Insights > Valuation Operations
Why Workflow Design Will Shape the Next Generation of Valuation Firms
One of the things I’ve noticed in conversations about AI lately is how much attention people are paying to the models themselves.
Which model is the most capable? Which one is the most reliable? Which platform should we be using?
Those are reasonable questions. And given how quickly the technology is advancing, I understand why people are focused on them.
But I think we’re missing a much bigger issue.
The model is only one part of the equation.
What matters just as much, and potentially more, is how we organize the work around it.
How is information collected and evaluated? How are assumptions reviewed? How is the reasoning behind a conclusion documented? And how do we maintain professional accountability when portions of the work are being performed by AI?
Those questions don’t get answered simply by adopting a more capable model.
They require us to reconsider how valuation work is actually performed.
And I think the firms that understand this distinction will be in a much stronger position as the technology continues to improve.
Making Existing Processes Faster Is Only the Beginning
For years, technology has helped valuation professionals perform their work more efficiently.
We’ve developed better spreadsheets, improved access to financial data, automated calculations, and introduced software to support various parts of the valuation process.
All of those improvements have been valuable.
But in most cases, we’ve been using technology to accelerate processes that already existed.
AI is creating an opportunity to do something different.
Rather than asking how we can perform each step faster, we can begin questioning whether every step needs to exist in its current form.
I think that’s a significant distinction.
Consider how a typical valuation engagement is organized.
We collect financial information, prepare analyses, construct models, review assumptions, develop conclusions, and produce a report.
Each stage involves a series of activities, often performed by different professionals with varying levels of experience.
Over time, firms have developed procedures to coordinate that work and maintain quality.
But those procedures were generally designed around the capabilities and limitations of the people and technology available at the time.
What happens when those limitations begin to change?
If AI can construct a sophisticated financial model, does it still make sense to organize the engagement around the time required to build that model manually?
If technology can identify potential financial normalization adjustments, should the initial analysis continue to be performed the same way?
And if systems can assist with drafting reports and evaluating supporting information, how should we structure the review process?
I don’t think the answer is simply to insert AI into every existing step.
The larger opportunity is to reconsider the steps themselves.
The Difference Between Automating a Task and Redesigning a Workflow
Let me give you an example.
Valuation professionals spend a considerable amount of time working with complex spreadsheets.
Those spreadsheets may contain hundreds or thousands of formulas, multiple interconnected schedules, and calculations that ultimately affect the conclusion of value.
Historically, we’ve developed review procedures to help identify errors.
An analyst may be asked to perform a mathematical review, checking formulas, tracing calculations, and confirming that the model works as intended.
Now suppose we introduce AI into that process.
One approach would be to use AI to perform the mathematical review more efficiently.
That’s useful. It could reduce the time required to check formulas and identify potential errors.
But I think we need to ask another question.
What if AI generated the model in the first place?
Would we still need the same review process? Would we need a different one? And could some of the controls be incorporated directly into the model-generation process rather than performed afterward?
I’m not suggesting we eliminate mathematical reviews or other quality controls. Right now, those controls remain essential.
But the point is that the workflow may need to change as the technology becomes more capable.
Instead of having one process for constructing a model and another for checking it, we may be able to design an approach where construction, validation, exception identification, and professional review are more closely connected.
That could fundamentally change how the work is performed.
And it’s a different way of thinking about AI than simply asking which existing tasks can be automated.
Professional Judgment Still Needs a Place in the Process
Of course, valuation work is more than a series of calculations.
It requires technical expertise, contextual understanding, professional judgment, and the ability to defend the conclusions.
Those responsibilities don’t disappear when AI performs portions of the analysis.
If anything, I think the design of the workflow becomes more important as technical execution becomes increasingly automated.
Consider a system that analyzes financial statements and identifies potential normalization adjustments.
It may recognize unusual expenses, compare historical results, and propose adjustments based on the information available.
But someone still needs to determine whether those adjustments are appropriate for the particular engagement.
What evidence supports them? What additional information is required? Are there circumstances the system hasn’t considered?
And ultimately, who is responsible for accepting or rejecting those adjustments?
A well-designed workflow needs to address those questions.
It should establish where professional review is required, how assumptions are evaluated, when exceptions should be escalated, and how decisions are documented.
Simply generating an answer isn’t enough.
For professional valuation work, we need to understand how the answer was developed and whether the underlying reasoning is supportable.
That’s especially important when the work may be reviewed by other professionals, regulators, courts, or clients who need to understand the basis for the conclusion.
I think this is where firms need to focus considerable attention.
The objective shouldn’t be to eliminate human involvement.
It should be to determine where professional involvement creates the greatest value and design the workflow accordingly.
Institutional Knowledge May Become More Valuable
There’s another aspect of workflow design that I think is particularly important.
For years, valuation firms have accumulated significant knowledge about how engagements should be performed.
That knowledge exists in templates, financial models, internal procedures, review checklists, training materials, and the experience of individual professionals.
But much of it isn’t necessarily organized in a way that can be consistently applied across the firm.
Some of it exists primarily in the heads of experienced professionals.
And when those people leave, retire, or move into different roles, firms can lose valuable knowledge about how the work gets done.
AI creates an interesting opportunity to address that problem.
But only if the underlying knowledge is sufficiently documented and structured.
In our own work, we’ve spent years documenting valuation use cases, including the inputs required, analytical steps, expected outputs, and review procedures.
That type of documentation becomes increasingly valuable as AI systems gain the ability to execute more complex processes.
Instead of relying entirely on individual professionals to remember how a particular analysis should be performed, firms can begin incorporating established methods and procedures into their operating systems.
That doesn’t mean every engagement becomes identical.
Valuation work will continue to involve unique circumstances, assumptions, and professional decisions.
But it does mean firms may be able to preserve and apply more of their institutional knowledge consistently.
And over time, that could become an important competitive advantage.
A firm that has documented its processes, established clear review procedures, and organized its knowledge will be better positioned to incorporate new capabilities than one that relies entirely on informal practices.
In that sense, the value may not come primarily from the AI model itself.
It may come from the firm’s ability to apply its accumulated expertise through a well-designed operating system.
Better Workflows Require More Than Better Technology
One of the challenges I see is that many firms are approaching AI adoption primarily as a software decision.
They evaluate different platforms, compare features, run demonstrations, and decide which tools to provide to their professionals.
There’s nothing wrong with that.
But purchasing technology is relatively straightforward compared with changing how an organization operates.
Redesigning workflows requires a detailed understanding of the work itself.
Firms need to identify which activities are repeatable, which decisions require professional judgment, what information must be available, and how outputs should be evaluated.
They also need to consider governance, confidentiality, data security, and quality control.
And perhaps most importantly, they need to determine how all of those elements fit together.
A powerful AI model can perform an impressive task in isolation.
But that doesn’t necessarily mean it can perform the task reliably within a professional valuation engagement.
The system needs appropriate inputs. It needs to operate within defined procedures. Its outputs need to be evaluated. And the people responsible for the engagement need to understand what happened throughout the process.
Those are workflow design problems.
They won’t be solved simply because the next generation of AI models becomes more capable.
In fact, as those models improve, I suspect these operational questions will become even more important.
The Operating Model Is Becoming a Strategic Decision
I think this also has implications for how valuation firms compete.
Historically, growth has often required additional professional capacity.
More engagements generally meant more analysts, more reviewers, and more people coordinating the work.
Technology helped improve productivity, but the relationship between headcount and production capacity remained fairly direct.
If AI can perform increasingly sophisticated portions of valuation work, that relationship may begin to change.
A firm with well-designed workflows, strong professional oversight, and access to quality data may be able to complete more work without expanding its team at the same rate.
But that outcome won’t necessarily come from simply providing everyone with access to AI.
It will depend on how effectively the firm integrates these capabilities into its operations.
Two firms could use the same underlying AI model and achieve very different results.
One might use it primarily to summarize documents and draft reports.
Another might redesign its engagement process around structured information collection, automated analytical execution, defined review procedures, and documented professional decisions.
Both firms could realize productivity improvements.
But the second firm may be creating a fundamentally different operating capability.
And I think that distinction will become increasingly important.
The model itself may not provide a lasting competitive advantage if competitors can access similar technology.
The way a firm organizes its processes, applies its expertise, and maintains professional accountability may be much harder to replicate.
What Should Valuation Firm Leaders Be Thinking About Now?
I don’t think every valuation firm needs to redesign its entire operating model immediately.
And I certainly don’t think we should assume that every new AI capability is ready for unrestricted use in professional work.
There are still important questions about reliability, consistency, confidentiality, and professional oversight.
But I do think firms should be taking a serious look at how their work is organized.
Where are we spending professional time today?
Which activities are primarily technical execution, and which require meaningful professional judgment?
How much of our institutional knowledge is documented and available for consistent use?
Are our review procedures designed around the way work has historically been performed, or the way it could be performed with emerging technology?
And if we were building our valuation practice today, with the capabilities that are becoming available, would we organize it the same way?
I think that last question is particularly important.
It’s easy to evaluate new technology through the lens of existing processes.
We’ve done things a certain way for years, so naturally we look for opportunities to make those same activities faster and more efficient.
But I think there’s a limit to that approach.
The more significant opportunity is to reconsider how the work should be performed in the first place.
That requires a different conversation.
It’s not just about which AI model a firm should use.
It’s about how information moves through the organization, how expertise is applied, how decisions are made, how work is reviewed, and how professional accountability is maintained.
Those are the elements that determine how effectively a valuation firm operates.
And as AI capabilities continue to advance, I think the firms that pay attention to those elements will be in a much stronger position.
The model matters. But the workflow matters more.
Better valuation operations begin with better questions.
Explore how ValueVision is building specialized agents, connected operating infrastructure, and practical advisory services for the future of valuation.