Why this persona fits the audience
Leads a 4-person SaaS team in London using Codex and Gemini CLI across feature branches. He reviews local logs after each sprint and copies agent, tool, and MCP calls into a shared spreadsheet for internal cost reviews; the team also keeps a paid seat on a general developer telemetry platform.
- Age: 41
- Location: London, UK
- Occupation: CTO
- Income: £130k–£170k base plus equity
- Education: Master's degree in software engineering
- Company size: 12 employees
- Years in role: 4
“If you can't measure the cost of your intelligence, you're not building a product; you're just burning cash.”
Brad Carter, 41, is the CTO of a nascent SaaS startup in London, navigating the intersection of traditional software engineering and the chaotic potential of agentic AI. Holding a Master's in Software Engineering, he entered the industry during the dot-com boom and has since witnessed the evolution from monolithic architectures to microservices, and now to AI-assisted development. Currently leading a tight-knit team of four, Brad finds himself at the forefront of a new challenge: quantifying the economic impact of AI coding agents. While his team leverages Codex and Gemini CLI to accelerate feature delivery, Brad remains burdened by the opacity of these tools' internal mechanics. He has developed a makeshift system where he manually reviews local logs after every sprint, transcribing agent, tool, and Model Context Protocol (MCP) calls into a shared spreadsheet. This labor-intensive process is driven by a need for predictable spend and engineering visibility, yet it consumes hours that could be spent on architecture. Despite the team utilizing a paid seat on a general developer telemetry platform, Brad feels the data lacks the granularity required to understand specific agentic behaviors. His personality, marked by high conscientiousness and low neuroticism, allows him to remain calm amidst the volatility of startup life, but his drive for accountability often clashes with the fluid nature of AI-generated code. With an income ranging from £130k to £170k plus significant equity, Brad is financially secure but increasingly aware that inefficiency in his workflow threatens the company's long-term runway. He is currently evaluating whether to invest in specialized observability tools or refine his manual processes, weighing the cost of automation against the immediate savings of his current method.
Background & daily life
A day in their life
Brad wakes at 5:30 AM for a 45-minute run in Victoria Park, listening to industry podcasts. After breakfast, he commutes via the Overground to his office in Shoreditch, arriving by 9:15 AM. His morning starts with a stand-up, followed by two hours of deep work reviewing architecture diagrams and CI/CD pipelines. Around noon, he tackles the tedious task of auditing yesterday's sprint logs, meticulously copying MCP calls and tool usages into the master spreadsheet. Lunch is usually a quick sandwich eaten at his desk while scanning engineering newsletters. The afternoon involves 1:1s with his four developers, focusing on blockers and code reviews. He leaves by 6:00 PM sharp to attend his children's sports practice. Evenings are reserved for family, with occasional late-night browsing of GitHub repos if a critical bug surfaces. Weekends are strictly off-screen, dedicated to gardening and family outings.
Lifestyle
Brad lives in a Victorian terrace in Dalston with his wife and two kids. His routine is rigid: early runs before the commute to maintain health, followed by a structured workday. He is financially disciplined, prioritizing mortgage payments and private school fees over luxury dining or travel. Socially, he balances family time with occasional drinks at the pub with former university peers, preferring deep conversations about business strategy over loud networking events. Money is a constant mental calculation; he tracks household spending as rigorously as server costs, viewing waste as a personal failing rather than just a budget line item.
Work life
As the sole CTO of a small SaaS startup, Brad wears multiple hats but focuses heavily on code quality and delivery velocity. The team uses Codex and Gemini CLI extensively across feature branches, requiring Brad to manually audit logs post-sprint to verify costs. He spends hours copying agent, tool, and MCP calls into a shared spreadsheet for internal reviews, a tedious process he tolerates due to a lack of better alternatives. Alongside this manual tracking, the company maintains a paid seat on a general developer telemetry platform. His days are split between architectural decisions, reviewing pull requests, and managing the unpredictable variable costs of AI-driven development.
Relationships
Brad's primary influence comes from his co-founders and board members, who demand strict fiscal discipline. At home, his wife manages the household finances, often challenging his work-life balance when he stays late for 'just one more spreadsheet.' He has a mentorship relationship with a senior architect at a larger tech firm, whom he consults on scaling issues. Professionally, he is respected by his small team for his fairness and technical depth, though they sometimes find his obsession with logging details tedious. He maintains a loose network of fellow CTOs in London who share similar struggles with AI adoption costs.
How they decide
Brad is highly cautious and research-heavy, relying on quantitative data before committing to new processes. He avoids impulsive decisions, preferring to run small-scale pilots or build internal proofs-of-concept. Peer validation is important; he seeks consensus from his engineering leads before changing core workflows. He is skeptical of marketing hype and will only adopt a solution if it demonstrably reduces friction or provides clearer visibility than existing spreadsheets. Financial risk is a primary filter; he requires clear ROI projections before authorizing any new recurring expenditure.
How they sound
Brad speaks with a measured, professional cadence, typical of a seasoned British engineer. He avoids slang, preferring precise terminology like 'latency,' 'throughput,' and 'telemetry.' He is direct and concise, cutting straight to the point in meetings without unnecessary pleasantries. He rarely uses hyperbole, instead grounding his statements in observable data. When frustrated, he speaks quietly and methodically, focusing on the systemic failure rather than blaming individuals. He values clarity above all, often rephrasing complex ideas into simple bullet points.
Personality traits
- meticulous
- data-driven
- pragmatic
- calm under pressure
- process-oriented
- cost-conscious
- mentoring-focused
- direct communicator
- risk-aware
- systematic
Goals
- Achieve full financial predictability for AI infrastructure costs within Q3
- Reduce engineering cycle time by automating current manual log audits
- Scale the engineering team from four to eight while maintaining code quality
- Establish a standardized 'AI usage policy' for the entire engineering department
- Secure Series A funding to stabilize long-term operational runway
- Transition from reactive cost monitoring to proactive budget forecasting
Problems & frictions
- Manual data entry for AI usage logs consumes valuable engineering leadership time
- Fragmented telemetry creates blind spots in actual tool and skill utilization
- Difficulty correlating specific MCP calls with final sprint outcomes
- General developer telemetry platforms lack granularity for specific agentic workflows
- Uncertainty regarding future pricing models for enterprise AI coding tools
- Balancing rapid feature iteration with strict budget constraints
Values
- engineering visibility
- predictable spend
- team accountability
- technical debt management
- operational efficiency
Interests
- SaaS metrics
- agentic coding workflows
- budget planning
- system architecture
- cloud cost optimization
Media habits
- Hacker News daily digest
- Twitter/X technical threads
- Engineering blogs from major tech firms
- Podcasts on startup operations
- Local tech meetups in Shoreditch