AI has a helpful role to play in housing procurement, writes Kathryn Irons, but warns that it also present risks
AI is becoming part of everyday working life.
But what happens when your work is public procurement and your sector is social housing – a highly regulated and scrutinised environment with heavily stretched budgets, where every purchase must be transparent, defensible and shown to positively impact tenants’ lives?
Using AI too freely risks undermining the accountability that is so important in my line of work. It provides reasoning that might be surface level or inconsistent, lacking the deeper, real-world judgement that allows housing providers to defend the procurement decisions they make every day.
Having worked in public procurement for many years, I’ve come to recognise that it’s the skills only humans can bring – contextual understanding, proportionality, nuanced thinking, ethical reasoning and relationship-building – that are key to ensuring public money is spent fairly.
Despite this, there is definitely a role for AI in social housing procurement, one that will free buyers from transactional tasks so they can focus on strategic thinking, partnership development and adding real value to their organisations.
Standardised approach
The challenge, however, is developing a standardised approach. Right now, housing associations and councils are dipping in and out, using AI for some procurement tasks, over-using it for others and shying away in certain areas due to risk of challenge.
There is little guidance to support and inform buyers so they can confidently harness these platforms with the right controls in place.
I’ve spoken to colleagues internally and externally about how they are using AI in public procurement, and I hope these insights prompt more debate.
One emerging theme is being clear on boundaries.
Let’s take tender specifications as an example. AI can be a research assistant as you consider the exact goods or services you need, along with supplier performance standards. It can find information on previous procurement exercises, identify similar specifications, research relevant products and suggest the latest supplier accreditations.
But that’s where the line should be drawn. A technical expert, someone with specialist category experience, who has considered the unique needs of your housing organisation, must draft the full specification.
The same goes for bid evaluations. AI is a useful investigation and analysis tool as you dig into bidders’ pricing, helping to benchmark their market position and surface any inconsistencies.
Once you and colleagues have scored a bid, AI can also verify your work, testing the robustness of your reasoning against scoring criteria. It offers an impartial perspective, acting as a neutral moderator.
Problems occur, however, when the line is not clearly drawn and AI moves from sense checker to decision maker. Using AI upfront for the evaluation of bids, for instance, exposes social landlords not only to legal challenge but also the risk of appointing the wrong supplier. Human oversight and governance are essential when scrutinising the capability, resources and capacity of a bidder, along with the case studies they provide.
Boundaries matter
Another area where boundaries matter is offer letters.
Feedback sent to unsuccessful suppliers must be written by an experienced procurement professional. Supplier challenges around these formal written notices are on the rise and it takes background knowledge and understanding to draft robust letters. It’s only right that suppliers who have worked on a bid for weeks receive a detailed outline of why they achieved certain scores.
AI can play a part in offer letters, such as summarising feedback from everyone involved in the quality evaluation. You may have operatives, tenants and service managers all responding in different ways and AI can analyse and consolidate central themes which helps with the time-consuming process of offer letter writing. But this role is consolidation and summarisation, not judgement or justification.
Under the Procurement Act’s new competitive flexible procedure, social housing buyers can, for the first time, enter into negotiations with suppliers: during preliminary market engagement, after initial shortlisting and just before final award. AI can help here too, extracting commercial insight from contracts, assessing and deconstructing pricing, and anticipating supplier responses to help buyers prepare for negotiations.
AI’s role as information interpreter and navigator is also playing out as procurement teams get to grips with the Procurement Act 2023. Checking what should come next after following a particular procedure or issuing a certain notice can be very useful. So can AI’s ability to bring together sector-wide best practice to inform the development of KPIs and other service level parameters within a contract.
Wider guidance
But it’s not only housing associations and councils that are using AI in procurement. Suppliers and contractors are increasingly leaning on AI to write bids, helping to level the playing field for smaller firms without bid writing expertise.
There are drawbacks though. AI-generated tender responses can oversell, leaving firms unable to deliver what they have promised. On the other side, housing procurement teams are receiving many more highly polished submissions which can feel generic or unsubstantiated. These seemingly ‘perfect’ bids take longer to check and differentiate, particularly around quality criteria, and they can push competitions into price-only races that rarely deliver good outcomes.
As a sector we need more communication with suppliers around using AI in bid-writing. Tender documents could include clearer wording reminding companies that they are contractually required to deliver any commitments in their responses – whether they have been drafted manually or via AI.
This type of clarification must be part of wider guidance that our sector needs. Current direction, including the government’s policy notice ‘Improving transparency of AI use in procurement’, published last year, is not specific enough to social housing.
By creating shared standards, we can ensure that housing procurement professionals and our supply chain partners make the most of AI so it supports their work and doesn’t drive it. The core principles of public buying: integrity, value for money, fairness, accountability, competition and transparency must always be led – and controlled – by human hands.
Main image: Kathryn Irons is head of key accounts at Procurement for Housing
Read next: Travelodge veteran to lead Orbit’s empty homes overhaul
Are you a social housing professional? Sign up for a FREE MEMBERSHIP to upload news stories, post job vacancies, and connect with colleagues on our secure social feed.
