The main challenge for development finance lenders is that each lender, in isolation, is not big enough to invest millions of pounds in a top-notch development finance platform. When it comes to tech startups, the priority is to develop technological solutions for mortgages, BTL and bridging lenders, as the market is much wider, and the product is faster, cheaper and easier to develop.

Another challenge is the need for deep knowledge of the end-to-end process and expertise when it comes to the development finance product.

Can we really trust AI to summarize a 200-page valuation report related to a development finance transaction and cross-check it against the Initial Monitoring Surveyor report, Report on title, drawings, planning permission, initial enquiry, credit paper, facility letter, and so on?

An average development finance transaction will have approximately 300 documents processed, reviewed, analysed, and/or produced by a lender.

It is obvious that we desperately need a solution to assist in streamlining the process. Not every step of the process, however, should be automated, and careful consideration should be given to which parts of the process will benefit from automation and which elements should remain intact.

Let’s look at the early customer journey, for example. Each transaction starts with an initial enquiry. These enquiries come in so many different shapes and forms: email, personal meetings on-site, phone calls.

The information is sent in attachments, including documents in Word, Excel, PowerPoint, PDFs, email write-ups, or voice notes. It would be almost impossible to create a technological solution that will collect data from all these mediums and summarise it for the lender to review.

•   The Finance Professional Show 2023: The Video
•   Uliana Kuzmis leaves HTB
•   Will 2025 chart a new course for UK housebuilding?

It is also not uncommon for the information provided within the initial enquiry to be inconsistent; for example, the write-up in the email might refer to a slightly different budget than the budget included in the attached development appraisal. This happens due to the fact that development finance transactions are fluid, and changes can happen almost on a daily basis. It is not uncommon to receive two different development appraisals in the same enquiry.

In these circumstances, the introduction of AI might not add much value, or even be detrimental to the process, due to the increased risk of inaccurate data interpretation.

Equally, a lender’s portal forcing brokers to populate the relevant fields will be seen as too onerous, as most brokers submit an initial enquiry to multiple lenders in order to present the client with a wide range of options before narrowing down to one lender of choice.

So, is there a solution?

As an industry, we would benefit from a uniform framework that defines the process and the required data each lender needs at each stage of the process. Perhaps a code of good practices and guidelines that lenders and brokers can lean on.

If data is uniform, it is easier to collect, input, and analyse, which will ultimately improve transparency and efficiency of the process and allow AI to assist both lenders and brokers on the journey.

Perhaps this is the perfect opportunity for trade bodies to come together and assist the development finance industry with the creation of this uniform framework.

The framework, in turn, will allow the development of a platform that will appeal to a wide range of development finance lenders, providing economies of scale for tech companies.

Food for thought for 2025?