"To my mind, the most obvious process that can be undertaken by AI in strategic planning is setting the context and identifying the ‘needs’ for development. Data gathering, performance analysis, development monitoring and statistical modelling are very much at the heart of AI"
- Pete Canavan - Carter Jonas
I have been a planner in local authority and in private practice. I have written Local Plans, scrutinised economics, and demographics, and I have guided the strategies which deliver significant development schemes.
In my career to date, I have been at the ‘coal face’ policy writing, and I have had to balance the competing (and often sharply political) priorities which abound. I have seen first-hand how this directly affects planning decisions, and ultimately the delivery of development. This to me, is the essence of strategic planning; balancing priorities across a wider area than in, and immediately adjacent to, the ‘red line.’
I believe the most important exercise is staying on top of the data – maintaining quality evidence – because whatever policies are in place, whether local or national, strategic planning is about places and the people that make up those places.
So, understanding that, and the relationship between the two is key and that understanding can respond to any policy or policy change. Change and growth can and should be a product of that understanding, which is guided by policy, not the other way around.
I am now becoming increasingly aware of the impact of Artificial Intelligence, in all the things that we do, but also specifically in planning. I have therefore put together some thoughts on the potential impacts of AI on strategic planning.
What is strategic planning?
The role of strategic planning is to engage with issues and manage the delivery of development at a wider spatial scale than a single site at a time. Where traditional development management planning – guided by planning applications – is focused on a building, or a site, defined by a red line on a map, strategic planning considers the interrelationships between sites, places, people, and infrastructure.
Strategic planning starts with considering why development is needed, and where it might most sustainably be delivered. The best strategic plans then move on to identify how that delivery will be achieved.
Successful strategic planning is positive and properly justified. It can demonstrate an understanding of social and economic needs; physical constraints and opportunities; of how to weave together infrastructure and other service delivery, balancing this with conserving and enhancing the best of the built and natural environment.
To deliver successful strategic planning one needs to have an effective mechanism to collect and interpret a range of demographic, economic, and other statistical data. This comes from a range of sources including the monitoring of development delivery, population changes as most reliably recorded through the Census, and economic analysis of housing markets, gross national outputs and the like.
This exercise also often includes the need to construct complex statistical models to predict how each element will change in the future (or indeed the creation of standard formulae to avoid any localised interpretation!).
Physical constraints and opportunities, across a wide area, are best understood and interpreted through mapped data. At Carter Jonas we have our own in-house geospatial tools which can quickly map a range of constraints and help us to build a picture of where development might be most achievable – avoiding areas of flood risk, or ecological sensitivity, for example – but also it can identify those locations that are close to existing services and facilities – such as public transport connections, shop, schools and doctors surgeries. This allows us to promote the best, and most deliverable location for development.
For a local planning authority, such tools will allow it to alight upon the most appropriate locations for allocation in its Local Plan.
To enable quality development, however, one needs to consider the infrastructure that is needed to underpin it. This could be improving water supply and sewerage systems, providing new roads or rail stations, new schools, or healthcare premises, and contributing to funding bus routes. These matters need to be considered at a cumulative level, and the delivery timescales and ‘trigger points’ for infrastructure should also be understood.
Infrastructure needs, and delivery, should be understood early on – and ought to be more rigorously scrutinised as part of the effectiveness and deliverability of Local Plans at examination – but all too often these matters are left for the development management process to resolve on a piecemeal basis.
The trick then is strong leadership and agency, quality coordination between stakeholders, and engagement with a range of stakeholders including individuals and organisations. In plan making this is an important role for the local planning authority, but not just in writing a plan and identifying places for development.
The Local Authority should lead the way in understanding when infrastructure is needed to support growth, and using its agency it must bring together those who are going to deliver it so the practicalities, and cost, of new infrastructure are better understood.
The importance of quality consultation
Engagement, of course, must be with those who live and work in the area for which the strategic plan is being created. This engagement must be regular, iterative, and relevant. The whole process of strategic planning must be a shared and transparent exercise, including the ‘why,’ the ‘where’ and the ‘how.’ Critically, the interrelationships between each of those three questions must also be explored, and also the relationships between the potentially competing priorities in the ‘planning balance’. A strategic planning engagement exercise must be focused on delivering an output, rather than identifying how many people and organisations favour one ‘topic’ or ‘theme’ over another.
Quality strategic planning tells a clear and coherent story, can demonstrate deliverability at an early stage, achieves buy-in through engagement, and does not stop with allocated development sites in the plan on a shelf. Quality strategic planning leads directly to sustainable development delivery.
To what extent can this process be undertaken using AI?
AI can be used to write, and respond to, Local Plans. But this feels like a very underwhelming use of its potential. One could gather information, and simply ask AI to formulate policies which allow for development in certain locations, and not others. A site promotor could simply engage AI to argue for slight changes to the wording, or for additional inputs to the data gathering.
This approach, however, would lead to homogeneous plans. It would also avoid the very point of policy writing, which is to articulate the challenges an area is facing, and what the solution to those challenges is, as understood by the people making the planning decisions.
Which processes might be improved upon due to the benefits that AI can bring?
To my mind, the most obvious process that can be undertaken by AI in strategic planning is setting the context and identifying the ‘needs’ for development. Data gathering, performance analysis, development monitoring and statistical modelling are very much at the heart of AI.
Of course, the quality of the inputs will directly correlate to the quality of the outputs but if an agreed approach to understanding such needs, and the inputs such as household projections, and economic growth, can be reached then it would avoid a substantial amount of debate and delay through Local Plan making.
To an extent, AI is already used in geospatial mapping and analysis. Improving the way that sites and areas are analysed for their ‘headline’ constraints would only aid efficient decision-making. If mapping tools can be used to appraise sites consistently and efficiently for their development suitability then land availability assessments could be quickly updated, and a large evidence base for plan making could be managed and maintained.
Infrastructure modelling too, ought to be something that can be achieved using AI. The phasing of development, and infrastructure delivery and the dependencies between the two would be something well suited to the powers of AI.
What will be left for humans to do?
It is that final step in strategic planning, however, that will be very difficult for AI to replace. Selecting a spatial strategy, and set of development sites that can deliver, in a place where people what to live and/or work is – in my opinion – a step too far for AI and requires personal engagement and professional judgement.
One could argue that removing the personal or human element makes for more efficient plan making. If one could use AI to rule out areas where development could not be achieved and can gradiate other areas - avoiding more ecologically sensitive areas for example - then link together all those places that were left (and perhaps call them a development zone), it would be very straight-forward to simply allow development in those places.
This approach, however, forgets the very essence of strategic planning; that it is about maintaining and creating places for people. The importance of planning judgement – borne out of the shared experience - is that it allows space for inherently ‘messy’ considerations, such as how people interact with the spaces and buildings around them, and of course for politics and changing priorities.
I also believe that engaging with individuals and organisations (and consultations) is best and most valuable when it involves discussions between people. True engagement, and agency, come from empowering people and helping them to understand the competing priorities and the potential outcome and helping all stakeholders understand how they can affect the final development that is delivered. This is something that I simply do not believe that AI can achieve.
There is a further risk, which is that AI – or those setting the parameters for AI – becomes too effusive with identifying or mapping constraints and we forget that development can be a way to overcome certain constraints or to improve a situation.
This again removes the element of judgement – or site promotion – that might allow for development to enable the better preservation of a heritage asset for example or deliver new service and facilities in a location that is currently under served.
When will the change happen?
Some use of AI – or at least enhanced use of IT – has made its way into the planning system. Geospatial tools are in widespread use, alongside large databases of (relatively static) information. The future challenge is to make these data stores more responsive and have better up-to-date monitoring.
These tools should also be able to auto-report the suitability of a site or area for development given an objective and simple set of criteria, upon which further judgements can be made, and discussions had.
Increasingly Local Plan ‘consultations’ are moving to online platforms, and no doubt AI is being used to analyse interactions with these platforms (for example, to see how long each visitor is spending on each subject page).
However, I am yet to see such a platform that has successfully demonstrated the interrelationships between the three strands of sustainability: social, economic and environmental needs.
Instead, most consultation platforms prefer to treat each theme, or even subtheme, separately, which can only lead to siloed responses and a division of priorities rather than any sophisticated confluence of ideas.
Other uses of AI – in modelling – are likely to increase in their use quite rapidly in the next few years. I see this as a positive if it can help us to understand what change might look like and open up discussions about how we manage that change.
How do we adapt to the change?
Clearly, we are all going to have to adapt to the use of AI in all walks of life. Plan-makers and decision-takers especially will need to be aware of its influence on people’s opinions, and on the way that people respond to consultations.
However, in a world in which we have a resourcing crisis in planning, data handling ought to remain a key driver for policy making and strategic planning in particular. In this case, then investment in AI which is focused on evidence gathering, development monitoring and statistical modelling can be very beneficial.
AI can set the context for growth and change and free up the time to engage with individuals and organisations. Collectively then, conversations can be had about what ‘good change’ looks like and AI-driven models should be easily updated to respond to these conversations.
Planners will have more time to apply judgment about spatial options and weigh the benefits and harms of development across a wide area. AI can also be used to test assumptions about the deliverability of spatial strategies and effectively knit together infrastructure needs and when these will be provided through strategic development.
