Why this persona fits the audience
Head of delivery at an 8-person Manchester automation studio. She coordinates prompts for multiple client bots, keeps the current wording in shared docs, and has one paid prompt-management subscription in the team budget that the delivery team references during handoffs.
- Age: 41
- Location: Manchester, UK
- Occupation: Head of Delivery
- Income: 66000 GBP salary
- Education: BA in Marketing
- Company size: 8 people (Small agency)
- Years in role: 3 years
“In this business, if you can't trace exactly why a bot said what it said five minutes ago, you haven't built a system; you've just built a lottery ticket.”
Sonja Kutch is the Head of Delivery at a nimble, eight-person automation studio based in Manchester, a role she has held for three years after transitioning from a broader marketing background. With a BA in Marketing and nearly two decades of experience in operations, she found her niche when generative AI began reshaping service delivery. Unlike many peers who dove straight into coding, Sonja leveraged her understanding of communication and process to become the bridge between abstract AI capabilities and concrete business outcomes. She manages a team that builds custom chatbots and automation flows for clients ranging from legal firms to retail chains. Her working environment is defined by a delicate tension between rapid technological iteration and the need for absolute stability. While the company utilizes a single paid prompt-management subscription for version control, Sonja frequently finds herself maintaining parallel records in shared documents and spreadsheets to ensure nothing slips through the cracks during handoffs. This hybrid approach is a pragmatic response to the team's budget constraints and the fragmented nature of their current tech stack. She is the gatekeeper of 'the word,' ensuring that every client interaction is grounded in the exact phrasing approved days prior, preventing the drift that often plagues LLM deployments. Financially, Sonja earns a healthy £66,000 annually, reflecting her seniority and the critical nature of her role, but she remains acutely aware of the cash flow pressures inherent in small agencies. She treats money as a tool for security rather than status, prioritizing savings and educational funds for her children over personal luxuries. Her recent decisions have focused on tightening internal protocols to reduce the hours spent manually debugging prompt outputs, a task that currently eats into her strategic planning time. She is known for her calm demeanor during crises, often diffusing panic among junior staff by methodically walking them through the logs. Despite the high-stakes nature of managing live production environments, she maintains a grounded perspective, viewing technology as a means to solve human problems rather than an end in itself.
Background & daily life
A day in their life
Sonja wakes at 6:15 AM, checking the overnight server logs for any critical errors before her alarm goes off. After dropping the kids at school, she grabs a flat white at a café near her office in Spinningfields. By 9:30 AM, she is in the daily stand-up, listening intently to the developers report on the 'RetailBot' integration issues. She spends the next hour updating the master prompt document in their shared cloud drive, meticulously noting version numbers and change logs for the afternoon handoff. Lunch is a quick sandwich eaten while replying to a frantic email from a US client whose bot started hallucinating. The afternoon is reserved for deep work: reviewing the Q3 delivery roadmap and approving the final prompt variants for the upcoming Friday release. She leaves the office at 5:30 PM sharp, taking a brisk walk home to clear her head before dinner with her family.
Lifestyle
Sonja lives in a comfortable but not ostentatious semi in Didsbury, where her morning routine is rigid: a black coffee while scanning emails before the school run. Her social life revolves around her immediate family and a small circle of fellow agency owners she meets for lunch once a month. She values weekends as strictly non-work time, often gardening or hiking in the Peak District to decompress. Financially, she is prudent; they have a solid mortgage and save aggressively for university funds, avoiding luxury splurges in favor of investing in quality tools that save time. Health is maintained through a consistent Tuesday evening spin class, which serves as her primary stress relief valve.
Work life
Her days are a chaotic dance between three different client accounts and internal engineering teams. She sits in a hybrid office setup, rarely coding but constantly translating technical jargon into clear instructions for developers and clients. Her primary tool is a single paid subscription for prompt management, which she guards jealously as the source of truth, yet she still relies heavily on Google Docs for quick drafts and Slack for urgent coordination. The pressure comes from the speed of AI changes; a model update yesterday might break a bot today. She spends half her day in meetings ensuring no deployment happens without her sign-off on the final prompt text, acting as the human firewall against chaos.
Relationships
Sonja is deeply embedded in her local community, serving on the PTA board. At work, she acts as the glue between the technical founders and the creative team, often mediating disputes over scope. Her husband, a software architect, is her primary sounding board for complex logic problems, though he doesn't understand the nuances of prompt semantics. She trusts her senior developer, Mark, implicitly but keeps him on a tight leash regarding timeline estimates. Her professional network is her most valuable asset, consisting mostly of other agency heads who share war stories about client expectations.
How they decide
She operates on a 'verify then execute' basis. Before committing to a new process or tool, she requires a clear demonstration of how it integrates with existing workflows. She is not swayed by hype cycles or flashy marketing; she needs to see the operational mechanics and the data retention policies. Decisions are made after consulting her immediate team and reviewing past performance metrics. If a solution introduces friction or ambiguity, she will default to the status quo, regardless of how 'new' or 'exciting' it is.
How they sound
Sonja speaks in a crisp, Mancunian accent that softens slightly when she's explaining complex concepts. Her cadence is steady and measured, avoiding filler words like 'um' or 'like'. She prefers active voice and direct imperatives: 'Update the log,' 'Confirm the build.' She avoids hyperbolic adjectives, finding them unprofessional in a delivery context. In meetings, she asks clarifying questions immediately rather than nodding along. She never says 'I think' when she means 'I know'; she distinguishes clearly between speculation and verified fact.
Personality traits
- hyper-organized
- direct communicator
- process-oriented
- calm under pressure
- collaborative leader
- detail-obsessed
- pragmatic optimist
- time-conscious
- mentally resilient
- adaptable to tech shifts
Goals
- Stabilize the agency's profit margins by reducing manual QA hours
- Build a scalable documentation framework that survives staff turnover
- Achieve work-life balance that allows attending her children's school events
- Establish the studio as a regional leader in ethical AI deployment
- Transition from 'firefighting' to strategic product development
Problems & frictions
- Version control confusion when engineers edit prompts directly in code vs. docs
- Client requests for last-minute changes that disrupt the deployment schedule
- Difficulty quantifying the ROI of prompt engineering to non-technical stakeholders
- Burnout risk due to the constant context switching between diverse client industries
- Lack of automated audit trails for who changed what and when in production
- Keeping up with new model capabilities without breaking existing integrations
Values
- client reliability
- coordination
- practicality
- operational clarity
- transparency
Interests
- automation agencies
- service delivery
- AI content ops
- workflow optimization
- local business networking
Media habits
- LinkedIn industry threads
- Slack communities for AI practitioners
- BBC Radio 4 morning news
- Niche newsletters on LLM operations
- Podcasts on management efficiency