9ine Insights | Latest news from 9ine

Forward-Deployed Engineers: Why schools will need more than an AI code review

Written by Mark Orchison | Nov 17, 2026, 10:35:48 AM

As AI makes it easier for schools to develop their own applications, I believe we are approaching a point where the challenge will shift from creating software to ensuring that it can be securely deployed, properly managed and maintained over time.

Over the past twelve months, there has been a significant increase in investment in what the technology industry describes as Forward-Deployed Engineers (FDEs), with some of the world's largest AI companies recognising that providing access to increasingly capable models and development tools is only one part of helping organisations realise the benefits of artificial intelligence. The more difficult challenge, and one that I believe is often underestimated, is taking that technology and integrating it into the operational environment of an organisation, where it needs to work alongside existing systems, meet security and compliance requirements, and remain reliable as both the technology and the organisation's needs change.

In May, OpenAI announced the creation of its Deployment Company, supported by a consortium of private equity firms and investors including TPG, Advent, Bain Capital and Brookfield, alongside plans to acquire Tomoro and bring approximately 150 experienced Forward-Deployed Engineers and deployment specialists into the organisation. Anthropic has been pursuing a similar approach, launching Ode with Anthropic in July, backed by Blackstone and Hellman & Friedman, followed by an announcement on 3 November that private equity firm Hg and Anthropic were investing in a new European AI consultancy built around Forward-Deployed Engineering. Amazon Web Services has also committed $1 billion to developing its own FDE capability, reflecting a wider recognition that organisations need engineering expertise working directly alongside their teams if they are going to move beyond experimenting with AI and start embedding it successfully into their operations.

I find this particularly interesting because, while Forward-Deployed Engineering is becoming an increasingly established approach within the commercial technology sector, it is not yet a recognised model of support within education, despite the fact that schools are beginning to experience many of the same challenges that are driving this investment elsewhere. In fact, I would argue that the way AI is changing software development within schools makes the case for this type of service particularly compelling, especially when we consider the capacity and specialist skills available within most school technology departments.

Schools are starting to develop their own applications

Over recent months, I have attended a considerable number of conferences and worked with school leaders and Technology Directors across the UK, USA and elsewhere internationally, and one of the more significant trends I am seeing is the increasing number of applications being created by teachers, students, school leaders and support staff using tools such as Claude, ChatGPT, Lovable and Replit. Some of these applications are relatively straightforward, perhaps automating an administrative process or providing a particular teaching resource, while others are considerably more sophisticated, introducing workflows, managing information or providing functionality that schools might previously have purchased from a commercial software supplier.

For the most part, I think this is a positive development, as the ability for someone who understands a particular operational or educational problem to create a working application without needing extensive development experience has considerable potential. Teachers have always been good at finding creative ways to overcome the limitations of the technology available to them, and AI now provides an opportunity to take that creativity much further, allowing them to develop applications that are specifically designed around the needs of their students, departments or wider school community.

The challenges tend to emerge when an application that was originally developed for personal use starts to attract interest from other people within the school, perhaps because it is particularly effective or addresses a problem that colleagues have also been experiencing. What began as an experiment within a personal ChatGPT, Claude or Lovable account may subsequently be presented to the Technology Director with a request to make it available across a department, integrate it with existing school systems or allow other staff and students to begin using it.

At that point, the application starts to move from being an individual productivity tool to something the school has a responsibility to govern, particularly where it begins processing personal information relating to pupils, staff or families, or where it is being used to support activities that are important to the school's operations. The questions that need answering are no longer simply whether the application does what its creator intended, but whether the school understands its architecture, the security arrangements, where information is stored, how access is managed, who owns the intellectual property and what would happen if the person who developed it were no longer available.

This is becoming a growing challenge for Technology Directors, who are often being asked to take responsibility for applications they did not commission, were not involved in designing and which may have been created without any of the documentation, testing or development controls they would normally expect when introducing a commercial application into the school's technology environment.

Why a privacy assessment, cybersecurity review or code review may not be enough

At 9ine, we have recently been developing our Application Assurance service to address the risks associated with schools creating their own applications using AI, recognising that an application developed internally should be subject to an appropriate level of scrutiny before it is adopted more widely, particularly where it processes personal information or connects to other school systems. This includes considering the application's privacy and data protection arrangements, undertaking relevant risk assessments, reviewing the security of the application and its underlying infrastructure, examining the source code and understanding the contractual, intellectual property and operational responsibilities that arise from its development.

These activities remain important, and I believe they will become an increasingly necessary part of how schools manage technology, but I am also conscious that undertaking an assessment and identifying a series of weaknesses does not necessarily provide the school with the capacity or expertise needed to address them. If a code review identifies shortcomings in authentication, database security, application architecture or the way information is transferred between systems, somebody must still take responsibility for resolving those issues, testing the changes and establishing the arrangements required to operate the application securely.

In many cases, this work will fall to the Technology Director and their team, who may already be managing a complex technology environment, responding to operational demands and delivering other strategic projects, without necessarily having specialist software engineering or AI application development skills available internally. The challenge becomes even greater where an application has been created through a series of prompts to an AI development tool, as the person who originally developed it may understand what the application is intended to do without having a detailed understanding of how the underlying code works, the architecture supporting it or the dependencies that need to be maintained.

I think this is where schools will begin to recognise that a privacy assessment, cybersecurity review or code review, while necessary, may only address part of the problem, because an application that is to become part of the school's operational environment needs considerably more thought around how it is designed, deployed, maintained and supported. It requires a documented architecture, appropriate hosting and ownership arrangements, controlled development and testing environments, secure integrations, version management, backup and recovery arrangements, monitoring, clear responsibilities and processes for introducing changes without compromising its security or functionality.

It also requires training, not simply for the people who will use the application, but for those who will be expected to manage and support it, so that the school is not left dependent upon the individual who originally developed it or an external service provider that nobody internally knows how to manage.

Where Forward-Deployed Engineers can make a difference

A Forward-Deployed Engineer is someone who works directly alongside an organisation to understand its requirements, design and develop applications, integrate them with existing systems and support their deployment into the operational environment, rather than working solely from a specification or providing a report on what needs to be improved. The value of this approach is that the engineering work is undertaken in the context of the organisation's actual processes, existing infrastructure, security requirements and long-term responsibilities, with the intention of creating something that can be operated and sustained rather than simply demonstrating that an application is technically possible.

Within a school, I can envisage this starting at several different points, depending on how far an application has already developed and what the school is trying to achieve. It might involve working with a teacher who has created something useful in Lovable or Replit and now wants to make it available to other colleagues, helping the Technology Director understand what has been built and then designing an appropriate architecture for its wider deployment. Alternatively, it might involve working with a school leadership team that has identified a particular operational challenge and wants to understand whether an internally developed AI application could provide an appropriate solution.

In either case, the work should begin with understanding the intended purpose, the information being processed, the people who will use the application and how it will interact with the school's existing technology environment, before moving into the architecture, development, testing and deployment arrangements. This should include establishing school-controlled development and hosting environments, appropriate access and security controls, documented integrations, ownership of source code and intellectual property, and a process for maintaining the application as functionality and requirements change.

There is also an important governance component to this, particularly where applications involve students, personal information or AI-generated outputs, as the Technology Director should not be expected to make decisions about educational suitability, digital safety or privacy in isolation. Our Diamond Formation approach, bringing together Academic, Technology, Privacy and Digital Safety, provides a structure through which the relevant school functions can understand the proposed application, contribute to the assessment of its risks and establish the conditions under which it may be used, with the Head and wider leadership team maintaining appropriate oversight.

I believe the opportunity for a Forward-Deployed Engineer is to work alongside this structure, taking the agreed requirements and governance decisions and translating them into a properly designed, implemented and documented application, while ensuring that the school has the knowledge and arrangements needed to manage it after deployment. This is a different proposition from simply providing a developer to write code, because it recognises that the application must become part of an existing operational environment and that the responsibility for its ongoing functionality, security and development needs to be understood from the outset.

Why we are introducing Forward-Deployed Engineering at 9ine

There is another reason why I believe this is an important direction for 9ine, which comes from our own experience over the past year of concepting, designing and developing an AI-first platform that we are preparing to launch. While we have been providing technology, cybersecurity, privacy and governance services to schools for many years, developing an AI-first platform has required us to bring those disciplines together in a different way, particularly as we consider how AI interacts with information, how it operates within the permissions and controls established by the platform, and how that functionality can be supported and maintained as the underlying technology develops.

Our work has involved much more than integrating an AI model into an existing application, because we have had to think through the overall architecture, the relationships between different areas of the platform, the management of data and user permissions, the security of the infrastructure, the testing and deployment processes, and how we provide ongoing assurance that the functionality continues to operate as intended. We have also had to consider how AI-enabled capabilities will change over time, including the implications of new models, additional functionality, changing dependencies and the need to introduce improvements without undermining existing controls or creating unnecessary operational risk.

One of the lessons from this experience is that creating an application is only one part of developing a sustainable software platform, and that the engineering arrangements around it are just as important as the functionality being delivered. This is particularly relevant for AI-first applications, where rapid changes in the underlying technology can introduce opportunities to improve functionality, but can also create additional dependencies, security considerations and ongoing maintenance requirements.

Building on this experience, alongside our established capabilities in application development, technology architecture, cybersecurity, privacy and governance, we will be introducing 9ine Forward-Deployed Engineering as a service for schools that are developing their own applications, including those created using AI. Our intention is to provide access to engineers who can work directly with school teams, helping them move from an initial concept or working prototype through to an appropriately designed, documented, securely deployed and sustainable application, while supporting the training and ongoing development arrangements needed to maintain it.

For schools that already have an application, this may involve reviewing what has been created, establishing whether the existing architecture is appropriate, identifying what needs to change and undertaking the engineering work required to prepare it for wider use. For others, we may become involved much earlier, working with the school to understand the problem it is seeking to solve, define the functional and technical requirements, consider the relevant governance implications and develop an application that is designed from the outset to operate within the school's environment.

Importantly, I do not see this as creating a model where schools become permanently dependent upon 9ine to understand or operate their own applications, as our objective should be to provide the appropriate architecture, documentation, training and knowledge transfer so that the school retains control and understands the ongoing responsibilities associated with the software it is using. Where specialist engineering or maintenance support is required over time, that can be provided through a defined service arrangement, rather than leaving the school with an application that nobody is equipped to support.

We would like to hear what schools are building

I think we are still at a relatively early stage in understanding how AI-assisted software development is going to change the way schools use technology, but the examples I am already encountering suggest that we will see a considerable increase in applications developed by individuals and teams within schools, many of which will have been created without following a traditional software development process. Some of these applications will undoubtedly solve problems that schools have struggled with for years, while others will raise difficult questions about ownership, security, data protection, sustainability and the responsibilities that arise when something originally developed for personal use becomes part of the school's operations.

Rather than simply waiting for these applications to arrive with Technology Directors for approval, I believe there is an opportunity to help schools approach their development more deliberately, preserving the creativity and flexibility that AI provides while introducing the engineering and governance arrangements needed to support their wider adoption. This should enable schools to benefit from the technology without creating a growing collection of applications that are poorly documented, difficult to maintain or dependent upon individuals whose responsibilities may change.

As we develop our Forward-Deployed Engineering service, we would really like to hear from schools about the applications they are already creating, the technologies they are using and the challenges they are encountering, whether that involves something built by a teacher in Claude or ChatGPT, a more substantial application developed using Lovable or Replit, or a school-led project that has reached the point where specialist engineering support is required.

We would also welcome conversations with Technology Directors who are increasingly being asked to take responsibility for applications they have not been involved in developing, as understanding the practical challenges they are facing will be important in shaping how we provide the service and where our engineers can add the greatest value.

Our experience of building and preparing to launch our own AI-first platform has reinforced my view that the availability of increasingly capable AI development tools does not remove the need for good engineering, appropriate architecture or operational governance, and that these disciplines will become more important as schools begin relying on applications they have created themselves.

If your school is already developing applications using AI, or you are beginning to encounter the challenges associated with deploying and maintaining them, we would welcome the opportunity to hear what you are building and have a conversation about where 9ine's Forward-Deployed Engineers could help.

You can get in touch with us at 9ine.com.