Cleverbit Software

15/10/2026

Tomorrow at 15:00 CEST, we're talking accountability in AI-assisted delivery. Bring your questions, as there'll be time for Q&A with Eman and Ethan at the end. https://bit.ly/4yPF1nI

12/10/2026

Already running ISO 27001? You don’t need a second system to govern agentic AI. Our latest blog explains how to extend your existing ISMS to cover models, agents and AI-specific risks, plus a 90-day roadmap to get there.

Read it here: https://cleverbit.software/blog/iso-27001-ai-development/

29/09/2026

Welcome to Cleverbit, Enrico!

We're delighted to welcome Enrico Cerri to the team as our new Senior Application Support Engineer. Enrico is originally from Italy and now calls Poland home, where he relocated to settle down with his wife.

He's friendly, outgoing and approachable, with great soft skills and a strong ethical grounding rooted in solid family values. Enrico turned his passion for tech into his career.

Outside of work, Enrico is a leather artisan, turning his hands to a craft that mirrors the same care and precision he brings to his job.

28/09/2026

When an AI agent writes the code, who is accountable for the decisions inside it? Commit history used to answer that question, but agents now make decisions of their own and the commit no longer shows who decided what.

Join Cleverbit CTO Eman Zerafa and AI Solutions Engineer Ethan Zammit on Friday 16 October at 1500 CET for "When the agent decides, who answers? Accountability in AI-assisted delivery." They'll cover why code review alone can't keep up, which controls teams have lost and a practical model for putting human judgement where it matters. https://events.zoom.us/ev/AhoGM80UCknJ60hCeKYyi-WYcU8NwoypLN4oxfnSA1SbB0czhsvs~AiBsZirvOUjUMOj3rl9LigkpoEp6PRrAPI7jkauTlH2XrjV2J86EGXer5g

24/09/2026

Thinking about bringing AI agents into your software delivery? Start with a pilot, not a company-wide rollout. Our latest guide walks CTOs through a five-step pilot for an agentic SDLC, covering how to set the scope, which artefacts you need, what to measure and when to pause. Read it here: https://cleverbit.software/blog/5-step-pilot-for-agentic-sdlc-governance-first-playbook-for-ctos/

16/09/2026

Today Eman Zerafa and Ethan Zammit took the stage TechTok with a question every engineering leader will need to answer sooner or later: when the agent decides, who answers?

Source control shows who committed the code, not who made the decisions behind it, and that gap is widening as AI agents take on more of delivery. Their session made the case that everything downstream of an agent should be treated as AI-influenced and checked with the same consistent approach, including a live demo of how teams can keep real accountability as agents take on more of the work.

The room responded with genuine interest, especially around Nissy https://nissy.ai and how it fits into that accountability picture.

Proud of the whole team today, and proud to see the Maltese developer community turn out for a day full of strong sessions.

09/09/2026

Jaques and Evan are at Fintech Week London this week, catching up with teams who are deep into agentic AI adoption or just starting to ask the hard questions about it.

If your engineering team is experimenting with AI in the SDLC and you are not sure how to keep human review and governance in the loop, this is a good week to talk it through. The teams getting real value from agentic development are the ones who built control into the process from day one.

Around and want to grab a coffee, or want to set up a call for after the event? Send us a message, we would love to catch up.

02/09/2026

A firewall either blocks traffic or it doesn't. An AI system can drift from its intended behaviour for weeks without tripping a single alert, and that difference is why most traditional incident response playbooks don't hold up for AI systems.

Our new blog post is a practical guide for SOC and IR teams: what to do in the first 60 minutes of an AI incident, the containment steps for the most common archetypes (prompt injection, data leakage, model poisoning, hallucination, agent runaway), and how to build ownership across security, engineering and legal before the first incident hits.

Read it here: https://cleverbit.software/blog/ai-incident-response/

28/08/2026

A copilot suggests. An agent acts. Mixing those two up is where a lot of AI-in-DevOps strategy goes wrong, because an agent that plans steps and executes changes across systems needs a very different level of oversight than a tool that proposes a line of code.

Our new blog post breaks down where AI is actually earning its keep in DevOps right now (incident triage, log summarisation, runbook generation, infrastructure-as-code) and the guardrails that keep it from becoming a silent source of technical debt: schema validation, sandboxed writes, approval gates scaled to risk, and audit logs nobody can quietly edit.

Read it here:

Discover how AI in DevOps enhances incident management, reduces alert noise, and boosts productivity through practical implementation strategies.

26/08/2026

Most agentic AI failures come down to architecture, not the model itself. If an agent has write access to a pipeline and enough freedom, it will eventually do something nobody planned for, so the real question is whether you can see it happen and step in in time.

Our new blog post lays out a governance-first blueprint for agentic AI architecture: a clear autonomy envelope, version control for every agent-facing artefact, enforcement built into the pipeline itself, and a pilot-first rollout that earns wider scope with evidence.

Read it here: https://cleverbit.software/blog/ai-architecture-design/

Explore a governance-first approach to AI architecture design, ensuring autonomy and accountability while preventing drift in agent behavior.