Every quarter, we publish what most consultancies keep private: the AI tools we use, what we dropped, and the measurable impact on delivery speed and code quality.
Get each report in your inbox
Latest Report
The full quarterly edition: headline metrics, the complete tooling stack, and what changed across delivery.
84% AI-authored code, up from 51% last quarter. Agent workflows, expanded MCP integration, and measurably faster delivery across every active project.
Between Reports
Shorter observations published between the quarterly reports, on tooling shifts and decisions worth surfacing early.
In April we credited OpenSpec for our 84% AI-authored code. Two months on we have moved away from it, and the reason is about visibility, not tooling.
Why we run a 3-month review cycle on our entire AI stack, and what a $60 billion Cursor acquisition means for the tools behind our delivery numbers.
Talk Think Do runs a structured quarterly cycle: evaluate AI tools, deploy what works, measure outcomes, adapt. Most consultancies do some version of this internally. We publish ours.
The AI Velocity Report tracks our AI velocity metrics every quarter: the share of code that is AI-authored, delivery speed compared to traditional build, and defect rates. These are real numbers from production delivery, not marketing estimates.
Transparency in AI coding is something the industry needs more of. Most consultancies make vague claims about AI-augmented software delivery. We show the data: what changed, why, and what it meant for clients.
Every line of AI-generated code at Talk Think Do is reviewed by senior engineers, validated by ISTQB-qualified QA, and delivered within our ISO 27001-certified security framework.
Book a free 30-minute consultation. We will give you an honest assessment of your options.
Book a consultation