The lack of automation ideas in the UK Gov Green Paper on procurement reform

The final entry on this series about the Green Paper on public procurement reform is on focused on the disappointing lack of ideas around automation. For a once in a lifetime "root and branch" operation on public procurement rules, I would have expected to see an effort in moving the ball forwards in terms of what menial work could be automated away as not adding value to the procurement exercise. It does not help that there is nay a peep on electronic platforms in the green paper.

In general I propose an "automate first" or "automatic by design" approach to public procurement, effectively making information move from system to system automatically and only being stopped for human intervention where necessary. I'm not asking for a pie in the sky approach, ie "let's drop AI on this" but simply an integration/communication between data silos that already exist their connection with electronic procurement platforms. Unfortunately, that is not really the road the green paper decided to travel.

The problem with automation is that it does not blend well with flexibility or discretion which are two of the main objectives of the green paper (well, at least the first explicitly). Tasks to be automated cannot rely on judgment or provide an infinite number of alternatives. For example, the idea of creating a procedure which can be morphed into whatever form the contracting authority deems appropriate is orthogonal to a system built for simplicity (as argued before) and also for automation.

Even with the disclaimer above it is possible to conceive a couple of areas where automation should be considered without cutting across the main proposals of the green paper. The first is menial work; the second, activities which do not add value to the identifying the best bid.

Menial work

There is plenty of work in public procurement which is done manually and that could be automated away if systems were designed as such. Out of the top of my mind I can think of notices as a prime candidate for automation. These should be populated and sent automatically from the electronic platform to whatever publishing platforms they are supposed to be made available, in effect transforming them from a concept of 'notice' to one of 'notification' (as conceived by mobile operating systems that is). These include not only the initial notice at the beginning of the procedure but also those at the end with the information which is mandatory to be made public, effectively ending the current time lag between contract award and contract information being made available. In this vein, contract modification notices should also be automated away, something that is beautifully done in Paraguay for example.

Non-value added activities

As for activities which do not add value to procurement, that is those which do not require an evaluation nor contribute to identify the best bid, we have administrative checks and virtually all exclusion grounds. These should only be mandatory and automatic, and as I suggested in this previous entry they should be taken away from the hands of contracting authorities and centralised. And at least for those depending on administrative data, any queries should be resolved exclusively on the public side without encumbering neither the economic operator nor the contracting authority.

In effect, what I'm proposing amounts to a centralised supplier registration system which would move the assessment of pass/fail conditions from contracting authorities to central government which already holds (most) of the data, thus internalising the cost on the public body pushing for compliance with the exclusion grounds.

I will posit as well that the proposed move from standardised procedures to a flexible one will lead to an expansion of non-value adding activity.

Procurement work done by procurement officers (sorry, 'commercial teams') should be focused only on the activities that help them find the best bid: collecting and parsing information before starting the procurement process; defining selection and award criteria; evaluating those. That's it.