Building an AI-Ready Services Business
A practical guide to moving beyond AI experimentation and building the operating foundation, workflows, and organizational behaviors required to create real business impact.
Most services firms are already experimenting with AI - they are piloting tools, building agents, automating pieces of existing workflows, and finding ways to produce work faster. In many cases, those experiments are working as teams are seeing research be completed more quickly, proposals take less time to draft, meetings are easier to summarize and distribute findings, and information that once took hours to compile can increasingly be produced in minutes.
Those are meaningful improvements, but the larger opportunity is to rethink how a services business operates: how information moves through the organization, how work gets done, how leaders understand what is happening in the business, and ultimately how decisions get made.
For a services business, those decisions are consequential to business performance and growth potential:
When should we hire, and for which capabilities?
Does our pipeline actually support the revenue plan?
How aggressively should we invest ahead of demand?
Where should we deploy scarce capacity?
Which clients, projects, and capabilities are actually generating the most value?
Where is margin being created, and where is it quietly being lost?
Which parts of the business require leadership attention now, rather than at the end of the month?
AI has the potential to help answer these questions faster and with considerably more evidence behind the answers. It can also automate much of the work required to gather that evidence in the first place, allowing teams to spend less time collecting, reconciling, and reporting information and more time interpreting what it means and deciding what to do about it.
But there is an important catch: AI cannot compensate for an operating model that the organization itself has not clearly defined.
If pipeline stages mean different things to different people, project data is inconsistently maintained, utilization is calculated three different ways, resource plans live in spreadsheets outside the PSA, or leadership decisions depend on knowledge that exists only in someone's head, adding AI does not eliminate those problems.
This is why becoming AI-ready is about much more than technology. It requires the business to establish the data, definitions, context, processes, management practices, and organizational behaviors that allow AI to operate reliably inside the company.
And increasingly, we believe this is one of the most important implications of AI for services firms: AI readiness and operating-model maturity are becoming inseparable.
Step 1: Start with the decisions, not the technology
One of the easiest mistakes to make with AI is to start with the tool. A new platform becomes available, someone sees an impressive demonstration, or a team identifies a process that could be automated, and the organization begins experimenting and eventually accumulates dozens of potential AI use cases without necessarily having a clear view of which ones matter.
A better starting point is to ask a different question:
What are the most important decisions we need to make better, faster, or more consistently?
For most services firms, the highest-value opportunities sit at the intersections between functions rather than neatly within them.
A hiring decision, for example, is not simply a human resources decision. It depends on pipeline quality, expected conversion, project timing, current utilization, available skills, delivery commitments, labor costs, margin targets, and the firm's willingness to hire ahead of demand.
Similarly, forecasting is not simply a finance exercise. A reliable forecast depends on the relationship between pipeline, backlog, project schedules, staffing, historical delivery patterns, and revenue recognition.
These are precisely the kinds of decisions where AI can eventually become incredibly valuable because it can continuously synthesize information across the business rather than requiring a person to manually assemble it every time a decision needs to be made.
Starting with these decisions also forces the organization to define what AI is actually supposed to accomplish. Instead of pursuing "AI adoption" as an objective, the business can focus on tangible outcomes: better forecasting, more effective capacity planning, improved project economics, faster management visibility, or more disciplined investment decisions.
That provides the foundation for everything that follows.
Step 2: Define how the business actually runs
Once the important decisions are clear, the next step is understanding the operating system behind them.
Every services firm has processes connecting commercial activity to delivery and ultimately to financial performance. Opportunities become projects, and projects require people. People have capacity and also cost, and delivery decisions affect utilization and project economics. Those economics ultimately determine margins, cash flow, and enterprise value. The challenge is that these processes are rarely as standardized as leadership assumes.
Different teams may estimate work differently. Sales and delivery may have different interpretations of when an opportunity becomes real enough to staff against, and Project Managers may maintain forecasts differently. Teams may use different assumptions about utilization, backlog, revenue, or project health.
Human organizations are remarkably good at compensating for this ambiguity, but AI does not inherently possess that institutional knowledge. Experienced leaders learn whom to ask, which spreadsheet to trust, which number needs adjustment, and where exceptions usually occur.
Before asking AI to automate or improve an operational process, the organization therefore needs to make the process itself explicit. How does an opportunity move from pipeline into delivery? When does the business begin planning capacity against it? How are projects estimated, staffed, and managed? How is revenue forecast? How are utilization and margins calculated? Who owns each decision, and when is it made?
This is where AI transformation begins to look less like a technology initiative and much more like operating-model transformation.
Because if the organization cannot clearly explain how the business is supposed to work, it will be very difficult to teach AI how to support it.
Step 3: Connect the data, and agree on what it means
Most services businesses already have most of the data they need. The problem is that it exists across systems designed to manage individual functions. For instance, CRM platforms understand pipeline, a PSA and project management platform understands delivery, staffing, and time, and a financial system tracks revenue, expenses, and profitability. Each provides an important part of the picture, but individually, none understands the business in it’s entirety.
The real value emerges when those systems can be connected, as pipeline can then inform capacity planning, and capacity constraints can influence hiring. Delivery performance can affect forecasts, and Staffing mix can be connected to profitability. With these connections, even financial results can be traced back to the commercial and operational decisions that produced them.
But simply moving all of this information into one place is not enough, rather the organization needs consistent definitions for what the data means.
What constitutes a qualified opportunity? What counts as backlog? How is utilization calculated? Which labor costs are included in project margin? How should probability-weighted pipeline be treated in a capacity model? What constitutes a healthy project? When should future demand begin influencing hiring decisions? These are some of the definitions that an organization needs to spend time agreeing to and documenting.
These questions sound mundane compared with the excitement surrounding generative AI and autonomous agents, but they determine whether an AI system can produce information that leadership actually trusts.
This is the role of the semantic layer: creating a consistent business language that sits across underlying systems and establishes how important metrics, entities, and relationships should be understood. Without it, AI may be able to access more data, but it still cannot reliably determine what that data means.
Step 4: Give AI the context to understand your business
Even clean, connected, consistently defined data is not enough - AI also needs context.
Modern models already understand concepts like utilization, gross margin, pipeline coverage, project profitability, capacity planning, and revenue forecasting. What they do not inherently understand is how those concepts should be interpreted inside your business.
For instance, a 72% utilization rate might represent a problem in one firm and a deliberate operating choice in another. Three times pipeline coverage might be sufficient for a business with highly predictable expansion revenue and completely inadequate for a firm pursuing mostly new logos. A decline in project margin might indicate poor delivery execution, or it might reflect an intentional investment in a strategically important account. This is all business context that the numbers alone cannot tell you, but are critical for implementing AI systems that have the opportunity to drive meaningful impact.
A context layer provides the additional information required to interpret them. It captures how systems relate to one another, how metrics should be interpreted, what goals the organization is pursuing, what thresholds matter, how leadership evaluates tradeoffs, and how decisions are actually made.
Consider a seemingly simple question from a COO: Should we hire now or wait? Without business context, AI can summarize current utilization and pipeline. With context, it can evaluate pipeline quality and historical conversion rates, expected project start dates, current capacity by role, planned attrition, staffing mix, margin targets, hiring lead times, and the organization's appetite for carrying bench ahead of demand.
This difference is significant - in the first scenario, AI retrieves information, and in the second, it helps leadership reason through a decision. That is where AI begins moving from a productivity tool toward an operating capability.
Step 5: Redesign the workflow, rather than simply automating the old one
Once the foundation exists, there is another temptation to avoid: using AI simply to make the current process faster. Some processes absolutely should become faster, but the more interesting question is whether they should continue to exist in their current form at all.
Consider the traditional monthly management reporting process. Someone exports information from multiple systems, reconciles it, updates spreadsheets, builds a presentation, distributes it, and then leadership meets to discuss what happened. AI can make several steps in that process faster, or the process itself can change.
If operational data is connected and consistently interpreted, the business can continuously monitor performance, identify material changes, surface emerging risks, explain likely causes, and bring those issues to leadership when they require attention. Instead of spending management time assembling and reviewing information, the organization can spend more of that time deciding what to do.
And this same principle applies throughout a services business: A resource manager should not have to manually compare a pipeline spreadsheet against future capacity every week; A finance leader should not have to spend days reconciling forecast changes before explaining them to the CEO; A project leader should not have to discover margin erosion after a monthly close if the underlying delivery signals were visible weeks earlier. The objective is not to remove human judgment from these processes, instead AI should remove more of the work surrounding judgment so people can spend more time exercising it.
That is a much more meaningful form of productivity than simply generating the same deliverable faster.
Step 6: Treat change management as part of the architecture
It is possible to build the data foundation, connect the systems, establish the semantic layer, introduce sophisticated AI capabilities, and still see very little impact. This is because organizations do not change simply because better technology becomes available.
Communications is as much of a component for change as the technologies and the processes themselves. People need to understand what is changing, why it is changing, how their roles will be different, which decisions AI should support, where human judgment remains essential, and why they should trust the information being put in front of them, and all of that requires deliberate change management.
The people who actually perform the work should be involved in defining how processes operate today, where they break down, which exceptions matter, and how a redesigned workflow should function. Leadership needs to establish clear expectations for how new tools and information will be used, and teams need the opportunity to test outputs, challenge assumptions, and help improve the system before they are expected to depend on it.
Importantly, management behaviors need to change alongside the technology implementation. For example, AI identifies capacity risk but staffing decisions continue to happen through informal conversations, little has changed within the organization to realize the benefits of the new capabilities. If teams are given new tools but are still measured, incentivized, and managed according to the old workflow, adoption will likely stall.
This is why change management should not be treated as the final phase of an AI initiative, but rather the training program that happens after the technology has been built. In practice, that means starting with specific workflows and decisions where the organization can demonstrate value, involving the people who use them in designing the new approach, creating feedback loops that improve both the technology and the process, and establishing new management practices that reinforce the desired behavior.
Over time, AI stops feeling like another tool employees have been asked to adopt and begins becoming part of how the organization operates.
What an AI-ready services business looks like
The goal is not to have AI everywhere, instead it is to build a business in which AI can be applied reliably to the places where it creates the most value.
In an AI-ready services organization:
Operational information is connected rather than fragmented
Important metrics have consistent definitions
Business context is explicit rather than trapped in individual experience
Management processes are designed around the decisions leadership needs to make
AI is embedded into workflows rather than isolated in another application
People understand how their roles and ways of working need to evolve alongside the technology
When these pieces are in place, the business becomes less dependent on manual processes and institutional knowledge, and leadership gets earlier visibility into risks and opportunities. Ultimately, the organization becomes more measurable, predictable, scalable, and adaptable, which are characteristics that matter not only to operating performance, but to enterprise value.
Perhaps most importantly, it creates the foundation for whatever comes next. Models will continue to improve, agents will become more capable, and new tools will emerge. The specific technology that firms are experimenting with today will almost certainly look different several years from now, but the organizations best positioned to take advantage of those advances will not necessarily be the ones that adopted the most AI tools first, rather they will be the ones that did the harder work of making their businesses ready for them.
Continue exploring
This guide builds on several ideas we have been exploring about what it takes to create meaningful AI impact inside a services business:
Strategy and AI must work together.
AI requires domain-specific business context.
A semantic layer creates a consistent language across the business.
AI transformation requires changes to workflows, management practices, and human behavior, not just technology.
Explore each in more detail:
Strategy and AI Working Together to Accelerate Operational Transformation
The Missing Link in AI Adoption: Why Your Agents Need a Domain-Specific Context Layer
Why Your AI Data Analyst Won’t Work Without a Semantic Layer
If you’re looking to make AI work in your services business, the starting point isn’t the tool. It’s the foundation behind it. That’s what determines whether AI becomes a useful experiment, or a real operating advantage.
At Form & Function, we help services firms rethink the operating model behind AI - connecting data, defining business context, redesigning workflows, and helping teams adopt new ways of working so that AI can move from experimentation to meaningful operational impact.