AI in social housing procurement:
where it helps, where it doesn’t, and what your team needs to get right
AI in social housing procurement: where it helps, where it doesn’t, and what your team needs to get right
Across the social housing sector, procurement teams are using AI tools daily. Some are doing it deliberately, with policies and oversight. Many more are doing it informally, using ChatGPT to draft a specification, asking Copilot to summarise a tender response, or running supplier data through an AI-powered analytics tool without fully considering what that means for compliance and accountability.
This is not a future concern. It is happening now, on both sides of the procurement process, and it is moving faster than most organisations’ governance frameworks can keep up with.
The question for social housing procurement teams is not whether to use AI. That ship has sailed. The question is how to use it in a way that strengthens your procurement practice rather than quietly undermining it.
What AI is already doing in housing procurement
It is worth being specific about what is actually happening, because the conversation about AI often stays at a theoretical level while the practical adoption races ahead.
On the supplier side, contractors and product manufacturers are using large language models to write tender responses. For SMEs in particular, this has been a genuine leveller. Firms that previously lacked the resource to respond to complex tenders can now produce well-structured, articulate submissions. That is a good thing for competition and for the sector’s ambition to open up procurement to a wider supply base.
On the buyer side, procurement professionals are using AI to draft specification documents, structure evaluation criteria, summarise lengthy tender submissions and categorise spend data. Some are using it to help interpret the Procurement Act 2023, which introduced new terminology, new procedures and new transparency obligations that many teams are still working through.
None of this is prohibited. The Government’s Procurement Policy Notice (PPN 017), updated in February 2025, explicitly states that suppliers’ use of AI in bid writing is not banned. What it does require is transparency: contracting authorities should ask suppliers to disclose their use of AI, and those disclosures should be used for information, not scored as part of the evaluation.
The challenge is not that people are using AI. It is that the guardrails around its use have not kept pace with the adoption.
Where AI creates risk for procurement teams
Three areas stand out as particularly exposed.
Tender evaluation and scoring. When every supplier has access to the same AI writing tools, the quality gap between responses narrows significantly. Bids become polished, well-structured and comprehensive, regardless of whether the firm behind them can actually deliver what has been promised. This is a real problem for evaluation panels, who now face a stack of responses that read beautifully but may not reflect operational reality. The risk is that competitions collapse into a price-only contest, because there is so little differentiation on quality. That rarely produces good outcomes for housing organisations or their tenants.
Evaluators need to be alert to this. Probing questions, clarification interviews and reference checks become more important when the written submission alone can no longer be taken at face value. AI has raised the floor on bid quality, which means the methods procurement teams use to assess genuine capability need to rise with it.
Feedback and standstill letters. After a contract is awarded under the Procurement Act, unsuccessful suppliers receive feedback explaining their scores and the rationale behind them. This is a legal obligation, not a courtesy. Getting it wrong exposes the contracting authority to challenge.
It is tempting to use AI to speed up the drafting of these letters, and in principle there is nothing wrong with using it as a starting point. But the final letter must accurately reflect the evaluation panel’s reasoning, reference the specific scoring criteria and respond to the individual submission. A generic, AI-produced letter that does not align precisely with the moderated scores is a liability. Human review and sign-off are not optional here.
Confidentiality and data handling. This is the risk that gets the least attention but may carry the most consequence. When procurement professionals paste tender documents, supplier pricing or internal strategy papers into a public AI tool, that information may be used to train the model. PPN 017 is explicit on this point: contracting authorities should put controls in place to prevent confidential information, or information not already in the public domain, being used as training data for AI systems. Most housing organisations do not yet have clear policies on which AI tools are approved for procurement use and what data can be shared with them.
Where AI genuinely adds value
The risks above are manageable, and they should not discourage procurement teams from using AI where it is genuinely useful. There are several areas where it can strengthen rather than weaken the procurement process.
Specification drafting. AI is effective as a first-pass drafting tool. It can pull together a structure, identify relevant standards and accreditations, and flag common requirements for a given category. The key is that a technical expert then reviews, refines and owns the final specification. AI does the scaffolding. The expert does the building.
Sense-checking and moderation. One of the more promising applications is using AI to review your own work. After an evaluation panel has scored a set of tender responses, AI can be asked to assess whether the scores and written feedback are consistent with the published evaluation criteria and weightings. It is not making the decision. It is checking that the decision-makers have been fair and consistent. That is a useful safeguard, particularly for organisations managing multiple concurrent procurement exercises.
Spend analysis and contract management. Social housing procurement teams manage large volumes of spend data across many suppliers and product categories. Categorising thousands of individual material lines, spotting off-contract purchasing and identifying price anomalies are tasks that AI can perform faster and more accurately than manual processes. This is where AI’s capacity to process and classify data at scale delivers the clearest productivity gains, and where the risks of error carry less consequence than in evaluation or legal correspondence.
Regulatory navigation. The Procurement Act 2023 introduced a new regime that many teams are still learning. AI can be a useful reference tool for clarifying procedural questions, checking timelines and understanding obligations. But it is essential to verify any AI-generated guidance against the legislation itself or authoritative sources like the Government’s Procurement Act guidance. AI tools can confidently present incorrect information, and a procurement decision based on inaccurate legal interpretation is a serious risk.
What governance looks like in practice
The gap in most housing organisations is not awareness. Teams know AI is powerful and that it carries risks. What they lack is a practical framework for managing its use within procurement.
That framework does not need to be complicated, but it does need to exist. At a minimum, it should cover which AI tools are approved for procurement use and which are not, what types of data can and cannot be entered into AI systems, where in the procurement lifecycle AI-generated content requires mandatory human review before it is finalised, and how the organisation will comply with PPN 017’s transparency requirements when engaging with suppliers.
The LGA’s guide to responsibly buying AI is a useful starting point, as is PPN 017 itself, which includes example disclosure questions that can be incorporated into tender documentation. But the practical application of these principles within a housing organisation’s procurement team is something each organisation needs to work through for itself.
Getting the balance right
AI is not going to replace procurement professionals. The judgement, context and accountability that sit at the heart of public procurement cannot be automated. But AI is already changing how procurement work gets done, and organisations that ignore this, or that adopt it without governance, are equally exposed.
The social housing sector is under sustained pressure to do more with less. AI offers genuine help with that challenge. The task now is to use it with the same rigour, transparency and accountability that the Procurement Act demands of every other part of the process.
Rob Peck is Chief Commercial Officer at Procurement for Housing (PfH)
This article is based on an opinion piece published in Construction Management.
How PfH can help
Quantum, our spend and contract management platform, uses data classification and analysis tools to give you line-level visibility of spend across all your suppliers, helping you spot off-contract purchasing, benchmark prices and evidence value for money.
Consultancy support to help your team develop procurement specifications, structure evaluations and manage contracts with the right balance of efficiency and human oversight.
Frameworks and DPS solutions provide compliant, pre-tendered routes to market, reducing the procurement workload where AI governance is still being developed internally.
About Rob Peck
Rob Peck is a senior procurement leader with close to two decades of experience across the social housing sector, specialising in procurement strategy, solution design and commercial development.
Currently Chief Commercial Officer at the organisation behind Procurement for Housing (PfH), Rob leads the commercial strategy that supports more than 1100 social housing providers across the UK. He works with social landlords and local authorities to position procurement as a strategic function that delivers measurable value, driving efficiency and improving outcomes for residents and communities. He is a recognised voice on the intersection of technology and public sector procurement, having authored guidance on the responsible use of AI in construction procurement, published in Construction Management, drawing on his deep understanding of tender evaluation, contract management and regulatory compliance under the Procurement Act 2023.
Before taking on the CCO role, Rob spent six years as Director of procurement services and five years as Head of Procurement services at Inprova, where he led teams delivering procurement audits, organisational spending reviews, strategy development and outsourced procurement solutions for housing organisations. Earlier in his career, Rob held commercial and procurement roles at CEL Group, managing EU framework agreements for the social housing sector.
With expertise spanning category management, consultancy, stakeholder engagement, supplier relationship management and solution design, Rob brings a commercially minded, partnership-led approach to procurement, focused on helping housing organisations spend smarter and deliver better services for the people they house.
Sources:
Cabinet Office, PPN 017: Improving transparency of AI use in procurement, February 2025
Local Government Association, Responsible buying: how to build equality data protection your AI commissioning