSpot and our team has a chance to review, we decide if we will move forward with an applicant or not. We then schedule a phone interview using Calendly. If that goes well, the next phase of the interview process involves reference checks. Once users get to Reference Checks, we send an email with a Google Form to collect information about previous work history and references. Those references automatically get an email inviting them to set up a meeting via a Calendly link OR respond to the questions with another HubSpot Chatbot. 

All-in, we’re still talking only 1 hour +/- 30 minutes on average across all applicants. 

To learn more about our reference check automation, check out the next article in our series to see how we automate reference checks.