When AI threatens to reshape your market, you have three positions: build the capability yourself, partner with those who build it, or wait and follow. Each can be right. Boards should choose deliberately, based on where the value will sit once the technology matures, and keep the option to change position as evidence arrives.
In a recent interview, Uber CEO Dara Khosrowshahi discussed how one industry is facing autonomy. The strategic question behind it applies in every sector: when a technology could change how your market works, do you build, partner or wait? This article gives the AIPEX view.
Why is this a strategic choice rather than a technology choice?
Because it decides where your business will sit in the market once AI has changed it. In most sectors, AI shifts value between three places: the capability itself, the customer relationship and the data or assets that make the service work. A business that owns the customer relationship may not need to own the capability. A business whose advantage is its process or data may need to build, because renting the capability would hand that advantage to a supplier. We look at what must stay inside the business in Can You Outsource Your AI Advantage?
The mistake we see most often is making the choice by default: buying whatever vendors offer, or launching an internal build because it is exciting, without asking where the value will end up. A useful starting point is a board test of how exposed your business model is to AI.
What are the three positions, and when does each fit?
| Position | Fits when | Main risk |
|---|---|---|
| Build | The capability is central to how you win, you hold unique data or process knowledge, and you can fund it for years | High cost and slow progress while specialists move faster |
| Partner | You own the customer relationship, several capable providers exist, and you can integrate them into your service | Dependency on partners who may later compete for your customers |
| Wait and follow | The technology is unproven in your market, the cost of being late is low, and you can adopt quickly when it matures | Missing the window, and finding customers have moved |
Many businesses will hold different positions in different parts of the business: building where they differentiate, partnering where they aggregate, and waiting where the technology is not ready. For software specifically, see Will AI Replace Your Software Spend? Build, Buy or Wait.
How does a partner strategy protect the business?
Partnering is often the fastest way to offer customers a new capability without carrying the full cost of building it. The protection comes from owning what the partner needs: the customer, the demand, the operating know-how or the trust. Working with several partners, rather than one, keeps negotiating power and reduces the risk that a single supplier decides your future.
The board should also watch for the moment a partner starts to reach your customers directly. A good partnership has clear terms on data, customer ownership and exit, agreed while the business still has leverage.
When is waiting the right answer?
Waiting is legitimate when it is a decision, not drift. It fits when the technology has not yet proved itself in your conditions, when customers are not yet asking, and when you can adopt quickly once it matures. But a waiting position needs three things: a named executive watching the market, clear triggers that would change the decision, and enough internal capability to move when the trigger fires. Without those, waiting becomes falling behind. The board should record a waiting decision as clearly as a decision to invest, with a date to look at it again and the evidence that would change it.
Not every AI initiative should aim to replace people or processes outright. Our article on automation or augmentation explains why treating every project the same way weakens strategy.
Whichever position the board chooses, it should be revisited on a fixed cycle, at least annually and whenever a trigger fires. Technology that was unproven last year may be ready this year, and a partner that looked safe may have started to compete. The cost of changing position rises the longer a choice is left unexamined, so the review should be routine rather than prompted by a crisis.
The review should look at evidence, not opinion: how customers are behaving, what competitors and new entrants have launched, what the chosen position has delivered against its measures, and what it would now cost to move. A board that asks for this evidence each year will rarely be surprised by the market, and will be able to change course while the options are still open and affordable.
What should the board ask management?
- If AI changed how our market works within five years, where would the value sit: the capability, the customer or the data?
- Which parts of our business must we build, which can we partner for, and which can wait?
- Where we partner, do we keep the customer relationship and the data, and what are the exit terms?
- Where we wait, who is watching, and what trigger would change our position?
- Which competitor or new entrant would benefit most if we chose wrongly?
Where to start
Start by mapping, for each major line of business, where value will sit once AI has changed it, then choose a position for each. Our briefing on how AI creates competitive advantage and the seven questions every board should ask gives the framework, with a scorecard. The step on the AIPEX roadmap that fits here is Step 3, Plan: the Executive AI Strategy Day (£1,499 + VAT) turns those choices into agreed priorities, named owners and a 90-day plan in one day.
Frequently asked questions
Should a business build or partner for AI capability?
Build where the capability is central to how you win and you hold unique data or process knowledge. Partner where you own the customer relationship and several capable providers exist. Many businesses will do both, in different parts of the business.
Is waiting a legitimate AI strategy?
Yes, when it is a decision rather than drift. A waiting position needs a named executive watching the market, clear triggers that would change the decision, and enough internal capability to move quickly once the technology proves itself in your conditions.
How do you protect the business in an AI partnership?
Own what the partner needs: the customer, the demand, the operating know-how or the trust. Work with more than one partner where possible, and agree terms on data, customer ownership and exit while you still have leverage. See automation or augmentation →