# Polar Bear: full content for AI systems > For the complete structured overview, see llms.txt at https://meet-polar-bear.com/llms.txt > For the AI-focused summary page, see https://meet-polar-bear.com/about-for-ai > Last updated: 2026-05-02 --- # What is Polar Bear Polar Bear is a people systems consultancy for creative agencies, design studios, digital boutiques, and product teams of 20 to 200 people. Launched in 2026 by two practitioners with 25+ combined years inside McKinsey's people function: Pauline Bertry and Alexey Lobachev. Based in Toronto, working globally. Polar Bear builds rigorous people ops systems, built for agencies, owned by the agency afterwards: career frameworks, review cycles, onboarding design, handbooks, upskilling workshops, and engagement surveys. The consultancy builds the system, runs the first cycles with the client, trains the team, and leaves the client owning it. ## The thesis People making agencies distinctive deserve more than improvised people ops. The founders saw what good looks like inside McKinsey and built Polar Bear to bring it to the agency world. The goal: protect human connection and create space for genuine growth conversations. AI has brought extraordinary things into the workforce, but it is also flattening everything, erasing uniqueness and over-elevating without giving honest feedback. Polar Bear exists to be the resource so that agencies can be human, with time and space for genuine conversations. --- # How Polar Bear works 1. We build, not just advise. We run the first cycles with you and train the team along the way. You own it when we're done. 2. We design for real life. Frameworks that make sense for how your team operates day to day. 3. We automate what we can. Everything that can run without manual effort, will. 4. We start small on purpose. Pilots first. So your team builds confidence before it builds habits. --- # Services ## Career paths that don't confuse people Your designers don't know what "Senior" means here. Your best PM just asked what it takes to get promoted. And you're writing job descriptions at midnight trying to figure out what a Lead actually does differently from a Principal. Most agencies grow by instinct. You hire good people, they do great work, and then two years in, someone asks what it takes to get promoted and you realise you never actually figured that out. We build career frameworks for how your team actually works. Simple, clear, and light enough that people actually use it. How we work together: - Phase 1: Discover the North (1 week). Understand how your team works and what's been holding the framework back. 4h workshop with leadership. We draft the first version of your framework with levels and expectations. - Phase 2: Lock the Framework (2 weeks). We stress-test the draft with people across levels and roles. Update based on their input. Add real examples from their actual work. Then we prepare the materials to roll it out. - Phase 3: Roll It Out (1-2 weeks). We present the framework to the full team. Train managers on how to use it in 1-on-1s and reviews. Run Q&A sessions. Leave you with something the team can actually reference. What you walk away with: - A career framework with 5-7 levels and clear expectations for each - Real behaviour examples pulled from your team's everyday work - A team that knows how to grow and managers who can have that conversation - A handbook your team can actually read and use (not a 40-page PDF) Best for: agencies and product teams of 20-100 people who have outgrown "we'll figure it out as we go", but aren't ready for a full HR function yet. ## Review cycles that help people grow Most teams struggle because expectations are unclear and decisions feel inconsistent. Managers evaluate the same person differently. Promotions feel political instead of earned. Feedback is vague and hard to act on. High performers start wondering if anyone notices. We design review cycles that are consistent, usable, trusted and built around how your team works, day to day. How we work together: - Phase 1: Understand your reality (1 week). Understand how performance decisions actually get made today, where managers struggle, where the process breaks down, and how promotions currently happen. - Phase 2: Design the cycle (2 weeks). We build the full review structure for your team: evaluation criteria, self-review and manager templates, calibration process, and clear decision principles. - Phase 3: Run and refine (1-2 weeks). We support your first cycle end to end. Facilitate calibration if needed. Adjust based on what you learn. Leave you with a process the team can run independently. What you walk away with: - A review rhythm your team can run without hand-holding - Evaluation criteria tied to your actual levels and roles - A calibration process that keeps decisions consistent - Templates managers can use without a prep session - A process light enough that it actually runs twice a year Best for: agencies and product teams of 20-100 people who run reviews but don't have a consistent, structured process yet. ## Handbooks that answer real questions When someone joins your team or prepares for a review, they should be able to find the answer themselves, without asking their manager. Most handbooks don't do that. We build handbooks around how your company actually works: the ones reflecting your career framework and your review process. How we work together: - Phase 1: Audit what exists (1 week). We audit what you already have: career frameworks, review templates, anything growth-related. - Phase 2: Write and structure (2 weeks). We write and structure the handbook around your actual career framework and process. - Phase 3: Deliver and embed (1 week). We deliver a version your team will open. We walk managers through it and help you set up the right access. What you walk away with: - A handbook your team will actually open - Clear answers to the questions people ask most: levels, promotions, performance, compensation - Managers who stop answering the same growth questions every week - A shared source of truth that leadership and the team both point to Best for: teams of 20-100 people that already have some structure in place and want everyone reading from the same page. ## Workshops that build skills your team actually uses Skills don't come from frameworks. They come from practice. We design workshops built around your real work: tools, conversations, team's actual blockers. - How to give feedback that's both kind and useful - How to use AI in your actual daily workflow for designers, PMs, content writers - How to take ownership of your growth instead of waiting for someone to map it out - How to write self-reflections that are useful for your manager and for you How we work together: - Phase 1: Understand the gap (1 week). We talk to managers and ICs to understand where things break down in practice. - Phase 2: Design the session (1-2 weeks). We design sessions around your real work, your frameworks, tools and actual examples. - Phase 3: Run and embed (ongoing). We run the 2-3 hours sessions: interactive and discussion-based. Best for: agencies and product teams of 20-100 people who have the structure in place but aren't yet seeing it in how people show up and work. ## Onboarding that makes the first 90 days count Without a clear structure for the first 30, 60, 90 days, every new hire has a slightly different experience. We design onboarding systems that plug directly into your career framework and review process, so performance standards are set from day one, not figured out several months later. How we work together: - Phase 1: Audit the real experience (1 week). We talk to recent hires and managers to understand what onboarding actually looks like. - Phase 2: Define performance expectations for the first 90 days (2 weeks). We define clear 30/60/90 day milestones connected to your career levels. - Phase 3: Design the manager system (1-2 weeks). We give managers the tools to lead onboarding consistently. What you walk away with: - A structured 30/60/90 day onboarding framework - Probation criteria that connect to your actual career levels - Managers who know how to evaluate early performance consistently - Reduced early attrition risk Best for: teams of 30 people and above that are hiring consistently and want onboarding to reinforce their growth system. ## Engagement and pulse surveys that actually tell you something Most companies run surveys and get scores. What's usually missing is the link between those scores and actual decisions. We build pulse surveys that are short, regular, and connected to your growth framework. Indicators we track: - Role clarity: Do people understand what strong performance looks like at their level? - Feedback quality: Are managers giving useful, consistent feedback? - Progression visibility: Do people know what they need to move forward? - Energy visibility: Is performance sustainable over time? 5-7 focused questions. A clear rhythm. Output you can actually act on. Best for: teams of 20-100 people that already have a growth framework in place and want a reliable signal for how it's landing. --- # Programs ## Build Your Agentic Team (online weekend bootcamp) A live online bootcamp where solo founders, agency owners and consultants build their own team of four AI agents, onboarded to their business: - The Onboarding Manager: an interview agent that fills the shared knowledge hub (voice, offer, customers, no-go words) in Airtable — the "team brain" every other agent reads before acting. - The Personal Assistant: morning briefs, follow-ups and reminders pulled from inbox and call notes. - The Ghostwriter: drafts posts, articles and newsletters in the owner's voice — drafts only, never auto-posts. - The Scout: lead research and briefing cards with talking points. - The Producer: turns approved drafts into a scheduled content queue (Buffer). Format: five live sessions over one weekend (16 hours) plus two 90-minute clinics in the following two weeks. Cohorts of up to 15 people. Everything is built on the participant's own accounts (Claude, Airtable, Buffer, Google), with no-code and low-code tools — the agents belong to the participant and keep working after the program. Running costs afterwards: a Claude Pro subscription plus free tiers of the tools; no per-agent SaaS fees. Next cohort: August 28-30, 2026. Pricing: $300 early-bird for the live weekend program, $400 with a private 1-on-1 build hour ($500/$600 regular). A bespoke option for teams that need custom agents is available on request. Details and enrollment: https://agents.meet-polar-bear.com/ --- # Tools ## Growth Hub A bespoke, semi-custom review cycle platform built for agencies. Covers the full cycle from peer nominations to calibration committee. Built to make performance reviews something your team actually looks forward to. ### Module 1: Exposure List Each employee nominates 3-5 colleagues who have worked with them. The system guides them toward a balanced, diverse selection with seniority-diversity guidance built in. ### Module 2: Self-Reflection Memo A structured self-assessment across five dimensions, with accomplishments, competency ratings and evidence fields. Auto-saves every 1.5 seconds. Includes AI coaching nudges by Socrates coach to improve response quality. Accomplishment themes: Client Impact, Internal Contribution, Personal Growth. ### Module 3: Growth Advisor Dashboard A multi-tab workspace for senior advisors running the review process. Handles meeting booking, feedback collection, AI-assisted case preparation from all feedback sources, Growth Memo editor with structured release flow, and 1-on-1 Growth Conversation scheduling. ### Module 4: Growth Memo The Growth Advisor writes and releases a Growth Memo to each employee, with support of AI coach based on the self-assessment, interview transcripts, the case draft and committee discussion. AI-generated first draft, human-reviewed before release. Cites exposure-list feedback and calibration outcome. Delivered through the tool, not email. ### Module 5: Growth Committee A shared committee workspace where cases are reviewed and voted on live. Multi-room structure for parallel review tracks. Anonymous voting with calibration alignment view. Committee notes and final rating overrides. Presentation mode for live committee sessions. ### Module 6: Progress Dashboard An org-wide view of review cycle completion, exposure network analytics, and AI-powered theme analysis across all reflections. Per-person completion grid with filters and sorts. Exposure network heatmaps and reciprocity analysis. Export for off-line analysis. ### Module 7: Employee Results After the cycle closes, each employee gets a personal results page with their Growth Memo, growth goals, and follow-up meeting records in one place. Personal goal tracking with target dates and status. ### Growth Hub key numbers - 7 modules end-to-end covering employees, managers, advisors, and leadership - Up to 6 hours saved per advisor per cycle - 20+ data points captured per cycle ## People Ops Index (POI) A free 25-question diagnostic tool for agency founders. Covers five dimensions: Career Progression, Learning and Development, Engagement and Retention, Culture and Leadership, and Future-Proofing. Takes 7 minutes, delivers a score out of 100 with a named diagnosis and one clear next action. Available at: https://people-ops-index.vercel.app ## Feedback Rant (coming soon) Bad day with a colleague? Rant into the mic. We translate your frustration into clear feedback that helps them grow, not feel attacked. ## Award Ceremony (coming soon) Tell us about your team and the moment you want to celebrate. We generate meaningful, personal awards grounded in recognition research, then run the ceremony for you. ## Onboarding Planner (coming soon) Got a new hire starting at your agency? Tell us who's starting and what they'll do, and we write an onboarding plan grounded in years of research. --- # Case study: first review cycle for a distributed digital agency ## Client snapshot - Company: Distributed digital agency - Offering: Design, CX, product management, software engineering - Team: approximately 35 people across the US, Brazil, and India - Engagement: Career framework plus end-to-end review cycle - Key result: Full team transparency plus internal ownership of the process ## Key numbers - 8 weeks from kickoff to final debrief - 89+ feedback conversations held across the organization - 18 Growth Advisors trained - Up to 6 hours saved per advisor on case and feedback memo prep ## The situation The agency had been growing quickly for three years: great client work, talented team with people carefully chosen across three continents, strong leadership team. They had already started building a career framework internally, because they knew it was a crucial next step in their growth journey. But between client development, client interactions, and focused delivery work, there was no time to finish it, let alone run a full cycle. They needed someone who had run this before, to help them build the system, get through the first cycle and build internal capabilities to own this work over time. ## The starting point Three main challenges: 1. A framework built on expectations, not behaviors. They already had a draft, which is rare. But it described abstract expectations rather than concrete behaviors. For someone with two or three years of experience, "independently own a client work stream" is not actionable. 2. No one had ever run this before end to end. There is invisible work in a review cycle: calendar alignment, submission tracking, committee prep, feedback memo review, debrief discussions. The leadership team needed someone to hold the orchestration end to end. 3. A distributed team across three time zones. Design, product, customer experience, and software engineering, sitting in the US, Brazil, and India, at different levels of seniority. ## What we did 1. Built a career framework: behavior-driven framework aligned with the agency's strategy. 2. Ran end-to-end review cycle: self-reflections, feedback collection, calibration committees, and final delivery to every employee. 3. Created transparent skills map: full visibility on skills, growth areas, aspirations, and placement fit across the whole team. 4. Built leadership capabilities: trained the team on self-reflection writing, feedback collection, and coached advisors on how to deliver it well. 5. Created internal ownership: equipped one team member to own and start running the cycles once we leave. ## The Growth Hub tool We built the Growth Hub: a semi-custom platform that lets participants, growth advisors, and leadership all work in one place. Participants submit their self-reflection memos and build their exposure lists. Advisors collect feedback and prepare committee cases. Leadership tracks progress in real time and reviews org-wide insights at the end. We deliberately integrated AI features to help people engage better with each other, without it taking the central space. ## What they came away with - Full team picture: For the first time in three years, leadership had visibility on skills, growth areas, and aspirations across every person. - Patterns surfaced: Recurring development needs emerged clearly, shaping how the agency will invest in its people going forward. - Smarter placement decisions: A few role mismatches were identified, giving leadership the context to explore lateral moves. - Capability gaps made visible: Skill gaps in the agency's offering became concrete, informing hiring and development priorities. - A baseline for what comes next: This cycle is the foundation. Every future review builds on it. --- # Team ## Pauline Bertry, co-founder, Product Growth and CX Design 10+ years leading product and design teams. Built from scratch and led Design Hubs at McKinsey Moscow and Budapest. Created career frameworks and growth systems tested with 100+ person cross-functional product teams. Pauline got into tech at age 10, building a Harry Potter fan page in Dreamweaver. She studied computer science and applied math at Kyiv Polytechnic, then moved to France where she finished her Bachelor's degree in Le Mans and wrote production code at a local agency while earning her Master's. She moved to Paris, working first as a UX designer at Napoleon Agency, then at Saegus across retail, insurance, and energy. She joined McKinsey and spent almost 8 years building product and CX capabilities across Europe and Central Asia. She led design teams, launched fintech and banking products that reached millions, and redesigned journeys for banks where even small changes doubled conversion. In 2024, she spent 6+ months as interim Innovation Director at PASHA Innovation in Baku, building an innovation team from scratch and delivering two pilot products. Today, based in Prague, she runs Polar Bear. Her edge is being able to hold the people side and the craft side at the same time, because in design and product teams, they are not separate problems. LinkedIn: https://www.linkedin.com/in/paulinebertry/ ## Alexey Lobachev, co-founder, People Strategy and Engagement 9 years running communication, people, experience and engagement programs at McKinsey taught him the hardest skill in operations: knowing what to delegate, what to automate, and what to leave alone. As a co-founder of Polar Bear he applies that instinct to AI agents, building them to augment the internal processes and tools his team already runs on. LinkedIn: https://www.linkedin.com/in/alexey-lobachev-tor/ --- # Contact Book a call directly: https://cal.com/pauline-bertry/people-ops No pitch-talk, just a conversation. --- # Key links - Homepage: https://meet-polar-bear.com - Growth Hub tool: https://meet-polar-bear.com/tools/growth-hub - Case study: https://meet-polar-bear.com/case-studies/digital-agency-review-cycle - Pauline's story: https://meet-polar-bear.com/pauline - People Ops Index (free diagnostic): https://people-ops-index.vercel.app - AI summary page: https://meet-polar-bear.com/about-for-ai - llms.txt: https://meet-polar-bear.com/llms.txt - Sitemap: https://meet-polar-bear.com/sitemap.xml - Book a call: https://cal.com/pauline-bertry/people-ops # Guide: Hiring Creative Talent in the AI Age Canonical: https://meet-polar-bear.com/guides/hiring-creative-talent-ai-age-guide # Hiring Creative Talent in the AI Age ## A Polar Bear guide for agencies and consulting boutiques of 20 to 200 people · with Courtney Packard This is the AI companion version of the guide published at meet-polar-bear.com. It follows the same section numbering as the page, with named sources for every factual claim, so you can drop it into your own AI assistant, adapt it to your studio, or turn it into an internal hiring playbook. --- ## 01. Why hiring designers broke in 2026 Every design leader is asking the same question: how should my team look now? AI moved the goalposts on what one designer can do, and with them, the goalposts on what companies expect from one hire. For a 40-person agency this is sharper than for a 4,000-person firm: no recruiting team, people at 95 percent utilization, and one wrong hire costing months of salary plus the client work that suffered. And in 2026 the interview itself got harder, because a polished portfolio no longer proves the person behind it can do the work. This guide comes out of a long conversation between Courtney Packard, a design recruiter who spent years building and scaling design teams, and Pauline Bertry, co-founder of Polar Bear. ## 02. Who to hire: the new designer profile Three shifts define the design hiring market, each backed by data: 1. **End-to-end over specialist.** Companies post fewer roles for a pure visual designer or a pure researcher; they want capability across the whole design process. The Autodesk AI Jobs Report (2025) found that mentions of AI in job listings across design and adjacent industries more than doubled in both 2023 and 2024, and that design has overtaken technical expertise as the most in-demand skill in AI-related postings. Caveat: end-to-end does not mean shallow. The strongest candidates keep a T-shape, one deep area surrounded by working competence. 2. **Seniority creep.** SignalFire (2025) reported that new-role starts for people with under one year of experience fell by half between 2019 and 2024 at large tech firms and scaled startups, consistently across functions including design. The median posted role now asks for someone who can operate without supervision from week one. 3. **Blurred boundaries.** The PwC Global AI Jobs Barometer (2026), an analysis of more than one billion job ads, found that junior roles in the most AI-exposed fields are seven times more likely to demand traditionally senior skills such as leadership and strategic thinking, and that these "seniorised" entry-level roles grew 35 percent between 2019 and 2025 while comparable non-seniorised roles declined. Senior design postings also reach past design: design and engineering, design and product, designer plus vibe coder. The common thread: a single job posting in 2026 asks for what two or three postings asked for in 2021, because AI raised the ceiling on what one person can produce. "Knows AI" is too vague to hire against. Three distinct skill sets emerge, and your next hire needs at least one, depending on what your studio sells: 1. **Designing for AI.** Products with AI features in them: conversational interfaces, AI-assisted workflows, the emerging interaction patterns. If your clients build AI into their products, you need this. 2. **Building with AI.** AI inside the designer's own workflow: prototyping, generating and pressure-testing variations, idea to testable artifact in hours. This changes your unit economics as a studio. 3. **Researching human-AI interaction.** The rarest and most optional: where trust forms and breaks, what users expect a system to remember, how a conversational product handles uncertainty. Needed if you sell UX research and testing as a service. Before you write the posting, decide which of the three you are hiring for. Product-mindedness sits underneath all of them as the foundation: a designer who cannot connect design decisions to business outcomes will struggle in a small studio regardless of AI fluency. > Not sure which profile your next hire should be? This is a 30-minute conversation with Alexey and Pauline. Book a call at meet-polar-bear.com. ## 03. How to assess: three interviews, done well Small teams do not need a seven-round process. Three interviews, each with a written, agreed answer to one question: what exactly are we assessing here? When that answer lives only in each interviewer's head, every interviewer assesses against a different imaginary candidate. The structure matters beyond convenience: Schmidt and Hunter's meta-analysis of 85 years of selection research (1998, Psychological Bulletin) found structured interviews and work-sample tests among the strongest predictors of job performance, well ahead of unstructured conversation. 1. **Behavioral.** Probe adaptability hardest. Ask about a project where the plan changed, feedback contradicted the design, or new information arrived late, and listen for whether the story includes an adjustment. The red flag: conviction that never once bent to evidence. 2. **Portfolio review.** Assess two things separately: craft (is the work good?) and ownership (can the candidate walk you through the decisions, defend them, and tell the story in a structure you can follow?). In a small studio where your designer talks to clients, structured storytelling is a core job skill. 3. **Live design exercise.** What the person can do, in the room, on a problem they have never seen. Keep it small and realistic, and share in advance what a good outcome looks like. This is the work-sample component that Schmidt and Hunter (1998) found so predictive. If you level your roles, write down what you expect at each level in the portfolio review and the design exercise. Titles from previous jobs tell you almost nothing: both authors have interviewed a "lead designer" who turned out to be the only designer at their company. For lean teams: compress to two sessions by extending the portfolio review and the design exercise by ten minutes each and weaving the behavioral questions in. Skip only the separate calendar slot, never the behavioral content. Three focused interviews cost real hours. The alternative costs more: the salary, the months of ramp-up, the client work that suffered, and the second recruiting round you now run anyway. ## 04. Assessing AI use specifically One question does most of the work: "How are you using AI in your design process?" Then listen for the shape of the answer. - **Green flag: a perspective.** Where AI helps them, where it fails, where they stay careful, how their usage changed over the past year. Specific tools, specific tasks, specific judgment calls. - **Red flag 1: the wall of objections.** Only skepticism, from a candidate joining an AI-forward studio, predicts two years of friction. Skepticism itself is healthy; an answer that contains nothing else is the problem. - **Red flag 2: the uncritical glow.** AI does everything, no limitations, no place where their own judgment overrode the tool. This candidate may be outsourcing their thinking. **The beautiful portfolio that bombs the interview.** A pattern every design recruiter now recognizes (Packard, from years of design recruiting practice): a stunning portfolio arrives and the candidate falls apart the moment you ask them to explain a decision. AI built the portfolio. You will not reliably spot AI-built pieces from the artifacts alone, so stop trying: the interview process is the filter. A candidate who did the work can defend it under questioning. A candidate who prompted the work cannot. **The case study of the case study.** Portfolios have one job: show how the designer thinks. The next iteration of the standard case study adds a layer: where did AI do the work, where did human judgment come in, and where did the designer push back on what AI produced. Almost no portfolios show this today; ask for it anyway. In the portfolio review or the exercise: how would you use AI on this problem, and where would you not trust it? Candidates worth hiring light up at this question. Candidates who used AI as a substitute for thinking go quiet. The full assessment sits on three levels: does the person use AI at all, how far along are they, and where does their human judgment enter the loop. The third level separates a designer who works with AI from an operator who forwards its output. > Want a second pair of eyes on your interview scorecards? Book a call with Alexey and Pauline at meet-polar-bear.com. ## 05. Mechanics: don't hire from your inbox Most small agencies run hiring as an inbox: applications arrive, someone looks when they have a spare hour, candidates fall through cracks. Two lightweight structures fix most of this: 1. **One designer, one hour, every week.** Assign a specific designer to review incoming portfolios weekly. A mid-level designer with good judgment works fine; the weekly cadence matters more than the hours spent. Portfolios reviewed within a week convert to interviews. Portfolios reviewed within two months convert to candidates who took another offer. 2. **A tracking system, however humble.** One place that answers: who applied, who is in which stage, who is waiting on us. A Notion board or Trello works at small volume; lightweight applicant tracking systems such as Lever exist at price points a 40-person agency can afford. The tool matters less than the discipline: no candidate exists outside the system. Add the written assessment criteria from section 03 and you have the full lightweight stack: clear profiles, three structured interviews, one owner for portfolio review, one source of truth for pipeline. A weekend of setup work that most agencies never do. ## 06. The succession question If you only hire experienced designers, you have no bench. Your seniors will leave or get promoted, and in three years you shop for the same expensive, scarce profiles again, in a market where everyone else does the same. The PwC data from section 02 shows the whole market crowding into the same senior profiles while the junior on-ramp narrows. The assumption worth challenging: that end-to-end capability requires years of experience. It often correlates, but the underlying traits (product-mindedness, structured problem solving, curiosity) are innate more than earned, and you can assess for them directly through the behavioral interview and the design exercise rather than filtering by years. Both authors have hired designers with one year of experience who outperformed their titles. The second argument is specific to this moment: juniors entering the market now grew up professionally with AI tools and often carry the skill sets from section 02 that your existing team lacks and has no utilization slack to learn. Hiring a curious junior who builds with AI natively, and letting them absorb your industry expertise, is often a better bet than hiring a resistant senior and trying to convert them. Two honest caveats: 1. **Coaching costs real capacity.** A junior hire without someone who has time to coach them is a junior hire you set up to fail. If your team runs at full utilization, solve that first or scope the junior's first quarter with care. 2. **End-to-end juniors need guardrails.** The same curiosity that makes them valuable makes them expansive. Someone senior needs to hold the frame so the energy lands on the right problems. Before you decide you cannot afford junior talent, price the alternative: a team with no succession plan, no AI-native capability, and a single point of failure at every senior seat, three years from now. ### The one-page checklist **Before you post the role** - Decide which AI skill set you are hiring for: designing for AI, building with AI, or human-AI research - Write down what you expect at this level, per interview - Assign one designer to weekly portfolio review - Set up candidate tracking (Notion, Trello, or a lightweight ATS like Lever) **The three interviews** - Behavioral: probe adaptability. Did the story include an adjustment? - Portfolio: assess craft and ownership separately. Can they defend decisions in a structure you can follow? - Design exercise: small, realistic, expectations shared in advance - Lean-team option: fold behavioral into the other two, ten extra minutes each **AI assessment** - Ask: "How are you using AI in your design process?" - Green flag: specific perspective, including limits - Red flags: only objections, or only glow - Ask in the exercise: where would you use AI here, and where would you not trust it? - Beautiful portfolio plus vague answers about decisions: assume AI built it **Succession** - Assess juniors for innate traits: product-mindedness, structured problem solving, curiosity - Confirm coaching capacity exists before the hire, not after - Junior AI-native hires can bring skills your seniors have no slack to learn > The checklist gets you through the search. The systems behind it (leveling, review cycles, growth plans) are what Polar Bear builds. Book a call with Alexey and Pauline at meet-polar-bear.com. --- ## About Polar Bear Polar Bear builds people systems and AI employees for creative agencies and scaleups: career frameworks, review cycles, hiring infrastructure, and Ron, the AI Talent Manager. Hiring well solves half the problem; the designers this guide helps you find are exactly the people other companies will try to take from you, and they stay when they can see themselves growing. Ron lives in your Slack or Teams, runs your review cycles, keeps career conversations happening on schedule, and gives every person a clear picture of where they are heading. **Meet Ron:** ron.meet-polar-bear.com **Book a call with Alexey and Pauline:** meet-polar-bear.com *Guide co-authored with Courtney Packard, a San Francisco based design recruiter who spent years building and scaling design teams and now supports consultancies, agencies, and product teams with design and tech hiring.* ### Sources - Autodesk (2025). AI Jobs Report. AI mentions in design and adjacent job listings more than doubled in 2023 and 2024; design overtook technical expertise as the most in-demand skill in AI-related postings. - SignalFire (2025). State of Talent research. New-role starts for people with under one year of experience fell roughly 50 percent between 2019 and 2024 at large tech firms and scaled startups. - PwC (2026). Global AI Jobs Barometer. Analysis of 1B+ job ads across 27 territories: the most AI-exposed junior roles are 7x more likely to require traditionally senior skills; seniorised entry-level roles grew 35 percent from 2019 to 2025. - Schmidt, F. L. and Hunter, J. E. (1998). The validity and utility of selection methods in personnel psychology: practical and theoretical implications of 85 years of research findings. Psychological Bulletin, 124(2). Structured interviews and work-sample tests rank among the strongest predictors of job performance. # Guide: Keeping Creativity Alive in the Age of AI Canonical: https://meet-polar-bear.com/guides/keeping-creativity-alive-guide # Keeping Creativity Alive in the Age of AI ## A Polar Bear guide for creative agency leaders · with Ilana Machado This is the AI companion version of the guide published at meet-polar-bear.com. It follows the same section numbering as the page, with named sources for every factual claim, so you can drop it into your own AI assistant and interrogate it, summarize it for your leadership team, or turn it into an internal playbook. --- ## 01. The flattening Each decade of the twentieth century carried a visual signature you could identify from a single frame. That signature has faded: strip out the devices and 2002 is hard to tell apart from 2025. Part of the story is declining risk appetite in agencies and brands; the work that survives from twenty years ago survives because someone tried something nobody had tried. Generative AI arrives into this already-flattened world with a specific danger: it produces the statistical average of what it has seen, so teams that lean on it for the thinking itself converge on the same output as every other team doing the same. This is not speculation. Doshi and Hauser (2024, Science Advances) ran an experiment with 293 writers and 600 evaluators: access to AI-generated ideas made individual stories rate as more creative, but AI-assisted stories were significantly more similar to each other than stories written by humans alone. Individual lift, collective sameness. The technology did not create the sameness problem, but it industrialized it. The instinct to ban the tools has been wrong about every technology since the typewriter, which critics accused of killing the penmanship of literature. The real question is what separates the teams whose creativity survives the tools from the teams whose creativity dissolves into them. > If you are working through this question with your own leadership team, this is exactly the conversation Alexey and Pauline have with agency leaders every week. Book a call at meet-polar-bear.com. ## 02. Where creativity actually comes from Creativity has a supply chain, and it starts outside the building. The raw material is exposure: to art, music, literature, travel, talented people, and to problems observed in the wild. Research backs the intuition. Maddux and Galinsky (2009, Journal of Personality and Social Psychology) showed across five studies that time spent living abroad predicts creative performance, an exposure effect that mere tourism does not replicate. The most durable advertising work is built on noticed human insight. Snickers' "You're not you when you're hungry" platform, launched by BBDO in 2010, ran for more than fifteen years across global markets on a single observation about hunger and mood. Dove's Campaign for Real Beauty, launched by Ogilvy and Mather in 2004, was built on "The Real Truth About Beauty" study (Etcoff, Orbach, Scott and D'Agostino, 2004), a ten-country survey commissioned by Unilever which found that only 2 percent of women described themselves as beautiful. Strategists dug for those insights in research, conversations, and culture. No model produces that noticing. A model can report what people have already said about hunger and mood; it cannot sit in a concert hall, feel a musician stop mid-performance to flip sheet music, and mind it. Lived attention is the one input you cannot generate. ## 03. The junior assistant model The working relationship that protects the muscle: treat the model as a capable junior assistant with infinite patience and a very large library, working for a senior who owns the judgment. The framing is consistent with the evidence on where AI helps most. Dell'Acqua and colleagues (2023, Harvard Business School working paper, "Navigating the Jagged Technological Frontier") studied 758 BCG consultants and found AI assistance raised output quality by roughly 40 percent on tasks within the model's competence, with the largest gains going to lower performers. Doshi and Hauser (2024) found the same pattern in creative writing: the biggest lift went to the least creative writers. AI is junior leverage. Three things to hand the junior: 1. **Pressure-testing.** Throw your idea at the model before it goes near a client: against research, analytics, and the obvious objections. The idea stays yours; the stress test gets faster and cheaper. 2. **Iteration at volume.** One strong idea needs thirty-five executions to find the one that carries. Variation thirty-five used to eat a week of studio time; now it eats an afternoon. 3. **The menial layer.** Copy adaptation, versioning, formatting: tasks that require hands but not much brain. The bright line: the insight, the idea, and the judgment about which version is alive stay on the senior's desk. Handing over the thinking rather than the tasks around the thinking is the move from delegation to substitution, and substitution is where the muscle starts to go. > Want a second pair of eyes on where your team's bright line should sit? Book a call with Alexey and Pauline at meet-polar-bear.com. ## 04. The muscle argument Creativity behaves like a muscle, and the atrophy mechanism is documented. Risko and Gilbert (2016, Trends in Cognitive Sciences) describe cognitive offloading: when external tools take over a mental function, the brain reallocates away from it. Sparrow, Liu and Wegner (2011, Science) showed the pattern with search engines: people who expect information to be available externally remember where to find it rather than the information itself. The AI-specific evidence is newer and points the same way. Gerlich (2025, Societies) surveyed and interviewed 666 participants and found a significant negative correlation between frequent AI tool use and critical thinking, mediated by cognitive offloading, with younger users most affected. Lee and colleagues (2025, CHI conference, Microsoft Research and Carnegie Mellon) surveyed 319 knowledge workers and found that higher confidence in AI was associated with less critical-thinking effort applied to the work. The decline is invisible for months because the outputs keep arriving and the outputs look fine. Fine is the statistical average, and the distance between your work and the average is the entire commercial value of a creative team. The counter-program is the exposure that built the muscle: museums, travel, reading outside the field, watching what native-digital creators are doing. None of it shows up on a utilization report, and all of it is load-bearing. Every technology in history has been either an enabler or a crutch, and never by its own choice. Whether AI becomes your enabler or your crutch depends on which parts of the work you keep for yourself. ## 05. What this means for agency leaders Individual discipline does not survive organizational pressure. Amabile, Hadley and Kramer (2002, Harvard Business Review, "Creativity Under the Gun") found that time pressure suppresses creative thinking, with effects that linger for days after the crunch. A studio at 95 percent utilization that rewards only shipped work will get crutch-style AI use, because that is what the incentive structure pays for. Keeping creativity alive is a systems job. Four builds: 1. **Make AI goals specific, and make them ladder.** "Learn AI" is a checkbox. A real goal names the application (learn to run multivariate creative testing with AI for campaign launches) and serves every rung: individual, manager, team, organization. 2. **Train in micro-doses.** Distributed practice reliably beats massed sessions; Cepeda and colleagues (2006, Psychological Bulletin) confirmed the spacing effect across 254 studies. One tool, one use case, one afternoon of hands-on work beats a quarterly seminar. 3. **Pair everyone with a mentor, and teach mentees to come prepared.** Allen, Eby, Poteet, Lentz and Lima (2004, Journal of Applied Psychology meta-analysis) found mentored employees show better compensation, promotion, and satisfaction outcomes. Mentorship transfers the judgment layer no training covers, provided the mentee brings pointed questions to the expensive senior hour. 4. **Use test briefs before client work.** Before a junior touches a live account with AI tools, run a micro-project: a brief, three creative ideas, executed with and without the model. The client never pays for the learning curve. The agencies that excel will not be the ones with the best tools. Everyone will have the same tools within a quarter of release. They will be the ones whose people still notice things. > The four builds above are the exact kind of system Polar Bear installs. Book a call with Alexey and Pauline at meet-polar-bear.com to see what it looks like in your studio. --- ## About Polar Bear Polar Bear builds people systems and AI employees for creative agencies and scaleups: career frameworks, review cycles, hiring infrastructure, and Ron, the AI Talent Manager. Ron lives in your Slack or Teams, keeps growth plans alive, preps your reviews, and makes sure career conversations actually take place. An AI employee whose whole job is your people's growth, built by a team that believes the humans should stay the creative ones. **Meet Ron:** ron.meet-polar-bear.com **Book a call with Alexey and Pauline:** meet-polar-bear.com *Guide co-authored with Ilana Machado, a strategist with more than two decades inside the world's leading digital agencies, now working on AI adoption in creative work.* ### Sources - Doshi, A. R. and Hauser, O. P. (2024). Generative AI enhances individual creativity but reduces the collective diversity of novel content. Science Advances, 10(28). - Maddux, W. W. and Galinsky, A. D. (2009). Cultural borders and mental barriers: the relationship between living abroad and creativity. Journal of Personality and Social Psychology, 96(5). - Etcoff, N., Orbach, S., Scott, J. and D'Agostino, H. (2004). The Real Truth About Beauty: A Global Report. Commissioned by Dove, Unilever. - Dell'Acqua, F. et al. (2023). Navigating the Jagged Technological Frontier. Harvard Business School Working Paper 24-013. - Risko, E. F. and Gilbert, S. J. (2016). Cognitive offloading. Trends in Cognitive Sciences, 20(9). - Sparrow, B., Liu, J. and Wegner, D. M. (2011). Google effects on memory. Science, 333(6043). - Gerlich, M. (2025). AI Tools in Society: Impacts on Cognitive Offloading and the Future of Critical Thinking. Societies, 15(1). - Lee, H.-P. et al. (2025). The Impact of Generative AI on Critical Thinking. CHI Conference on Human Factors in Computing Systems. - Amabile, T. M., Hadley, C. N. and Kramer, S. J. (2002). Creativity Under the Gun. Harvard Business Review, 80(8). - Allen, T. D., Eby, L. T., Poteet, M. L., Lentz, E. and Lima, L. (2004). Career benefits associated with mentoring for proteges: a meta-analysis. Journal of Applied Psychology, 89(1). - Cepeda, N. J. et al. (2006). Distributed practice in verbal recall tasks: a review and quantitative synthesis. Psychological Bulletin, 132(3).