Data-Driven Pricing: How AI and FD Insight Transform Fees in Consulting and Agency Work
Most consulting and agency businesses believe they have a pricing strategy. What they actually have is a rate card that was built at a moment in time, updated inconsistently, and applied without enough visibility into what the work actually costs to deliver. The gap between what appears on the rate card and what each project ultimately returns is rarely discovered until the management accounts arrive, by which point the work has been done, the invoice has been sent, and the margin has gone.
This is not primarily a technology problem, though better tools can help. It is a data problem. Specifically, it is a problem of not connecting the information the business is already generating, timesheets, project records, write-off logs, invoice histories, to the decisions being made when the next proposal goes out. The data is there. The link to pricing is not.
AI tools are beginning to close that gap in useful ways, and this article covers where they add genuine value. But the more important shift is not technological. It is the decision to treat pricing as something that requires evidence, governance, and regular review, rather than something that was decided once and happens automatically ever since.
What you will learn
- Why scope creep is fundamentally a pricing problem, not a project management one, and how to address it at the right level
- What your time-recording data is telling you about pricing that nobody is currently reading
- Where AI tools add genuine value in surfacing underpriced work and patterns of margin erosion
- Why the psychology of discounting is a financial risk that most firms manage with instinct rather than policy
- How to build a pricing review into the FD’s regular cadence rather than leaving it to reaction
Scope Creep Is a Pricing Problem in Disguise
The conversation about scope creep in professional services and agency businesses is almost always framed as a project management or client relationship problem. The client keeps adding things. The account manager finds it difficult to say no. The brief was unclear. All of that is sometimes true. But underneath the project management conversation is a pricing conversation that is far more consequential and far less often had.
If a project regularly runs over its scoped estimate, in similar ways, across similar project types, the problem is not that this project was poorly managed. The problem is that the price for this type of work was wrong at the point of proposal. The scope was underestimated, the rate was too low, the contingency was insufficient, or some combination of all three. The project management problem is a symptom. The pricing problem is the cause.
The distinction matters enormously because the fix is different. A project management problem gets addressed through tighter scope documents, clearer change control, more assertive account management. These are useful disciplines. A pricing problem gets addressed by analysing what similar work actually costs to deliver, comparing that to what it is being charged, and adjusting the proposal process accordingly. The first set of fixes addresses the symptom. The second addresses the reason the symptom keeps recurring.
Most firms do both inadequately because nobody has connected the project data, what the work actually cost, to the pricing process for the next proposal. The estimator builds the quote based on their experience and a rough benchmark. The actual delivery cost sits in the timesheet system and is never systematically fed back. The cycle repeats.
What the Timesheets Are Telling You
Time-recording data is one of the most underused sources of pricing intelligence in professional services and agency businesses. It contains, in reasonably raw form, the information needed to answer the questions that pricing decisions depend on: how long does this type of work actually take, which activities routinely overrun their estimates, which teams or individuals deliver within budget and which do not, and what is the true cost of producing the outputs that clients are paying for.
The challenge is not collecting the data. Most firms with ten or more fee-earners have a time-recording system of some kind. The challenge is connecting it to anything useful. Timesheets are typically reviewed for billing purposes, to produce invoices and calculate WIP. They are almost never reviewed for pricing purposes, to understand whether the estimates used in proposals are accurate and whether the rates being charged are sufficient to cover actual delivery cost. Getting that connection right is fundamentally a question of processes and systems, not technology alone, and it is where an experienced FD can make a significant difference before any new tool is introduced.
A practical starting point, available to any firm with a decent time-recording system and someone willing to do the analysis, is a monthly reconciliation of actual delivery hours against estimated hours by project type. Not by individual project, which gets complicated and subjective quickly, but by category: a strategy engagement, a brand project, a litigation matter, a management consultancy retainer. What does this type of work actually cost to deliver, averaged across the last twelve months of similar engagements? How does that compare to what the proposals assumed?
The output of that analysis is usually instructive. There are typically two or three project types where the estimate is consistently wrong in the same direction, and the business has been absorbing the difference as a quiet tax on its margin for years without naming it. Naming it changes the conversation.
Where AI Tools Add Real Value
There is a version of this article that would breathlessly describe AI as the solution to pricing challenges in consulting and agency work. It would be wrong. The more honest version is that AI tools are useful in specific, bounded ways, and the businesses that benefit from them are the ones that have already done the foundational work: clean time-recording data, consistent project classification, a management accounting structure that connects revenue to delivery cost.
That said, within those boundaries, AI does add genuine value in two areas.
The first is pattern recognition at scale. An experienced FD reviewing monthly project data can spot the systematic underperformance in a particular project type. But doing that analysis manually across a portfolio of fifty active projects, month after month, is time-consuming. AI tools trained on project and timesheet data can surface those patterns automatically, flagging projects where actual hours are diverging from estimate early enough to trigger a review or a client conversation, rather than a write-off at the end.
The second is proposal benchmarking. Some more sophisticated project accounting platforms now include AI-assisted features that, given a new project brief and a set of parameters, can reference historical delivery data to suggest an estimate range. The estimate is only as good as the historical data feeding it, but used carefully it is significantly more rigorous than asking the account team to guess from memory. It makes the institutional knowledge accumulated in the timesheets accessible to the person writing the proposal.
What AI cannot do is replace the commercial judgement that decides how to use that information. It can tell you that projects of this type have historically taken 30% longer than estimated. It cannot tell you whether to hold the line on price, restructure the scope, or have a frank conversation with the client about what the work realistically costs. That judgement is the Finance Director’s domain, and it is where the real value concentrates.
AI surfaces the pattern. The FD applies the judgement. Neither is sufficient without the other.
The Psychology of Discounting
Most discounting decisions in professional services and agency work are made informally, in the moment, by the person closest to the client relationship. A proposal goes out. The client says it is a bit high. The account manager, who genuinely likes the client and wants the work, reduces the fee by ten percent. No analysis. No policy. No record of how often this happens or what it costs the business in aggregate.
The problem is not the individual discount. Sometimes discounting is commercially rational. The problem is the absence of a framework for deciding when it is and when it is not. Without that framework, discounting becomes the default response to any client price resistance, regardless of whether the resistance reflects genuine budget constraint, negotiating habit, or simply an expectation that the opening price is not the final one.
The data dimension matters here. Businesses that track discounts, by client, by project type, by the person who agreed them, typically discover that discounting is not evenly distributed. There are patterns. Certain clients get discounted almost every time. Certain team members discount more readily than others. Certain project types are discounted despite the fact that they routinely run over budget, meaning the business is simultaneously undercharging and overdelivering. None of this is visible without the data. And without the data, it cannot be managed.
A pricing policy does not need to be restrictive to be useful. It simply needs to set some parameters: the maximum discount that can be offered without sign-off, the information required before a discount is agreed, and a quarterly review of discounting patterns as part of the commercial reporting. That structure, embedded in the finance cadence, changes the quality of discounting decisions without eliminating commercial flexibility.
Building the Pricing Review Into the FD Cadence
Pricing in most firms is reviewed reactively. Someone notices the margin has dropped. A new business manager pushes back on the rate card because it is losing pitches. A key client announces a cost review. In each case the conversation happens under pressure, with incomplete information, and without the historical context that would make it productive.
The alternative is a pricing review that is built into the annual financial calendar as a matter of course. Not a lengthy strategic exercise, but a structured, evidence-based conversation that answers four questions: what did similar work cost to deliver in the last twelve months, what did we charge for it, what does the gap tell us about our current rate structure, and what are the market conditions that should inform the review?
The Finance Director owns the first two questions. The delivery and commercial teams own the third. The business development team owns the fourth. The conversation that results is qualitatively different from the reactive version, because it is grounded in data rather than anxiety, and because it happens at a time when the business can implement the findings rather than scrambling to respond to a crisis.
This is, ultimately, what data-driven pricing means in practice for a consulting or agency business. Not a software platform or an AI tool. A discipline: the decision to treat pricing as something the business thinks about systematically, with evidence, on a regular cadence, rather than something that was decided once and is now part of the furniture.
The Price Is Not Set in the Proposal
That is the shift in thinking that makes the difference. The price is set in the historical data: in the delivery hours, the write-offs, the time recorded against projects that went well and those that did not. The proposal is where the decision is communicated. If the information feeding that decision is incomplete or ignored, the proposal will keep underperforming in ways that feel like project problems but are actually pricing ones.
The businesses that close that gap are not the ones with the most sophisticated technology. They are the ones with a Finance Director who looks at pricing as a commercial question requiring the same rigour as any other financial decision, and who builds the review, the data, and the conversation into the regular rhythm of the business.
If your pricing feels more like inheritance than strategy, it might be time for a conversation. A short one can often reveal a great deal about where the margin has been going.
To find out how we can help your business scale its finance function, call today on:
+44 (0) 20 3848 1832
