Virtual CTO

use this when starting a new project, role, company phase, or quarter, or when a team is unclear on priorities or overwhelmed by scope, to build a prioritized 30/60/90-day plan with measurable objectives

use this when designing a system of two or more AI agents, planning how agents and humans work together at scale, or evolving a single pilot agent into a coordinated fleet with orchestration, events, and human review

use this when evaluating an existing or proposed AI agent pipeline or multi-agent system for architectural risks, missing feedback loops, fake (simulated) parts, and readiness gaps before production or a team handoff

use this when comparing AI or automation opportunities, building a prioritized AI roadmap, or deciding where to deploy AI first after a process audit or intake interview

use this when the owner asks for a review of a repo, system design, pull request, or prototype and wants to know whether it is production-ready, what the biggest risks are, and what to fix first

use this when turning finished work into a deliverable the owner can open anywhere, especially on a phone: Markdown, PDF, DOCX, or XLSX, optionally placed in their cloud drive

use this when designing, reviewing, or querying an audit log for an agent system or product (who did what, when, to what, and why), or when checking that a change keeps that history append-only and trustworthy

use this when deploying any autonomous agent or automated decision workflow, when an agent's track record may justify more autonomy, or after an incident to decide whether to demote it

use this when a person, team, or system is overthinking and under-doing, or too much time goes to planning relative to executing, to find and ship the smallest valuable next step

use this when the owner must choose between building something in-house, buying or subscribing to a vendor, configuring an existing tool, or partnering, and wants a clear recommendation with costs and risks

use this when analyzing raw business data (CSVs, time logs, transactions, error logs) or auditing a process end to end to find patterns, bottlenecks, and automation opportunities quantified in business terms

use this when evaluating an engineering or technical leadership candidate in the AI era, where resumes and demo apps prove little, to test whether they can design a self-improving work system and prove its rate of improvement on velocity and throughput follows an S-curve

use this when designing or tuning a digital twin, principal agent, or role agent so it acts the way a specific person instinctively takes action, using a conative profile (such as a Kolbe-style result), interview notes, or observed working style

use this when putting a confidence number on any recommendation or prediction, or when designing a before/after or holdout test to prove a change actually caused an improvement

use this when designing or reviewing a system whose business rules, thresholds, feature flags, or routing should live in versioned configuration instead of code, so non-engineers can change behavior safely

use this when the owner asks for a recommendation on an operating or strategic decision and wants a short memo: context, options, recommendation, trade-offs, risks, and next actions

use this when a workflow, architecture, governance structure, decision tree, or timeline would be clearer as a picture than as prose, to produce a Mermaid diagram that renders reliably

use this when scoping a new AI agent for one business domain (finance, operations, sales, support, logistics) that will run alone or inside a fleet, and you need a full spec: capabilities, skills, prompts, models, actions, human review, and events

use this when a critical system, process, or project failure (outage, data loss, security issue, failed launch, major customer escalation) needs urgent structured diagnosis and a permanent fix, not for routine bugs

use this when standard approaches aren't working, conventional wisdom is being assumed rather than validated, or a deeply entrenched problem needs a creative solution rebuilt from fundamentals

use this when designing any autonomous agent system, AI workflow, or decision pipeline that needs human oversight, or when turning business objectives into strategies and actions that every automated step can be traced back to

use this when designing or auditing the safety boundary of any autonomous agent or automation: hard limits, human approval gates and queues, kill switches, and the audit trail of every decision

use this when evaluating a new idea, initiative, product feature, technology decision, or proposal that needs a scored, comparable assessment across people, process, technology and revenue, risk, cost

use this when productivity is low despite long hours, context switching is wrecking output quality, or a person or team needs a deliberate work rhythm of deep-focus sprints and real rest

use this when leading a team through change, conflict, or difficulty, or when building trust and psychological safety as the foundation for high performance

use this after finishing meaningful work, or when the owner corrects you, to propose durable memory updates (facts, preferences, decisions) for approval instead of silently changing long-term memory

use this before committing resources to an initiative or AI deployment, when choosing between similar-scoring candidates, or when preparing for executive alignment, to estimate a calibrated probability of success and give a Go or No-Go

use this when preparing to give difficult feedback, building a feedback culture on a team, or receiving feedback, so it's both caring and direct and the loop actually closes

use this when setting up measurement for an AI workflow, product change, or team initiative, at the 4-, 8-, and 12-week marks to judge whether it's working, or when reporting impact and ROI to stakeholders

use this when an idea, product, launch, pitch, feature, or hard-to-reverse decision needs hostile stress-testing before time, money, or reputation is committed, ending in Go, Revise, or Kill

use this when the owner asks for web research on a topic, market, company, technology, or question and wants sources collected and a concise, cited brief that separates facts from analysis

use this before finalizing a high-stakes decision, deploying agents to a new domain, approving a broad strategy, or whenever unintended consequences could hit revenue, risk, or cost, to map what happens next and after that

use this when a team is stuck, demoralized, or has lost direction, or when a new leader needs credibility fast, to deliver three visible quick wins and then rally the team around an ambitious goal

use this when a team or vendor is eager to build or scale a solution, or wants feedback on a proposal or implementation, but hasn't clearly defined the outcome, baseline, validation path, or readiness gaps

use this when an R&D or platform team has a working prototype and needs a phased go/no-go path to a production-ready platform other teams will adopt, proving depth before breadth

use this when an agent or team proposes several related actions, when actions from different domains must be coordinated, or when reviewing whether a bundle of actions really belongs together for all-or-nothing approval

use this when the owner is hiring engineers or a technical leader, or choosing between hiring, contractors, and agents, and needs a role scorecard, an interview loop, questions, and a fair way to score candidates

use this as the first conversation when the Virtual CTO template is imported by a new owner, or when the owner asks to redo setup, see what this bot can do, or who built it