Choose Partners Who Can Defend and React
In the AI era, a technology partner must do more than ship software. It must carry the change, protect the old estate and the new one, and respond at machine speed when something goes wrong.
Software is becoming free. Change is not. Safety is not. Choose the partner who can deliver both.
1. Two Arguments, One Conclusion
In the first two articles of this series, I made two arguments. First, free software will not transform a business on its own. The real limits are risk, the cost of change and a lack of imagination. Second, every change a company makes with AI also opens new doors to attackers who now use AI themselves.
Put these together and one conclusion follows. The most important decision in a transformation is not which software to buy. It is which partner to trust with the change.
2. The Old Way of Choosing
For twenty years, enterprises chose technology vendors with a familiar list: features, price, references, the quality of the demo. Security was a questionnaire that procurement sent near the end. The vendor shipped the software, and the customer’s team owned everything after that.
That model does not survive the AI era. When software is nearly free, features and price stop separating one vendor from another. When attackers move at machine speed, a questionnaire at the end protects nobody. And when the hard part of a project is the change itself, a vendor that hands over the code and leaves has handed over the risk.
3. Five Questions to Ask a Partner
I suggest that every board ask five questions before it chooses a partner for AI transformation.
Will you own the change, not only the code? The partner should do the integration with the old systems, redesign the process and stay accountable for the result. A partner that is paid for outcomes has every reason to make the change succeed and to keep it safe.
Can you secure the estate I already have? Most attacks will not come through the shiny new AI system. They will come through the old ERP, the store devices, the factory controllers and the supplier links. A good partner maps these, closes the gaps and watches them, before it connects anything new.
Is security designed into the future estate from day one? AI agents need limits on what they can do, which data they can see and which actions need a human to approve them. Generated code must be tested before it runs. Every new integration needs an owner and a log. Ask to see how this is built, not only a policy document.
Do you understand how AI is used to attack? Prompt injection, AI-written phishing, deepfake voices, automated scans that find a flaw within hours of its release. A partner that has not studied these will not defend against them. Ask what they saw last quarter and what they changed because of it.
When something goes wrong, how fast do you react, and who answers the phone? Breaches will happen. What matters is detection in minutes, containment in hours and a clear plan for who does what. The partner must be able to switch off an agent, roll back a change and keep the business running while the problem is fixed.
4. Trust Is the Product
In my first article I wrote that, when software becomes abundant, value moves to what is still scarce. Trust is the scarcest of all. A customer will let an AI agent buy for them only if they trust it. A retailer will let a system change prices in a thousand stores only if it can see, approve and reverse each change. A board will approve a transformation only if it believes the company will be safer at the end, not more exposed.
This is why the partners who win this decade will be different from the vendors who won the last one. They will carry the change. They will defend both the estate you have and the estate you are building. They will react at machine speed when it matters.
Software is becoming free. Change is not. Safety is not. Choose the partner who can deliver both.
© 2026 Shopsense Retail Technologies Limited. All rights reserved.