Why is it important for organizations to consider using AI and automation in their recruiting?

To save money – recruiting is expensive. Smart use of AI and Automation can help your in-house recruiting team achieve more.

The game is changing – AI is shifting the way recruiters can access and analyze talent. If your recruiting function is not using the latest tools, technology, and processes, it could indicate a need for upskilling or leadership. At the same time, the level of sophistication candidates perceive can impact who you hire. Staying updated with industry trends and advancements is crucial to attracting top talent. 

Tools

What are the different AI/Automation tools that can help you through the recruiting process? 

ChatGPT/LLMs – this is a great starting point for founders, recruiters, and anyone in the early-stage startup space. If your company has less than 500 employees, ChatGPT is going to be a valuable asset in automating your recruiting process. 

Note: turn off chat history & training for ChatGPT use – on the bottom left of the interface, you can click on settings and beta → data controls and turn off Chat history & training. This ensures your data will not be used to train the model. Still, never include proprietary information in your prompts to ChatGPT.

Notion – this tool has several AI functionalities that can be useful in recruiting. It’s a great alternative to Google Sheets, offering more advanced features that can streamline your recruiting process. It allows features like:

• Dynamic templates – you can create dynamic templates that update your interview process or sourcing strategy across multiple roles with one click of a button. 
• Customizable prompts – it has built-in functionality that allows you to create your own prompts for use over and over again. These prompts can be kept in a Notion AI block within your templates and used every time you open a new role. This helps in creating sourcing strategies, writing job descriptions, and generating sample interview questions based on what the hiring manager said was most important. 

ATS (Applicant Tracking System) – an ATS (like Ashby, which offers ample automation, or Greenhouse) is a must-have tool in any recruiting process. It can help you with helpful time-saving automations like:

• One-click profile addition – you can go to a candidate’s profile, click one button, and add that candidate to your ATS. This functionality is in most modern ATS, but some recruiters may not be aware of it—capturing their full name, LinkedIn URL, or other source, and automatically finding their email address
• Ability to sync candidate data – being able to mass upload candidate data as a CSV is a form of automation that can be useful when native integrations don’t exist. 
• Boolean string searches in ATS – knowing that you can do Boolean search in an ATS allows you to create alerts for when you source or receive an application from a strong candidate that fits your criteria. 

Note: Hold regular ATS demos and look for updates – even if you’re not in a position to change your ATS or other major tools, it’s beneficial to get a quick demo of something else out there once a quarter to stay updated on developing functionalities. Look for functionality when selecting like:

• Automating sequences – your ATS should allow you to automatically send a series of sequences as yourself or as hiring managers. If it doesn’t offer this functionality, it may not be the best choice.
• Integration with third-party tools – the ATS you choose should integrate with the third-party tools you’re already using. This will help you automate a lot of your processes.
• Boolean search functionality – this is a fundamental aspect of recruiting. Your ATS should allow you to perform Boolean searches and set up automations, alerts, and triggers based on these searches.

How can you leverage LLM APIs to integrate AI functionality into your tech stack? 

LLM API + Google Sheets to create a JD and start to find candidates– the OpenAI API can be connected to Google Sheets to automate parts of the hiring process. For instance, when a hiring manager has a requisition that has been approved, they can provide certain information before the intake call. This information can be input into a Google Sheet, which, when connected with the OpenAI API, can automate the next half of the process. This includes suggesting job titles based on the description provided by the hiring manager, writing the first round of a job description, and creating Boolean strings to find potential candidates on LinkedIn or other sourcing platforms.

LLM API + RecText for personalized texting – you can set up automations to send a personalized text to candidates when they apply and at key steps using a tool like RecText. The text can use context found on the candidate’s LinkedIn, resume, or cover letter. If candidates are communicating with various people within the organization, you can ensure that all communications are centralized and easily accessible.

LLM API + your ATS for opportunistic filtering – in your applicant tracking software (ATS), you can open a role called “opportunistic hires”. This allows you to track candidates who might be a good fit for future roles. By setting up alerts in your ATS, you can be notified when a candidate applies with skills or a job title that match your open roles.

LLM API + Slack for Real-Time Alerts – APIs can be integrated with Slack to provide real-time alerts when certain conditions are met. For example, if a candidate applies to the “opportunistic hires” role and mentions a certain skill or job title in their application, you can receive an alert on Slack. This allows you to stay updated on potential candidates without having to constantly check your ATS.

AI in Recruiting Use Cases

How can you incorporate AI when opening up the spec for an open role? 

Before AI gets involved:

• Establish a requisition process – this prevents recruiters from spending time on tasks they shouldn’t be working on. An approval process can be as simple as clicking a button that says approve or decline in an Excel sheet. Once it’s approved, the hiring manager should receive a form to fill out before they get on a call with a recruiter. 
• Utilize an intake form to gather information about a role – the most important step to incorporating AI into your recruitment process when opening a new role is by honing in on an intake template. This should be your first step if you’re looking to add AI into your recruitment process. A hiring manager will be able to gather context, fill it out, click submit, and from there automate much of the rest of the process.

The easiest way to open kick-off is to create a Google sheet that will auto-populate with – using the OpenAI API, you can connect a google doc to ChatGPT and kick off a role with: 

• 10 job titles
• A job description based on a few sentences  describing the role, duties, and experience you’re looking for
• An interview scorecard
• Questions for each stage of the interview process based on the interview plan and topics being discussed at each stage
• A scorecard to evaluate responses
• Sequences for outreach
• Emails to schedule the interview process
• Ads for the role
• Boolean strings & a sourcing strategy 

Example Role Kickoff Google Doc

See Destiny’s video on building an AI-enhanced recruiting Google Sheet form.

Tips on writing a prompt for an LLM JD:

• Define the role clearly – start by stating the role you’re hiring for. For instance, “Write a job description for a graphic designer who will help with the rebrand of our website.”
• Specify required skills – clearly mention the skills required for the job. For instance, “The candidate must be familiar with using Figma.”
• Describe the ideal candidate – provide a brief description of the ideal candidate. For example, “The ideal candidate has helped with a rebrand of a B to B health tech startup that targets dental practitioners.”
• Incorporate company culture – reflect your company culture in the job description. For instance, “Our company culture is cheeky, feel free to be funny.”
• Provide context on desired experience – give more context to the job description. For instance, “Be sure to mention that we’re looking for someone with at least three years of experience using Photoshop and Illustrator.”
• Feed in a sourcing template – a sourcing template can provide a foundation for your job description. It can include sections about the company, the role, requirements, nice-to-haves, benefits, perks, and pay.

How can AI tools help in taking a role to market?

NOTE: Different platforms have different character limitations for boolean strings – for instance, LinkedIn Recruiter and Google have different character limits. Therefore, it’s important to format your Boolean strings according to the platform you’re using—ask ChatGPT to format it to the tool that you’re using, sometimes it knows the character limitations. 

Use AI for candidate research – ask an LLM for help finding communities where potential candidates hang out, identifying blogs they read, or compiling a list of competitors. This reduces the need for manual data scraping and organizing. 

Tip: Build a strong recruitment process before adding AI – if your recruitment process is flawed, AI won’t be able to improve it. Instead, it might exacerbate the existing issues. Therefore, ensure that your recruitment process is solid before integrating AI into it.

How can AI help with sifting through applications for a role and evaluating candidates?

Humans will still need to be involved in generating interview questions – expertise matters. While AI can generate a first draft, the final version should be reviewed and possibly revised by experts who do the job. The more specific you are in providing context, the less likely you’ll need to make substantial edits.

Other AI use cases in the evaluation process include: 

• Balancing out your team’s hiring needs- understanding the level of a candidate can also help balance the team. If you’re hiring a more junior person, you might need to hire a senior person to balance that out. AI can help you keep track of the levels of candidates you’re interviewing and hiring.
• Communicating with Candidates – once a candidate’s level is determined, you can communicate with them about potential compensation and whether they want to continue with the process. This helps to manage expectations and ensure transparency in the hiring process. 
• Creating a careers page from scratch – it can help design a careers page on your website for you.
• Reformat employment contracts – ask an LLM to reformat employment contracts for upload into the Applicant Tracking System (ATS). It can pull in dynamic information like the candidate’s name, start date, and salary.

What are examples of prompts that can help you use AI in your recruiting process?

At what points in the recruiting process should you avoid using AI? What use cases is AI bad for?

Despite the automation, humans play a crucial role in:

• Creating systems and processes – AI doesn’t appear in your organization, someone has to set it up.
• Orchestrating candidate feedback – the evaluators speak with each other, aggregate feedback, and come to a decision.
• Managing systems and integrations – they ensure integrations are working, and troubleshoot any problems. 
• Collecting overdue feedback and intake details – humans have to ensure that the right information has been collected throughout the process. 
• Offering to the candidate – candidates are unlikely to accept an offer from a robot. Human interaction is still necessary, especially for high-level positions.

What is all of the automation freeing up time for recruiters to do? 

With less time spent on repetitive tasks, recruiters are now freed up to:

• Format and manage data – the role has evolved to include a significant amount of data management. This involves gathering necessary information about qualified candidates, such as names, email addresses, and LinkedIn profiles, and uploading it into an ATS.
• Find creative ways to source candidates – not all candidates can be found on LinkedIn or other traditional platforms. Recruiters are spending more time discovering new and creative avenues to find potential candidates.
• Set up systems and tools – as companies grow and evolve, so do their recruitment needs. Recruiters are often tasked with setting up new systems, like their ATS, to streamline the recruitment process. This can involve onboarding team members to the new system and ensuring all necessary documents are in place.
• Ensure compliance and accuracy in offer letters – each role within a company may require a different offer letter. It’s the recruiter’s job to ensure that each candidate receives the correct offer letter, whether they’re a full-time employee, a contractor, or a commission-based sales rep. This can involve troubleshooting issues with HR and ATS systems to ensure data is correctly transferred and processed.
• Meet with executives – regular meetings with company executives are a crucial part of a recruiter’s role to stay aligned with the company’s hiring needs and strategies.
• Focus on training – training team members on how to conduct interviews is another important task for recruiters. This ensures that interviews are conducted in a compliant manner and everyone is on the same page.
• Improve candidate experience – a positive candidate experience is key to attracting and retaining top talent. This can involve creating detailed job descriptions, outlining benefit packages, and ensuring the company’s offerings are clearly communicated.

Improving Recruiting Process

Where should recruiters look to implement more traditional automation in the recruiting process?

Maximize existing functionality before adding new tools – before getting excited about adding any other tool, AI, automation, or anything similar, examine how you’re utilizing the existing functionality in your sourcing activities, scheduling, running reports, offer letters, and sifting through candidates. Check for functionality in:

• Sourcing Activities for example, conducting a simple Boolean search in your Applicant Tracking System (ATS) can be a powerful tool for identifying potential candidates. 
• Setting alerts – once you’ve set up your Boolean search parameters, you can create an alert that notifies you when someone fits that specific search. 
• Offer Letters – the process of creating and sending offer letters can be manual and time-consuming. 
• Scheduling – automating scheduling can be a game-changer, but it requires diligence. Some people might be too lazy to add whoever needs to be scheduled, which defeats the purpose of automation. 

Establish a recruiting Slack channel – and create communication guidelines. Everyone should know what channels to use and how to communicate information. Be mindful of what information is shared and where, especially around sensitive topics like compensation.

Establish and automate the offer delivery process – when an offer is given, it requires a whole process. This includes notifying the candidate, prepping the offer, sending the offer, archiving the candidate, and doing the manual push in the HRIS where it can’t be automated. 

What kind of recruiting templates should you create and store in your organization’s knowledge management tool?

Store templates like:

• Job Descriptions, Candidate Profiles, and Interview Scorecards – these are more specific versions of the templates discussed earlier. They include job descriptions, candidate profiles to outline what you’re looking for, and interview scorecards for each round. 
• Sequences – these are templates for messaging that works for different roles, such as sales, engineers, founding talent, etc. 
• Job Ads – Create a template for job ads and tracking links, especially if you’re partnering with other departments like marketing. 
• Software Evaluation Tracker – keep track of the different companies that pitch you on their software or different software that you’re evaluating. 
• Offer checklist – to ensure that all the necessary steps are taken when delivering an offer. This includes discussing all the necessary points in a certain order and selling the candidate on the role. 

Setting up a recruiting locker – a recruiting locker is a centralized hub where all your hiring-related documents and processes are stored. This includes:

• Headcount and requisition documents – these are the first things to be set up in your recruiting locker. Every time a new role gets approved, you follow the process outlined in your requisition document.
• Sourcing Document Template – this is where your sourcing document template lives. It’s a blueprint for sourcing candidates for each new role.
• Weekly Hiring Sync Notes – these are the notes from your weekly hiring meetings. They contain specific things that you should be discussing each week. You can create a database in your knowledge management tool and use it as the hiring sync document.

How can you use AI in recruiting/HR org process improvement? 

Create meeting transcriptions for clarity and accountability – recording intake calls with hiring managers can be a game-changer. Transcriptions of these calls can serve as a record of the kickoff, capturing the exact words of the hiring manager, and the agreed-upon action items, timelines, and responsibilities. This information can be formatted as a CSV for upload or automatically pushed into a project management tool.

Automate time management feedback – by taking a screenshot of a recruiter’s calendar and aligning it with their OKRs and goals, recruiters can receive feedback on how they’re using their time. This can help them batch tasks, be more organized, and meet their targets more effectively.

Use gamification for recruiter motivation – gamifying tasks can make them more engaging and enjoyable. Seeing tasks disappear or achieving goals can be rewarding and motivating, leading to increased productivity and effectiveness.

AI can provide real-time interview feedback – it can automate the collection of feedback from various sources such as candidate experience surveys, Glassdoor reviews, and employee feedback. This real-time data can be used to identify issues, such as a hiring manager who is performing poorly in interviews, and take corrective action. 

Overall 

What are the most important things to get right?

Understanding and documenting your process is key – if you don’t know what you’re doing and lack context on how you’re doing it, it’s unrealistic to expect another person, let alone an AI, to understand. If you’re getting poor results, it’s likely because you’re not describing your process adequately.  If an AI struggles to understand you, it’s likely that people around you are struggling too. This can be a valuable opportunity to improve your communication and process documentation skills.

Creating a brand guideline can be helpful – if you’re unsure where to start, consider using AI to outline a brand guideline for you. This can provide a clear framework for your process and make it easier for both people and AI to understand.

Remember, it takes time and effort – leveraging AI and automation isn’t something that will happen overnight. It requires effort and understanding, but with consistent use, you’ll catch on quickly. 

What are common pitfalls?

Over-reliance on tools – companies often resort to investing in multiple sourcing tools when they struggle to find enough candidates. However, adding more tools is only beneficial when you’re ready to scale something. If you’re not getting good results with one tool, it’s unlikely that adding more will solve the problem. Instead, it’s crucial to focus on improving your process.

Failure to integrate existing tools – before investing in new tools, consider how you can make the ones you’re already using work together. This could involve using APIs or low-code/no-code tools like Zapier. 

Lack of professionalism – it’s important to maintain a high level of professionalism in all your operations. This includes proofreading all your outputs to avoid errors and ensure the quality of your work.

Pay attention to data privacy and security – ensure that the AI tool you’re using doesn’t use the information you share to train their data. Be mindful of using custom GPT keys as they can feed your entire interview process into their training data, potentially making it public information. When you use ChatGPT, you’re trading your data for access.