Problem First or Tool First: The Fork in the Road for Services Firms in 2026

Every firm in the world is trying to make decisions about AI right now.

It feels like progress. New tools. New workflows. Faster output. But underneath the activity is a choice that will matter a lot over the next few years.

I am seeing firms headed down one of two paths. Problem first or tool first.

The Tool First Path

The tool first path starts with capability.

• What can AI do?
• What can we automate?
• What can we launch?

This path feels productive for many because it produces visible output. AI copilots rolled out to delivery teams. Proposal generation automated. Internal agents built to summarize research, draft decks, or generate diagnostics. Leadership can point to adoption numbers and time saved.

In practice:

A marketing services firm uses AI to generate proposals in half the time. Proposal volume doubles. Win rate stays flat. Average deal size does not move. Pricing pressure increases because the work now looks easier to deliver.

A strategy consultancy builds an AI powered assessment. Clients like the output, but it is bundled into existing engagements at no additional fee. Delivery effort goes down. Revenue does not go up. Margins barely improve because scope is still loose and outcomes are still unclear.

An IT services firm automates research and documentation. Junior staff move faster. Senior staff still spend their time fixing work, managing scope creep, and stepping in to save accounts. Founder dependency remains unchanged.

AI scales whatever you already have. Clear systems or messy ones. Strong positioning or vague positioning.

If your ICP is fuzzy, your problem definition is weak, and your delivery model depends on tribal knowledge, AI will not fix that.

The Problem First Path

Problem first firms are starting somewhere else.

They do not ask how to use AI.
They ask what problem their best clients are already paying to solve.

• Who feels the problem?
• What does it cost them today?
• Why does it matter now?
• What happens if it stays unsolved?

One example.

A compliance advisory firm notices that their fastest growing accounts all share the same issue. Regulatory change cycles are accelerating, and internal teams cannot keep up with interpretation and execution. Misses are expensive. Delays are visible at the board level.

The firm designs a tightly scoped service around regulatory readiness, with a clear definition of done and a fixed time horizon. Only after selling it repeatedly do they introduce AI to speed up document analysis and consistency checks. AI reduces delivery cost. Pricing stays anchored to risk reduction, not hours saved.

Another example.

A RevOps consultancy sees that clients are not struggling with dashboards. They are struggling with decision latency. Forecasts arrive too late to act on. The firm reframes its offer around decision speed, not reporting. AI is used to surface anomalies and signals earlier, not to create prettier outputs. Clients pay more because the problem is framed in financial terms.

Problem first firms use AI to reinforce something the market already values. They do not ask AI to create demand.

The Problem First Path Applied Internally

A lot of firms right now are applying the “Problem First Path” to client problems, while applying the “Tool First Path” to internal problems.

• “We should use AI in sales.”
• “We need to automate delivery.”
• “Finance should be faster.”
• “Hiring should be more efficient.”

Problem first firms treat internal work the same way they treat client work. They start with the problem, not the tool.

• Where is the firm actually losing time, money, or trust today?
• Who feels the pain most?
• What does it cost us every month?
• Why does it matter now, not later?
• What breaks if this stays unsolved for another year?

Only after those answers are clear do they introduce AI, automation, or new systems.

Example: Sales Operations

Tool first motion looks like this.

A firm rolls out AI for outbound, proposal writing, and CRM hygiene. Activity goes up. Pipeline quality does not. Forecast accuracy is still weak. Senior leaders still do not trust the numbers.

Problem first motion needs to start elsewhere.

A real issue might be decision quality. Deals sit too long in pipeline. Qualification is inconsistent. Forecasts arrive too late to manage the quarter.

So the firm should define the problem as decision latency.

AI is then applied narrowly. Call transcripts are analyzed to surface deal risk. Qualification criteria are enforced automatically. Forecasts are stress tested weekly instead of quarterly.

Same tools. Very different outcome.

Example: Delivery Operations

Tool first motion focuses on output.

Automated research. Faster decks. More reusable assets. Delivery teams move quicker, but margins stay flat and scope creep continues.

Problem first motion starts with economics.

The problem here might not be speed but delivery variability. Projects are profitable on paper and unpredictable in practice. Senior leaders step in late to save work. Founder heroics remain baked into delivery.

So the firm should define this problem as margin leakage caused by inconsistent execution.

AI is then used to reinforce standards. Scope boundaries are flagged early. Past projects are analyzed to identify where margins eroded. Quality gates are enforced before work moves forward.

Again, same tools, different outcome.

Example: Finance and Forecasting

Tool first motion looks like automation for automation’s sake.

Close the books faster. Generate dashboards automatically. Push reports out more frequently.

Problem first firms need to name the actual issue. Financial visibility arrives too late to change behavior. By the time margins are clear, the work is already done.

AI is then applied to project level forecasting, not just month end close. Risk is flagged mid engagement. Pricing and staffing decisions are adjusted in flight, not after the fact.

The finance function moves from reporting history to influencing outcomes.

The Pattern

Internally, tool first firms ask, “Where can we use AI?”

Problem first firms ask, “Where are we bleeding?”

Time.
Margin.
Trust.
Energy.
Founder attention.

They define the problem precisely. They quantify the cost. They agree on what “fixed” actually means. Only then do they introduce tools.

AI can and will get noisy fast without defining and scoping a problem like it’s a project.  

The Internal Payoff

When problem first thinking is applied internally, firms get the same benefits they see externally.

• Clear priorities.
• Fewer initiatives.
• Higher leverage.
• Less founder dependency.
• Better operating margins.

Problem first is management posture.

Firms that get this right do not feel busier after adopting AI. They feel calmer. More predictable. More in control.

That is the signal you chose the right path.

AI Is an Accelerant

When AI is applied to a validated problem and a repeatable delivery model, AI compresses cycle time, improves consistency, and lowers cost to serve. Margins improve. Founder dependency decreases. Scale becomes possible.

When applied to unclear problems and undisciplined operations, AI creates motion. Lots of it. Very little real progress.

By 2026, every firm will claim to use AI.

What will matter is whether AI reinforces clarity or multiplies noise.

Where the Fork Shows Up Inside the Firm

You can spot the difference in each motion.

In sales, tool first firms use AI to generate more outreach and more proposals. Problem first firms use AI to qualify harder, surface real pain patterns, and disqualify faster.

In expansion sales, tool first firms chase AI detected signals they cannot monetize. Problem first firms define expansion paths first, then use AI to spot the right triggers.

In delivery, tool first firms automate output while outcomes stay flat. Problem first firms use AI to reinforce scope control, quality standards, and institutional knowledge.

In the back office, tool first firms adopt faster than governance can support. Problem first firms systemize workflows and controls before scaling usage.

Same tools. Different results.

The Divide Ahead

Over the next few years, the market will sort this out.

The winners will be the firms that stayed disciplined. They nailed a real client problem, proved willingness to pay, built a repeatable delivery model, and then used AI to scale something that already worked.

The Bottom Line

Problem first firms build leverage.
Tool first firms build activity.

Only one shows up in pipeline, margin, and valuation.