Loading...
Community presentation
# Maddie — Company Introduction & User Guide ## What is Maddie? **Maddie is Gaia Talent’s AI recruitment assistant.** Her purpose is to help us manage the early stages of recrui…
Create your own version
Use this visual direction as inspiration for a new Presenton deck.
Use this designCreator prompt
# Maddie — Company Introduction & User Guide ## What is Maddie? **Maddie is Gaia Talent’s AI recruitment assistant.** Her purpose is to help us manage the early stages of recruitment more consistently and efficiently — from receiving an application or speaking with a candidate, through screening and job matching, to recording useful information in Recruit CRM for the recruiter to take forward. Maddie is not intended to replace the recruiter or make final hiring decisions. Think of her as an **intelligent first layer between candidates, our live vacancies and Recruit CRM**. She helps us answer three questions: 1. **Who is this candidate and what are they looking for?** 2. **How strong is their profile, and how suitable are they for a particular vacancy?** 3. **If that vacancy is not right, do we currently have something better for them?** The recruiter then takes over with much more of the initial work already completed. --- # 1. How candidates enter the system There are two main routes. ### A. Maddie screening A candidate can speak directly with Maddie. This may happen because they are interested in a **specific vacancy**, or because they want to discuss opportunities with Gaia Talent more generally. Maddie conducts a conversational screening and collects the information needed to understand the candidate. ### B. Applications received by email Maddie also has a background application-processing system connected to recruiters' Gmail inboxes. When applications arrive by email, the system can identify genuine job applications, read the attached CV and extract the candidate's information automatically. This means we do not necessarily have to manually open every application, download the CV and create the candidate from scratch. Both routes ultimately feed the same candidate ecosystem. --- # 2. What information does Maddie collect? The objective is to create a useful, structured candidate profile rather than simply storing a CV or transcript. Depending on the information available, this can include: * Name and contact details * Current position and employer * Desired role * Seniority * Location * Skills * Years of experience * Relevant experience * Current salary * Salary expectations * Notice period * Willingness to relocate * Motivations / professional summary * Work history * Education * Certifications * Languages * CV * Fit assessment * Recruiter notes * Red flags Much of this information can then populate the corresponding candidate record in Recruit CRM. The candidate summary in Recruit CRM can also bring together information such as current position, desired role, seniority, experience, motivations, fit assessment, work history, education, certifications and languages. **This is important for users:** the quality of Maddie's output depends heavily on the quality of the information available to her. If salary, location, skills or another important field is missing or inaccurate, this can affect what Maddie understands about the candidate and, consequently, how effectively she can match them. --- # 3. Understanding the different scores This is one of the most important parts of the system to understand because **not every score means the same thing**. ## Initial Score — “How strong does this candidate look from their CV/application?” When a CV enters through the email/application route, the system can perform an initial assessment before a Maddie screening call takes place. The candidate receives an initial **0–100 assessment**, which is then stored in a score band: **0–20 | 20–40 | 40–60 | 60–80 | 80–100** The system also stores the rationale behind that assessment. This is primarily a **CV/application-based first assessment**. It should be treated as an early prioritisation signal, **not as the final verdict on the candidate**. For example, a CV may not clearly show salary expectations, motivations, relocation flexibility or experience the candidate has not written down. Maddie's conversation can subsequently provide a much richer picture. --- ## Candidate Score — “How strong is the candidate after screening?” After a Maddie screening conversation, the system can record a separate **Candidate Score from 0–100**. This is different from the Initial Score. The Initial Score is based primarily on the information available when the application/CV enters the system. The Candidate Score exists after Maddie's screening and reflects the richer information gathered through that process. So, as a simple rule: **Initial Score = initial CV/application assessment** **Candidate Score = candidate assessment after Maddie screening** Neither should be confused with the next score. --- # 4. Job Fit Score — “How suitable is this candidate for this particular job?” A strong candidate is not automatically right for every vacancy. Maddie therefore calculates a separate **Job Fit Score for individual roles**. For every open vacancy being considered, the matching system essentially asks two questions. ### Question 1: Does this role broadly make sense for this candidate? — 60% Maddie considers the overall meaning of the candidate profile against the job, including factors such as: **desired role + skills + seniority + location + salary** This uses semantic matching rather than simple keyword matching. For example, the system can understand that **“Junior Recruiter”** and **“Graduate Recruitment Consultant”** are conceptually similar even though the wording is different. ### Question 2: Does the candidate actually have the skills required? — 40% Maddie also compares the candidate's recorded skills against the skills required for the vacancy. Exact skills receive full credit, while related skills can receive partial credit. The two assessments are then combined: > **Job Fit = 60% overall profile match + 40% skills match** ### The current pass mark is 75/100 Maddie will only proactively recommend a vacancy when the combined Job Fit reaches **75 or above**. Below that threshold, the system is intentionally designed **not to offer a poor match simply because we happen to have an open vacancy.** This threshold can be adjusted as we learn from the system and our vacancy database grows. The 60/40 weighting and the amount of credit given to related skills are also configurable. --- # 5. Why accurate fields and skills matter This is particularly important when using Maddie. **AI matching does not mean the system can safely assume information that has never been captured.** For example, imagine somebody who is actually an excellent Environmental Consultant but whose profile contains very few properly recorded environmental skills. The overall profile comparison might recognise that the vacancy sounds right. But the skills component could still score poorly because the system does not have enough structured evidence. There is already an example in the matching logic where a candidate scored **83/100 for overall similarity but 0/100 for skills** because their skills had not been properly captured. The combined result was only **50/100**, meaning the apparently relevant vacancy was correctly not offered under the existing rules. For users, the practical lesson is: > **Good data in = better matching out.** Pay particular attention to **skills, desired role, seniority, location, salary expectations and experience** when reviewing a candidate. If something looks obviously wrong, consider the underlying profile information before assuming the matching engine itself is wrong. --- # 6. What happens when someone comes through a specific job? Suppose a candidate applies for **Role X**. Maddie knows which vacancy brought them into the process and conducts a screening focused on that opportunity. ### If the candidate is suitable for X Maddie can recognise the match and ask whether they want Gaia Talent to put them forward. There is no reason to search for another job simply for the sake of doing so. ### If the candidate is not suitable for X Maddie does not necessarily stop there. She can use what she has learned during the screening to search the other live vacancies for something more appropriate. However, there is an important protection built into this process: > **Maddie should only redirect someone to another vacancy if that vacancy is a better match than the one they originally chose.** So if: **Role X = 60% fit** **Role Y = 82% fit** Maddie can introduce Role Y. But she should not redirect someone away from their chosen opportunity to something the system considers even less appropriate. --- # 7. What happens when somebody hasn't applied for a specific job? Maddie can also conduct a **general candidate screening**. She first establishes who the candidate is and what they are looking for. Once she has enough information, she searches the current vacancies and evaluates the relevant opportunities. If she finds something that genuinely passes the matching criteria, she can explain the opportunity to the candidate and ask whether they would like to be represented. If nothing meets the required standard, Maddie can tell the candidate that there is nothing sufficiently suitable at the moment and retain their profile for future opportunities. **Finding no match is a valid outcome.** The objective is not to make Maddie recommend something to everyone. It is to make her recommend something **when we have good reason to believe it is relevant**. --- # 8. Candidate consent remains important Finding a match does **not** mean automatically submitting the candidate. Maddie introduces the opportunity and asks whether the candidate wants to proceed. If they agree, the relevant application can be created/progressed. This allows the automation to do the repetitive work without removing candidate choice from the recruitment process. --- # 9. One candidate can have different results for different jobs This is another important concept for users. A candidate does not simply become globally **“good”** or **“bad.”** They can be: **Rejected for Role X** while simultaneously being **Matched for Role Y** because those assessments answer different questions. The system maintains decided applications independently, allowing the candidate to be assigned to different jobs with the appropriate outcome for each one. That is why users should always look at **which score or verdict they are viewing and which vacancy it relates to**. --- # 10. How Recruit CRM fits into Maddie Recruit CRM is the recruiter's central operational workspace. The relationship works in both directions, depending on the information. ### Jobs: Recruit CRM → Maddie Live vacancies are pulled from Recruit CRM into the system. Maddie therefore uses the Recruit CRM job catalogue when searching for opportunities. Changes to jobs can also be synchronised into Maddie's database. **This makes keeping Recruit CRM vacancies accurate extremely important.** If a role is incorrectly left open, poorly populated or missing important information, that affects what Maddie has available to search and assess. ### Candidates: Maddie → Recruit CRM Candidate information and screening outcomes can flow back into Recruit CRM. That can include: **candidate details + CV + scores + assessment + application/job assignment + hiring stage + screening activity** Candidate records can also carry Maddie as their source, making it possible to distinguish candidates coming through the AI screening workflow. --- # 11. What happens to the Maddie call? The screening conversation does not disappear when the call ends. The system can create a Recruit CRM call log containing: * AI summary * Key points * Full transcript * Fit assessment * Recruiter notes * Red flags * Motivations This gives the recruiter taking over the candidate much more context than simply seeing **“Maddie screened — passed.”** They can understand what was discussed and why the system reached its assessment. --- # 12. What happens with incoming applications? The passive Email Agent complements Maddie's live screening. It can monitor connected Gmail inboxes, recognise genuine applications, identify the source, extract the CV and candidate information, create or update the candidate, perform an initial assessment and mark the email as processed so the same application is not repeatedly handled. Where configured, sufficiently strong candidates can also be automatically published to Recruit CRM with their CV and initial assessment. Recruiters can also receive alerts for qualifying candidates. So Maddie isn't useful only when somebody actively speaks with her. The wider system can also help us **identify and organise candidates arriving through our existing application channels.** --- # 13. What Maddie does — and what the recruiter still does The easiest way to understand the division of responsibility is: | **Maddie / System** | **Recruiter** | | ------------------------------------------ | ------------------------------------------ | | Reads and structures candidate information | Builds the human relationship | | Conducts initial screening | Reviews important/complex cases | | Captures candidate preferences | Validates anything that looks questionable | | Performs initial CV assessment | Provides context AI cannot know | | Searches live vacancies | Manages client relationships | | Calculates job-fit scores | Makes commercial/recruitment judgements | | Suggests genuinely relevant alternatives | Advises candidates | | Records screening information | Manages interviews and processes | | Helps update Recruit CRM | Owns the candidate after handoff | | Handles repetitive administration | Makes the final human decisions | Maddie's scores are **decision-support tools, not replacements for recruiter judgement**. A 74% fit does not mean a recruiter is prohibited from considering somebody, just as an 85% score does not guarantee that a client will hire them. The scores give us a consistent way to **prioritise, understand and compare** candidates and opportunities. --- # 14. The simplest way to think about Maddie The complete journey can be summarised as: **Candidate applies or speaks to Maddie** ↓ **System understands who they are** ↓ **Initial profile and relevant fields are created** ↓ **Maddie screens the candidate** ↓ **Candidate Score helps us understand the candidate** ↓ **Job Fit Score assesses them against individual vacancies** ↓ **Good fit? → Offer the opportunity and ask for consent** **Poor fit? → Search for something genuinely better** **Nothing suitable? → Keep the candidate profile for future opportunities** ↓ **Candidate, CV, assessment, application and screening information flow into Recruit CRM** ↓ **Recruiter takes over** --- ## The three things every Maddie user should remember **1. Understand which score you're looking at.** **Initial Score** is the early CV/application assessment. **Candidate Score** follows Maddie's screening. **Job Fit Score** is about the relationship between one candidate and one specific vacancy. **2. Data quality matters.** Skills, salary, location, desired role, experience and other fields influence what Maddie can understand and match. If a result looks strange, check the underlying candidate and job information first. **3. Maddie supports recruitment judgement; she doesn't replace it.** Her job is to screen consistently, organise information, identify promising matches, avoid obviously poor recommendations and reduce repetitive administration. **The recruiter remains responsible for the relationship, judgement and recruitment process.** The goal is ultimately very simple: **less time spent processing information, more consistent screening, better visibility of our candidates and vacancies, and more recruiter time spent actually recruiting.** Follow Design: {"palette":["Gaia forest #174C3C — primary headings, anchors, and trust signals","Sage green #8FB7A3 — secondary diagrams and supportive states","Warm ivory #F7F4EC — main presentation background","Charcoal #202824 — body text and high-contrast labels","Digital teal #20A89A — AI actions, process flows, and active states","Consent amber #E6A23C — cautions, thresholds, and human-review moments"],"fonts":{"Manrope":"https://fonts.googleapis.com/css2?family=Manrope:wght@200..800&display=swap","Inter":"https://fonts.googleapis.com/css2?family=Inter:ital,wght@0,100..900;1,100..900&display=swap","IBM Plex Mono":"https://fonts.googleapis.com/css2?family=IBM+Plex+Mono:ital,wght@0,100;0,200;0,300;0,400;0,500;0,600;0,700;1,100;1,200;1,300;1,400;1,500;1,600;1,700&display=swap"},"type":"Manrope in 700–800 weights for sentence-case headings with a clear 40–52 pt slide-title scale; Inter at 400–600 for concise body copy with generous 1.35–1.5 line height; IBM Plex Mono for scores, thresholds, field names, and workflow labels. Use bold sparingly to distinguish key rules and decision points.","layout":"A 12-column operational grid with generous 6–8% outer margins and an 8-point spacing rhythm. Use left-aligned titles, short text blocks, and modular two- or three-column compositions. Reserve the rightmost column for recurring section markers; keep score definitions and process stages in consistent horizontal positions across slides.","framework_treatment":"Warm ivory backgrounds with restrained white cards, soft 12 px corners, fine sage borders, and minimal shadows. Use teal connectors for automated actions and forest connectors for recruiter-owned actions. Present scores as distinct gauges or labeled chips rather than one generic visual. Use simple candidate, vacancy, consent, email, and CRM icons; diagrams should show clear handoffs and decision gates. Covers use one large conversational workflow illustration; content slides use modular cards and annotated flows; closing slides return to a simplified human-plus-AI handoff motif.","feels_like":"A trusted people-operations handbook blended with a calm, modern AI product interface"}
Presentation overview
This community presentation was shared by Steff Henrique as a complete 18-slide example. # Maddie — Company Introduction & User Guide ## What is Maddie? **Maddie is Gaia Talent’s AI recruitment assistant.** Her purpose is to help us manage the early stages of recrui… The preview lets you review the full sequence in order, rather than judging the design from a single cover image.
Use it as a reference for planning your own deck: notice how the amount of information changes from slide to slide, where visual emphasis appears, and how repeated design choices help the presentation feel connected. Community examples are inspiration, not locked templates, so you can keep the ideas that fit your audience and replace anything that does not.
Learn from the example
Look at the job each slide performs. Identify where the deck introduces its subject, develops the main points, adds evidence or examples, and moves toward a conclusion.
Notice which element attracts attention first and how headings, supporting text, images, and data are separated. Strong hierarchy makes the intended reading order obvious.
Compare spacing, typography, color, and repeated components across the slides. Consistent rules help varied content feel like one presentation instead of unrelated screens.
Make it your own
Begin with your audience and the decision, lesson, or action the presentation should support. Replace the topic, examples, evidence, and visuals with material you can verify, then edit every slide for one clear takeaway. You can borrow the visual direction without copying the creator's wording or message. Presenton will use this community deck as a design reference while you develop content for your own purpose.
