Is AI essential to optimise your flow?

Our opinion: Not yet.

AI is now part of many people’s workflow, from curating emails to writing complex code, it is hard to avoid it. There is no doubt it is a helpful addition and enhancement to productivity when used correctly (and with the right governance).

Our opinion from our experience using AI directly and working with companies embedding AI into their workflows is that AI is not essential to optimise flow in your engineering system to deliver value. At this stage anyway. This is likely to change in the future. 

For this reason we have decided not to add AI into the Hierarchy of Engineering Needs as a need in its own right.

However we do believe that AI can and should be used to support and lift the maturity of existing needs. Therefore the effective use of AI can be an indicator that these needs are more mature within an organisation. We are proposing the next release of the HoEN model to include AI usage across needs such as these. 

It would be easy to include AI across nearly all of the needs in some form, but we are concentrating on practical and proven usage of AI that we believe lifts the maturity of the need significantly. Here our thoughts on key areas of the model we are looking to update and we are keen on your thoughts, so please share your comments.

Basic Needs

Technical Capability

AI acts as your always available pair to lift technical capability, whether you're coding, designing, or testing. In development, AI pair-programming tools offer contextual suggestions and reduce time writing boilerplate code. In testing, generative AI helps create test cases, detect flaky tests, and speed up feedback loops, enabling teams to ship with confidence and less manual toil.

AI enhances and speeds up your learning of technical concepts, providing contextual information to help understand and provide solutions to complex problems. AI assistants summarise docs and surface relevant context, helping engineers onboard faster and find what they need without digging. They can help you understand that legacy code no one has touched in years, reducing the risk of technical debt.

Local Dev & IDE

Correctly integrated with techniques such as with instruction files, AI can provide tailored and context specific assistance. AI in editors speeds up coding with real time code suggestions, flags bugs early, reducing defects for higher quality code, and improves adherence to style, architectural patterns and standards. However if given too much freedom the non deterministic nature of AI can reduce consistency and coding efficiency.

Environment Management

AI can generate and anonymise reference data, making it easier to create realistic environments with minimal overhead.

Workback Log

AI helps extract and refine requirements from documents or conversations, analyse customer feedback at scale, and suggest data informed prioritisation.

Managed Work

Quality Engineering

AI generated tests, risk based prioritisation through analysis of historical data, and smart failure diagnosis are improving test reliability and reducing release blockers.

Information Management

AI tools that auto generate and summarise documentation (from code or meetings) help ensure decisions and knowledge are captured clearly and shared efficiently. Model Context Protocol (MCP) servers for documentation and tool discovery.

Security Controls

AI continuously scans codebases for vulnerabilities, flags policy violations, and supports compliance audits with auto generated evidence and suggested mitigations.

Alerting

AI analyses logs to detect anomalies, reduce noise through smart grouping, and propose likely root causes, shortening MTTD/MTTR.

Effective Ownership

Templates and Goldenpaths

AI can learn from successful past projects and existing codebases to dynamically suggest or generate tailored templates and golden paths for new services or features, ensuring best practices are embedded from inception and reducing boilerplate.

Continuous Integration

AI predicts build failures, resolves merge conflicts, prioritises critical tests, and delivers lightweight automated reviews.

Static Analysis

AI enhances traditional static analysis with fewer false positives and intelligent fix suggestions, making code reviews more effective.

Sustainability

Product Metrics

AI identifies themes and sentiment from user feedback, predicts churn, and surfaces product insights that inform roadmap decisions. Deep research agents help with user research analysis.

Career Growth

AI can act as a career coach and open up a lot of information. It can help provide both technical and professional skills growth.

Optimised Flow

Hypothesis Driven

AI can strengthen hypothesis-driven development by acting as an accelerator of insight generation, experimentation planning, and data analysis.

Chaos / Game Days

AI plays a supportive role in designing chaos scenarios, detecting failure patterns, and automating response validation, making chaos practices more proactive and insightful.

AI: A Force Multiplier, Not an Underlying Need

AI doesn’t replace strong engineering practices, it amplifies them. As Paul Meyrick has already made clear in his article (AI Can Code, But Can It Worry Like a Mother?), this can only happen with clear AI governance in your organisation.

Mature teams are using AI not as a shortcut, but as a force multiplier for code quality, operational speed, and decision clarity. That’s why, in future versions of the HoEN model, you’ll see AI indicators embedded across key needs not as a new pillar, but as a sign of maturity already in motion.

As always we value community feedback to help enhance our HOEN model, if you are using AI in your engineering system we’d love to hear how it is lifting the maturity of your engineering needs.