Pricing is a delivery decision, not a proposal decision

Most consulting firms don’t lose margin because partners can’t calculate a day rate.

They lose it because a new project gets priced from a partial memory of the last one. A partner remembers the headline fee, the scope that was sold, and the client relationship. They don’t always remember the two unplanned workstreams, six rounds of stakeholder interviews, or the senior review time that turned a 45 percent gross margin project into a 22 percent one.

That gap gets expensive quickly.

For consulting and advisory firms in the $1 million to $25 million range, we usually see annual revenue leakage from underpricing, scope ambiguity, and unmeasured delivery effort land somewhere between $80K and $300K. That doesn’t mean every project is badly priced. It means a handful of projects each year can absorb far more senior time than the firm budgeted.

The problem gets worse as a firm grows. At 10 people, a founding partner may remember every engagement. At 35 people, the real commercial history is spread across CRM records, proposal folders, project plans, timesheets, invoices, slide decks, and the heads of practice leads.

Pricing then becomes a negotiation between intuition and urgency.

A client asks for a proposal by Friday. The commercial lead has 20 to 40 hours of pitch work ahead of them for a meaningful opportunity. They search for a similar deck, update slides, borrow a fee table, ask delivery leaders for a quick view, and try to get something out the door. The proposal may win. The delivery economics may still be wrong.

AI can improve this, but only if it is connected to your actual project history. A generic prompt that says “suggest a consulting fee” won’t know your delivery model, your team mix, or which client segments create expensive decision cycles. An AI pricing system needs to learn from completed work.

What makes a consulting project profitable

A project fee should reflect more than the number of weeks on a Gantt chart. The best firms already understand this, even if the calculation lives informally in a partner’s judgement.

The practical inputs fall into five areas.

Direct delivery cost

Start with the people who will do the work. This is not just their billed rate. It is their loaded delivery cost, expected hours, utilisation assumptions, and the level of senior intervention likely to be required.

A strategy project may be sold as 12 weeks with one partner, one manager, and two analysts. But historical data might show that projects in this category need 40 to 60 partner hours, not the 20 hours in the initial plan. That difference has a material impact on the fee or the team design.

Use actuals wherever possible:

• Budgeted hours versus submitted hours by role
• Internal cost rates by role and location
• Contractor and specialist costs
• Travel, research subscriptions, data purchases, and workshops
• Unbilled rework after client feedback
• Write-offs and delayed approvals

If your time data is imperfect, don’t wait for perfection. Most firms have enough invoice, staffing, and project-management evidence to identify which types of work are routinely under-scoped.

Complexity factors

Two projects with the same broad deliverable can have radically different economics.

A market assessment for a founder-led business with one decision-maker is not the same as a market assessment for a global company with five business units, an internal strategy team, and legal review in three countries.

Complexity often shows up in places that don’t appear in a standard scope statement:

• Number of stakeholder groups
• Number of interviews, workshops, or locations
• Quality and availability of client data
• Integration with client systems or internal teams
• Regulatory exposure
• Need for primary research
• Speed of decision-making
• Required executive review cycles
• Dependence on third parties

Your pricing model needs a way to score those factors. It doesn’t need to pretend they are perfectly scientific. A simple low, medium, and high complexity classification can improve consistency if it is grounded in completed projects.

Client segment and buying behaviour

Client segment affects both the work and the commercial process.

Enterprise clients may carry higher fees, but they may also require procurement negotiation, security reviews, extra reporting, and more senior stakeholder management. Private equity clients can move quickly, but may expect compressed timelines and rapid access to senior people. Mid-market owner-led businesses may value speed and directness, yet need more support translating recommendations into action.

The right question isn’t, “What can this client afford?”

The better question is, “What has it cost us to deliver successfully for clients with this operating model, decision structure, and expectation of access?”

This distinction protects margin without treating pricing as an arbitrary mark-up exercise.

Scope risk

A profitable price needs clear boundaries. Many consulting firms list deliverables but fail to define the work required to produce them.

“Develop a growth strategy” sounds clear until the client expects 30 executive interviews, competitor research in four markets, a board workshop, a financial model, and three iterations of the final deck.

A pricing model should flag scope language associated with past overruns. Phrases like “as needed,” “ongoing support,” “full stakeholder alignment,” and “implementation support” are not necessarily wrong. They do need a commercial decision behind them.

Target margin

The final input is the margin you are deliberately protecting.

A firm might target 45 to 55 percent gross margin on a conventional advisory engagement. A complex transformation project with a higher delivery risk may need more buffer. A small diagnostic engagement could have a lower margin if it is a deliberate entry point to a larger programme.

The key is to make that trade-off explicit. A partner can choose to price a strategic account differently. They should be able to see the cost of that choice before the contract is signed.

Why historical data beats partner memory

Partner judgement matters. It is often the reason a consulting firm exists in the first place.

But judgement improves when it has evidence behind it.

Historical project data reveals patterns that are hard to see deal by deal. You may discover that public-sector work is profitable only when discovery is separately funded. You may find that projects involving more than 15 interviews consistently consume manager time. You may learn that proposals with low initial fees generate a disproportionate share of scope negotiations later.

AI is useful here because it can read and organise a large body of unstructured material. It can connect proposal language with delivery records and then identify recurring signals.

The data typically sits in places such as:

• CRM opportunity notes and account records
• Past proposals, statements of work, and change requests
• Timesheets and staffing plans
• Project budgets and actual cost reports
• Invoices and payment records
• Kickoff notes, workshop outputs, and final decks
• Post-project reviews, if the firm runs them
• Meeting transcripts and internal delivery channels

A good system does not simply average historic fees. That would repeat past mistakes at scale.

Instead, it compares projects based on the factors that drive effort. It might identify 18 completed projects similar to a new opportunity, adjust for client segment and complexity, then recommend a fee range, role mix, and scope assumptions. The commercial lead gets the supporting evidence, not a black-box answer.

That is the practical role of Omni ops. It turns repeated operational work into structured agent workflows, while keeping commercial judgement with the people accountable for the deal.

What an AI pricing agent does end to end

An AI agent for profitable pricing should work before the proposal is written, not after the fee has already been agreed in principle.

Here is what the workflow looks like.

1. It reads the opportunity brief

The agent starts from the CRM record, discovery-call notes, email thread, or a short intake form. It extracts the basics:

• Client industry, size, and segment
• Problem to solve and expected outcomes
• Proposed deliverables
• Timeline and decision deadline
• Stakeholders and likely workshop needs
• Data availability
• Geography and travel requirements
• Known procurement or compliance requirements

Where information is missing, it flags questions for the commercial lead. It doesn’t invent assumptions and bury them in a fee number.

2. It finds comparable work

The agent searches prior projects using a combination of structured fields and document content. This is where basic folder search falls short.

It should recognise that a previous operating-model engagement may be comparable to a new transformation diagnostic even if the project titles are different. It should also distinguish a $100K engagement that involved four weeks of analysis from a $100K engagement that included six months of change support.

For each comparison, the agent pulls the original scope, sold fee, expected staffing, actual effort, delivery cost, margin outcome, and any scope issues.

The Knowledge Agent supports this work by reading the decks, documents, and meeting transcripts your firm produces. Instead of asking three partners, “Have we done this before?”, the team can ask questions across the firm’s project corpus and retrieve the relevant evidence.

3. It scores complexity and delivery risk

The agent applies the complexity factors that your leadership team has agreed matter. It may assign a score or use descriptive labels, but it should show why.

For example, it could identify that a proposed project has:

• A compressed six-week timeline
• Three business units that need alignment
• Client data with uncertain quality
• A board-level decision point
• Ten stakeholder interviews
• Two on-site workshops

Those conditions may suggest a higher partner and manager allocation than the initial outline assumes. The agent should make the risk visible before it becomes an unplanned staffing request.

4. It creates a pricing recommendation

The output is a range, not one falsely precise number.

A useful recommendation includes:

• Recommended fee range
• Target gross margin and likely margin at each fee point
• Suggested team structure by role
• Estimated hours and direct delivery cost
• Comparable projects and what happened on them
• Complexity drivers
• Scope exclusions and change-control triggers
• Questions that must be resolved before final pricing

For a project where the evidence supports $145K to $170K, the agent might recommend anchoring at $165K, with a lower boundary of $150K if the scope is reduced. It can also show that the client’s requested $125K would be acceptable only if interview volume, workshop count, or implementation support are removed.

That is a much better commercial conversation than simply saying, “Our standard rate is higher.”

5. It drafts the proposal around the economics

The fee and the proposal have to agree. If the scope changes in the drafting process, the pricing needs to update.

The Proposal Generation Agent pulls relevant past proposals, case studies, approved positioning, and historical pricing into a tailored first draft. It gives your senior people a strong starting point, rather than asking them to recreate a 30-page pitch from scratch.

This can reduce the 20 to 40 hours that major proposals often consume. More importantly, it makes the scope, staffing model, commercial assumptions, and fee structure consistent.

The partner still reviews the narrative. The delivery lead still challenges the staffing. The agent handles retrieval, comparison, drafting, and first-pass analysis.

Build the data foundation before chasing accuracy

You don’t need a data warehouse to begin. You do need an agreed minimum set of fields.

For every completed project, aim to capture the sold fee, client segment, service line, duration, team mix, planned hours, actual hours, direct costs, scope changes, and outcome. Add a short project review with two questions:

1. What did we underestimate?
2. What would we price or scope differently next time?

That last step is often where the best commercial intelligence sits.

The Research Agent can also reduce the repeated research burden that affects project economics. It runs structured company and industry research at the start of an engagement, producing sources, summaries, and a one-page brief. That means teams don’t spend the first two weeks repeating desk research that another team completed for a similar client six months earlier.

You can find more operating ideas through our guides and practical resources, but don’t treat this as a content exercise. Start with one service line and one repeatable project type. A firm that does commercial due diligence, for example, can build a pricing model around target sector, company size, data availability, timeline, and interview requirements.

Then validate recommendations against partner judgement for the first 10 to 20 opportunities. Track what was recommended, what was sold, and what the project ultimately cost to deliver. That closes the learning loop.

The commercial controls that stop margin erosion

AI recommendations are only useful if your operating model gives them a place to land.

Put a few controls around pricing:

• Require a margin view before proposals above an agreed threshold go out
• Name one partner accountable for approving exceptions
• Use a scope-risk checklist in every proposal review
• Include change-control language tied to interview volume, workshops, locations, and revision rounds
• Review actual versus estimated hours at project midpoint, not just after completion
• Feed project lessons back into the model monthly

These controls should not slow down sales. They should prevent last-minute pricing debates and delivery surprises.

One trades-business owner in our network described the equivalent problem clearly. His team knew which jobs made money, but only after the crew had finished them. Consulting firms face the same issue when they rely on retrospective margin reports. By then, the proposal is signed and the team is committed.

If you want to map the exact decision points, data sources, and first agent workflow for your firm, Book a 60-min Omni Audit. It is a working session, not a presentation. You leave with three outputs: the highest-value use case, the workflow design, and a practical path to deploy it.

Start with a focused first agent

Don’t try to automate every proposal or every delivery process at once.

Choose a project type with enough history and a real margin problem. This might be a market-entry assessment, operating-model review, diligence engagement, or strategic planning programme. Aim for 15 to 30 prior projects. That is often enough to test whether the patterns are useful.

Your first agent should answer a narrow commercial question:

Based on comparable projects, what should we charge, what team should we staff, and what conditions need to be in the scope?

That agent can produce an internal pricing brief before proposal writing begins. Once the workflow is trusted, connect it to proposal drafting, project intake, and delivery reviews.

If you need a practical way to define that first workflow, download the Deploy Your First Business Agent worksheet. You can also access the direct asset here: Deploy Your First Business Agent.

The point is not to replace the partner who understands the client. It is to give that partner better evidence, faster. Your firm has already paid to learn what makes projects profitable. The opportunity is to stop leaving that knowledge in old project folders and people’s memory.

For a view of how this applies across commercial work, delivery, research, and firm knowledge, see Omni for consulting firms. We can identify where your proposal process, project data, and delivery records can support better pricing without forcing a major system replacement.

The firms that protect margin well are not always the ones with the highest rates. They are the ones that know the cost of their work, recognise complexity before they sell it, and hold the line on the parts of scope that create unplanned effort.

If that is a priority for your next quarter, Book a 60-min Omni Audit. You can also review the AI audit for consulting firms before the call.