Client story AI Adoption Consulting and Fluency Coaching
CLIENT STORY
AI adoption consulting and fluency coaching
Client
The world’s leading education provider and the UK’s largest awarding body engaged Insight AI to build AI fluency in its English Language Learning engineering division. The programme coached 60+ engineers, QAs and tech leads.
Industry: Education
Challenge
A baseline survey showed that AI use had spread faster than AI skill. 74% of surveyed teammates reported using AI daily, but depth had not kept pace with breadth. Security and compliance anxiety was the single most cited barrier (19 mentions in survey responses) across every role and squad. Engineers unsure what was safe to share with AI were either avoiding it entirely or second-guessing every prompt.
Outcomes
Over three months, Insight AI designed and delivered a structured AI fluency programme grounded in three modes of AI interaction: Automation, Augmentation and Scalable Agents. It aimed to develop four core competencies: efficient delegation of work to AI, effective prompt and context engineering, discernment of AI outputs and behaviours, and diligent AI use that is ethical, secure and safe. Results included a 39% average reduction in lead time to first code commit, a 69% average reduction in PR approval time, and AI agents built with Insight AI’s coaching guidance.
Insight AI was selected for its technical depth and its ability to work within engineering teams’ delivery context
The client needed a partner who could diagnose the real AI adoption gaps rather than assume them, and design an intervention that would land with technically sophisticated, time-poor teams. Insight AI began with a structured baseline survey rather than an assumptions-led curriculum, giving an evidence-based picture of where the cohort actually was across all four fluency competencies. That consulting rigour, and the ability to communicate credibly to engineering leadership, established the trust that made embedded coaching possible. Insight AI was chosen because it could coach the difference between AI adoption and AI fluency, and make that difference matter to the people doing the work.
Our solution
Over three months and four phases, Insight AI combined workshops, embedded coaching and reusable standards to build AI fluency across the engineering teams.
Key highlights:
- Workshops: A four-week “AI Fluency” workshop series and assessment, showing that AI fluency is tool and discipline agnostic. The same principles applied in a QA context (test case documentation) and an engineering one (code refactoring).
- Embedded coaching: Six weeks working with engineers and QAs in their actual codebases and test suites on real delivery tasks, with demos and example scenarios.
- Standards and artefacts: An AI usage governance guide, a task decision framework, role-specific prompt libraries, best-practice playbooks, tooling and integration guidance and more, published for all squads and designed to outlast the engagement.
- Champion network: The most advanced practitioners were identified and given visible ownership of emerging standards and a teaching role in the embedded phase.
Benefits and outcomes
The programme delivered measurable gains in delivery speed, testing efficiency and team confidence, and produced agents that move effort towards higher-value work.
- Faster delivery: 39% average reduction in lead time to first code commit and 69% average reduction in PR approval time
- Test coverage: 9% average improvement in automation test coverage (E2E and API testing)
- QA efficiency: 2–8 hours saved on test creation per change, 10–30 minutes per bug report, and meaningful time saved on test documentation and data generation
- Perceived impact: 100% of developers report that “AI tools save me real time in actual work I deliver” (mean 6.06 on a 7-point impact scale)
- Reusable standards: 14 reusable AI adoption standardisation artefacts created
- Recommendation: 85% of teammates would recommend the coaching and content to others, per the closing survey responses
From adoption to scalable agents
The agents built with Insight AI’s coaching guidance demonstrate durable capability and adoption upskilling, with effort moving from repetitive manual execution towards higher-value review, governance and judgement.
- Accessibility Testing Agent: Parses UI context, executes compliance checks and auto-generates Playwright and axe test scripts, with a QA engineer retaining responsibility for reviewing flagged violations. Manual effort that previously took 20–38 hours per page has been condensed to approximately 30 minutes. Scaled across five pages, that collapses an estimated 4–5 weeks of work into roughly half a day.
- Unit Testing Agent: Automatically writes and improves unit test suites for React components, producing test code that matches existing frontend conventions. Time to produce a unit test suite dropped from up to 8 hours for high-complexity work to under 10 minutes. Frontend test coverage also increased from below 5% to ~51%.
