Contents · 6 min read
01 / 05 · The constraint
What is holding the business back?
In short
Name the constraint first. Do this before you compare firms.
A vertical software company is preparing to seek bids. Several types of firms could help, but each starts with a different need. Its leaders have not yet decided what AI changes at the core of the business. That makes strategic judgment the first constraint, not the build.
The firm must decide what AI changes about the business before it chooses technology. That is the implication of its vertical software position.
Use four checks before you call a firm.
Position
Can leaders state how AI changes competition?
Specification
Does the firm know exactly what to build?
Fluency
Can the leaders judge the choices they face?
Delivery
A strong business owner is in place. Does the firm have the tech skill to build?
The first meeting should reveal that starting point.
For this software firm, the first need is strategic judgment. This is a choice for the top team.
02 / 05 · Strategic judgment
Has the leadership team decided how AI changes competition?
In short
If the answer is no, buy help with strategy first. The work must help the team choose a clear stance. It must also build skill with AI. The outside firm may guide the choice. It cannot own the choice.
The software firm is not ready to pick a system. Its leaders have yet to agree on what AI changes at the core. So they cannot write a sound brief for a build.
A strategy firm helps the leadership team decide what AI changes about the business and how the company will compete. It then helps the team choose what to build. A tool plan does not supply the missing market position.
The top team must take part. If it hands the task to IT, the work can turn into tool use. It can also focus on speed and cost. Those gains may help. Yet they do not set the firm's place in the market.
The CEO need not learn how to code. The team must be fluent enough to set direction.
Fluency does not make the key choice. The team must still decide which need AI should serve.

Good help should leave two gains behind. First, the leaders can state their stance in plain words. Second, they can choose what to build first.
That is where a real AI strategy starts. Once the team can state its stance, it can decide what to build.
03 / 05 · The specification
Can the company say exactly what to build?
In short
A clear stance changes the need. The team can now write a clear brief. The main constraint is the build itself. Tech skill now matters most.
The software firm has made its choice. Its leaders can say what the system must do. They can also say why the firm needs it.
Now the firm can buy tech help. An AI agency can build a specific system. A data scientist can build a specific model. A tech vendor can give the firm a platform.
Those roles differ from deciding what AI changes about the business.
A tech bid can now solve the right task. The builder does not need to choose the firm's place in the market. The top team has made that choice.
A clear brief sets the result. The company still needs someone accountable for directing the build toward it.
04 / 05 · The owner
Who should direct the build?
In short
Put a strong business owner in charge. Use the firm's tech team if it can do the work. Add a builder when a key tech skill is not there.
At the software firm, a product lead can own the build. That person knows which choices hold the most value. The tech team can then work from a business choice. It does not start from a tool request.
Some firms need no outside help at this stage. They have a strong business owner. They also have the tech skill to build.
Other firms lack part of that tech skill. A builder can fill the gap. Both paths can be right.
Outside help may be useful. It may even be needed. Yet it must not replace the owner inside the firm. The business owner must still point the technology at the result that counts.

Keep AI skill in the firm. Keep the key business choice there too. The firm should also own what gets built.
These choices let the software firm stay in charge. That remains true after the builder leaves.
05 / 05 · The scale
Does the partner match the work?
In short
Choose a large consultancy when global scale or regulatory program management is the main constraint. Otherwise choose the narrowest partner that fits the work and leaves the decision with your company.
If the software firm needs a program run at global scale, a large consultancy fits that constraint. The same is true when regulatory program management is the main job.
Most needs are narrower. If the company still needs a market position, the work must start with the business. If it has a clear build brief but lacks tech skill, the work can start with the system. If its own business owner and tech team can deliver the brief, it needs no outside firm.
Then test how the partner works. Ask who will do the work. Ask who will own what gets built. Ask whether the company can stop after the first step without penalty.
An honest comparison of the alternatives should make those differences clear.

Name the constraint
The CEO writes one line on what prevents the work from moving forward.
Define the required result
The team states its market stance or its build brief.
Invite the matching category
The firm asks only the right type of partner to bid.
Test the partner
The CEO uses the five questions for choosing an AI consultancy. The team writes down each answer before it weighs the bids.
The next step is to test the short list before you sign.
Test the fit
Your next task is to test whether the partner category fits the constraint holding your business back before you sign. A 30-minute conversation with GRAIL lets you talk through that constraint and the kind of help it calls for, while keeping the choice with your leadership team. If GRAIL is not the right fit, it will say so and point you elsewhere.
Choose a 30-minute slot. If we are not the right fit, we say so and point you elsewhere.
Questions CEOs ask about this
Do we need a strategy firm, Big Four consultancy, AI implementer, boutique, training provider, or no external help at all?
Choose AI help based on the constraint holding your business back. Decide if leaders need a market position or a build brief, since tech help too soon can solve the wrong need. If both are clear, your own team may be enough. Keep the key choice inside the company. The choice is what help the company needs first.
Which type of partner fits when our main problem is choosing the right AI bets?
Choose help that starts with what AI changes in your business. Your top team should end with a clear stance. It should also know what to build.
When does technical implementation capability matter most?
It matters once the firm knows just what to build. Before then, strong tech skill may give you a good build for the wrong need.
Can one provider credibly cover the full path?
A broad service list cannot answer that. Check where the provider starts. Then ask who does the work and who owns what gets built. Ask whether you can stop after each step.
When is doing it ourselves the better choice?
Do it in-house when your stance is clear. A strong business owner must lead. Your tech team must also be able to build from the brief.
Related reading
Sources
- AI in Vertical Software: why the work starts with the business and leaves the skill with the firm. (accessed October 8, 2026)
- How to Choose an AI Consultancy: how firm types differ and what to ask before you sign. (accessed October 8, 2026)
- AI for Leadership Teams: why the top team needs enough skill to own the choice. (accessed October 8, 2026)
- How to Build an AI Strategy: why strategy starts with what changes at the core of the firm. (accessed October 8, 2026)
- What Is AI Consulting?: when to buy help with strategy and when to buy help with a build. (accessed October 8, 2026)
- Copilot Is Not an Agent: when a business owner can lead the work with no outside help. (accessed October 8, 2026)
- GRAIL vs the Alternatives: when each main type of AI partner fits. (accessed October 8, 2026)
Next scheduled review: January 20, 2027
