AI for Project Finance

I’m not sure if/how AI will transform project pricing for complex service providers, but the use case is far more clear to me for how complex service providers manage their project finances.

Conversations around the industry tell me that AI functionality is being discussed and in some cases developed across numerous project finance systems, but that actual live functionality, let alone tangible results, isn't there yet.

Here are a few use cases which, once satisfied, would tell me that AI is truly making a difference in managing project finances.

Revenue Variances

Right now, humans have to download revenue reports that compare monthly revenues generated from the system and compare them, typically in a manual manner, against what was expected, what occurred last month, etc. The resulting variance reports are then sorted, highlighted, and so forth to call out what variances need researched, reconciled, explained, escalated, and so on.

Humans then need to present on these results and, in discussions with senior management, eventually settle on final revenue numbers for the month.

I’ve been both the analyst preparing the revenue reports and the VP Finance working with the CFO to determine when we can call the month closed.

Several AI use cases for me are:

• The user says “compare this revenue report against the forecast and provide me the variance analysis”, and the report is ready to review.
• The user asks “why did projects XYZ123 and ABC123 come in below forecast?” and receives a breakdown of which revenue components missed expectations (e.g., the Clinical team didn’t achieve as many units as expected).
• The user asks “which of our top 100 projects are consistently missing revenue estimates each month?” and receives the results in a concise format.
• The CFO asks “are some analysts estimating better than others?” and receives a report showing forecasting accuracy, sorted by best to worst variances.

These are just a few examples, but a project finance system that can’t perform more than 1 or 2 of them does not, to me, possess much AI capability.

Project Profitability Trends

In similar fashion to revenue variances, researching project profitability is a time-intensive experience where the high-levels number might be easy to find, but the root cause is not. The high-level numbers can often deceive, as a project that appears to running perfectly fine could actually be massively efficient in one area but massively deficient in another, and remedying the poorly performing area could make the project a top performer.

This same concept applies across group of projects, and the company at large. It’s good to know if Department A is performing inefficiently on the project in front of you; it’s even better to know if that inefficiently is a Department A issue across most other projects, or projects meeting a certain criteria.

I want to see AI make the research process far faster and more automated, so that the humans can focus on actioning what will improve a project’s profitability. Research, in and of itself does not improve profitability- only successfully actioning its results!

Several AI use cases for me are:

• The user says “for project XYZ123, tell me how efficiently each department is delivering its services”, and a useful report summarizing such is provided.
• The user says “tell me how well Department B is performing against budget across all projects that started in the last 6 months.”, and receives the resulting summary.
• The user asks “is there any (timesheet, unit, etc.) data missing for project XYZ123 this month that would be expected by now?” and receives a bullet point list of recommendations based on recent project activity.
• The user asks “based on units invoiced to date, are there any units that will likely exceed the contract?” and receives a report that could lead to creating a change order to keep the budget up to date.

If a project finance system AI can’t provide actionable recommendations to improve project profitability, you’re going to have a hard time selling me on that AI’s ROI.

Cash Flow

I’ve always thought of each project as its own mini-business. A project generates revenue, incurs costs, and, with respect to cash flow, either generates positive cash flow the business can use, or needs “financing” from the business when cash flow turns negative.

Few service providers I’ve worked for or with ever review project-level cash flow reporting. Part of the reason is because too many service providers focus exclusively on EBITDA and other P&L type reporting, but part of the reason is the failure of project finance systems to provide cash flow metrics at the project level.

Negative project cash flow, however, creates significant revenue risk. Cancelled projects with negative cash flow leave you heavily exposed to revenue writeoffs. It’s hard to stop work on a client in financial distress if you’re upside down on cash position.

So you need dynamic reporting for cash flow performance, and insights into what projects are most at risk. Several AI use cases for me are:

• A user says “provide me a list of project with the most negative cash position, and for each project list the invoices most past due.”, and receives the resulting summary.
• A user says “provide a list of projects with unbilled services and expenses, with a brief description for each project of those services and expenses which are unbilled.”, and receives the results.
• A user says “tell me if the cash position for projects A, B, and C have improved or declined over the past 6 months.”, and receives a trend report or visual.
• A user asks “tell me which projects are consistently invoiced by the 5th business day of each month” and receives the resulting list.

A project finance AI that brings greater awareness and insight into project cash flows would possess a subtle but important differentiator compared to where AI’s seem to be focused for the time being (revenue and profit reporting).

----------------

I didn’t even get into balance sheet reconciliations and other areas that you often see performed by G/L accountants instead of project finance professionals. The sky’s the limit when it comes to the possibilities of AI, it’s just a question of who will actually deliver valuable functionality that improves profitability.

----------------

What’s been your experience with project finance AI’s thus far? What promises are these companies making to you, and what has actually come to fruition? What are the use cases most important to you? Let me know.

----------------

I specialize in pricing and financial strategies for service and technology providers. Contact me to discuss solutions for your organization.