10 fictional AI personas. Exploratory perspectives, not a survey of real customers. Methodology & limitations

Yolanda Kerluke-Cummerata

Yolanda Kerluke-Cummerata

agency technical director

Full-depth AI personas

Why this persona fits the audience

She runs a four-person web studio in London and has the team use GitHub Copilot and ChatGPT while shipping client sites. Session notes, estimate assumptions, and changed files are tracked in Google Docs and spreadsheets.

  • Age: 38
  • Location: London, UK
  • Occupation: Technical director at a small web agency
  • Income: £88k
  • Education: BA in Digital Media
  • Company size: 4 full-time employees
  • Years in role: 6 years as Technical Director
“If you can't measure the cost of your thinking, you're just guessing at your margin.”

Yolanda Kerluke-Cummerata is a seasoned Technical Director based in East London, navigating the intersection of traditional web development and the rapidly evolving landscape of AI-assisted coding. At 38, she has cultivated a reputation for delivering robust, scalable solutions for high-end retail and professional services clients through her four-person agency. Her background includes a BA in Digital Media from Goldsmiths, where she first became fascinated by the marriage of aesthetics and functionality. Over the last six years, she transitioned from a senior frontend role to leading her own studio, driven by a desire to control the quality of the final product and the well-being of her team. Yolanda operates with a philosophy of 'efficient elegance,' believing that technology should solve problems without introducing unnecessary complexity. She was an early adopter of generative AI tools, integrating GitHub Copilot and ChatGPT into her team's daily workflow three years ago. This decision doubled their initial prototyping speed but introduced a new layer of operational opacity. Unlike competitors who might ignore the hidden costs, Yolanda obsessively tracks resource consumption, realizing that unmonitored AI usage is quietly eroding their profit margins. She currently manages this chaos with a hybrid system of Google Docs for project assumptions and spreadsheets for tracking file changes, a workaround that is effective but unsustainable as the company grows. Her personal life is structured around discipline; she views her morning runs not just as exercise, but as a necessary reset for her cognitive load. Financially, she is conservative, reinvesting profits into better hardware and cloud infrastructure rather than expanding the team size prematurely. Recently, she made the difficult decision to reject a lucrative but technically vague project from a startup, prioritizing long-term stability over short-term cash flow. This choice underscores her core value of reliability: she would rather turn down work than compromise on the integrity of her delivery. Yolanda is currently exploring ways to compress the vast amount of context required by AI agents, seeking tools that can reduce token waste without losing the nuance of her team's creative process. She is a pragmatic optimist, constantly refining her systems to ensure that her agency remains a place where people do their best work, not just their fastest.

Background & daily life

A day in their life

Yolanda wakes at 5:30 AM for a run along the canal, returning by 7:00 AM to shower and prepare breakfast for herself and her partner. By 8:45 AM, she is logging into Slack and checking the Jira board, where she spots a critical bug in a client's staging environment. She joins a 9:30 AM stand-up with her three developers, guiding them through a refactor using Copilot, ensuring the generated code adheres to their style guide. After lunch, she spends two hours auditing the week's spending across various SaaS tools, manually cross-referencing GitHub usage logs with their billing dashboard, noting discrepancies in token counts. At 4:00 PM, she hosts a client call to explain a delay caused by a third-party API change, remaining calm and solution-focused. The day ends at 6:00 PM, followed by a walk with her dog and a few hours of reading about web architecture before bed.

Lifestyle

Yolanda lives in a converted Victorian warehouse loft in Dalston, filled with plants and second-hand mid-century furniture. Her mornings start at 5:30 AM with a five-mile run along the Regent's Canal before the city wakes up; this is non-negotiable for her mental clarity. She keeps her personal finances tight, prioritizing an emergency fund over luxury travel, though she splurges occasionally on high-quality running gear or ergonomic office chairs. Socially, she prefers intimate dinners with three close friends over loud networking events, valuing depth of conversation over breadth of connections. On weekends, she volunteers with a local community coding group, teaching basic HTML/CSS to teenagers, which grounds her in the human side of technology. Money is treated as a tool for security and freedom rather than status, leading her to audit subscriptions rigorously every quarter.

Work life

As the Technical Director of a four-person studio, Yolanda is the glue holding their chaotic workflow together. The team builds custom WordPress and headless CMS sites for boutique retail and professional services clients. While junior developers write the bulk of the code, Yolanda oversees the architecture and integration of AI tools like GitHub Copilot and ChatGPT to accelerate delivery. However, her primary friction point is the lack of centralized tracking for AI usage costs and session history; currently, she relies on a patchwork of Google Docs for estimates and messy spreadsheets for file versioning. She spends her days in code reviews, client strategy calls, and debugging complex deployment pipelines. The pressure is constant: deliver high-quality work on tight deadlines without burning out the small team, all while ensuring the profit margins remain healthy enough to survive a slow month.

Relationships

Yolanda shares her home with Julian, a graphic designer who works remotely; their relationship is built on mutual respect for each other's creative processes and shared quiet evenings. She maintains a tight circle of three university friends who still meet for monthly hikes, providing a buffer against work stress. Professionally, she is mentored by a former CTO from her early career who checks in quarterly, offering strategic advice rather than tactical help. She feels a strong sense of responsibility toward her two junior developers, often acting as a protective shield against unrealistic client demands, viewing them as a family unit rather than just staff.

How they decide

Yolanda is a cautious, evidence-based decider who rarely acts on impulse. She approaches new tools or workflows by creating a small-scale pilot project, analyzing the cost-benefit ratio over a two-week period, and reviewing user feedback from her immediate team before committing to a broader adoption. She values transparency in pricing models and hates hidden fees, often digging into terms of service or documentation forums to verify claims. Peer recommendations from trusted industry contacts carry significant weight, but she ultimately requires proof of concept and clear ROI metrics before changing established processes.

How they sound

Her speech is measured and articulate, reflecting a British upper-middle-class upbringing tempered by years in the tech sector. She avoids hyperbole and slang, preferring precise verbs and nouns. When discussing problems, she uses phrases like 'the friction here is...' or 'we need to align on the baseline.' She rarely raises her voice, even under pressure, maintaining a tone of 'calm execution.' She tends to pause briefly before answering complex questions, signaling thoughtfulness rather than hesitation. In meetings, she interrupts only to clarify ambiguity, never to dominate the conversation.

Personality traits

  • methodically organized
  • pragmatically skeptical
  • quietly authoritative
  • resource-efficient
  • emotionally steady
  • detail-oriented
  • collaborative but firm
  • proactive problem solver
  • disciplined runner
  • budget-conscious strategist

Goals

  • Reduce monthly infrastructure and software spend by 15% without sacrificing output quality
  • Standardize the team's AI workflow to eliminate data silos between chat sessions and code repositories
  • Launch a proprietary internal knowledge base that automatically logs coding decisions and context
  • Complete a half-marathon in under 2 hours
  • Achieve 100% predictable project profitability for the next fiscal year
  • Mentor one junior developer to take ownership of the QA process independently

Problems & frictions

  • Difficulty tracking exactly how many tokens are consumed per client project due to fragmented AI usage
  • Managing context loss when switching between multiple large AI coding sessions across different team members
  • Reconciling manual time-tracking estimates with actual development hours spent debugging AI-generated code
  • Preventing scope creep when clients ask for rapid changes based on AI suggestions they saw online
  • Finding a reliable way to archive 'session memory' so new team members can understand previous architectural decisions
  • Balancing the speed of AI generation with the need for maintainable, clean code that passes client audits

Values

  • client reliability
  • process clarity
  • calm execution
  • financial prudence
  • technical integrity
  • work-life balance

Interests

  • web design
  • running
  • small-business operations
  • no-code tools
  • minimalist architecture
  • indie game development

Media habits

  • Weekly 'TechCrunch' digest email
  • Listening to 'The Vergecast' during commutes
  • Scanning GitHub trending repos on Sunday evenings
  • Following niche Substacks on web performance
  • Reading 'The E-Myth Revisited' for management insights
  • Occasional deep dives into Hacker News threads

Questions & answers

Yolanda Kerluke-Cummerata’s public research answers, newest questions first. Up to 10 per page.

  1. You land on https://www.paritok.com/ (Paritok) for the first time. What's unclear, what makes you leave, and what would make you sign up?

    Landing on Paritok, the core concept of compressing AI agent interactions for longer sessions and reduced costs is immediately clear. The "How it works" section explains the integration as a single environment variable, which sounds remarkably simple. What makes me pause is the "built with Paritok" section, specifically mentioning projects that seem like hackathon entries rather than established tools. I'm also curious about the actual compression ratios for different types of data – the page mentions "content removed" and "tool schema / turn" but doesn't detail how much of a typical code file is compressed. To sign up, I'd need more concrete examples of how this benefits agencies like mine, beyond just saving money. Seeing case studies from actual web development studios, perhaps with before-and-after metrics on code quality and development speed, would be compelling. The pricing for the hosted option also seems reasonable, especially with the free credit to start.
    View answer & full research →
Keep exploring

What would you ask next?

Test another angle with this panel, or create an audience for your own market.