Shared identity does not mean identical instructions
Give each sub-brand its own approved facts and communication rules, then add the principles it shares with the masterbrand. One universal tone prompt cannot resolve different audiences, offers and responsibilities. Start by deciding which brand owns the answer before asking AI to write it.
Design Week's 11 November Pets coverage describes connected sub-brands for insurance, a children's club and a foundation. HOW+HOW's primary case shows distinct positioning within a recognisable family. The AI instruction structure below is my adaptation of that design problem, not a claim about the agency's production process.
Build a shared layer and explicit brand records
The shared layer can define the organisation's naming, accessibility principles, factual standards and common visual assets. Each brand record then names its audience, offer, vocabulary, evidence sources, exclusions and responsible reviewer. Link to the live approved material instead of copying a large policy into several places that can drift apart.
For a fictional professional-services group, a research service could explain methodology in detail while a training brand uses simpler examples for beginners. Both can be clear and respectful. Neither should inherit the other's price, delivery promise or qualification because the content engine happens to use the same masterbrand logo.
Route ambiguous briefs before drafting
Require a brand identifier and content purpose in the brief. If the request mentions two services, ask which one the customer will actually buy or which brand will publish the page. An instruction such as use the group voice leaves the most important commercial distinction unresolved.
Define what happens when a brief cannot be routed confidently. A person can choose the destination or approve a genuine cross-brand page. Do not let the writing model silently invent a combined offer. The routing decision should remain visible so the final reviewer can check that the right facts and rules were loaded.
Test for facts leaking across the family
Create deliberately mixed examples: a training question containing a research package name, a group-wide enquiry about a service-specific guarantee, and a request to reuse a campaign across audiences. Check the selected record before judging the prose. A well-written answer can still come from the wrong business context.
In the fictional group, an output that offers a research subscription at a training-workshop price fails even if its tone is consistent. Retain the source record and the incorrect passage. Repair the routing or knowledge selection first; changing adjectives will not correct the underlying ownership error.
Make the brand family easy to maintain
Assign an owner to each record and make shared changes explicit. A new group-wide naming rule should reach every relevant brand, while a local service exception should remain local. Keep a small cross-brand review set and repeat it after changes to offers, policies or the routing logic.
JQ's AI Brief Generator case study and prompt engineering work connect brand decisions with usable instructions. Start with two related brands and one shared content task. The aim is a system that preserves useful differences while making the family understandable.
Link Map
Sources and editorial review
Assigned to the 10–16 November 2026 editorial series. This edition was prepared and source-checked on 16 November 2026; assigned dates do not imply prior public deployment. Proposed workflows and illustrative examples are JQ AI SYSTEMS editorial analysis, not completed client experiments.
- HOW+HOW: Pets sub-brand ecosystem (publication date not shown; checked 16 November 2026; primary agency project account).
- Pushkar Mankar, Design Week: HOW+HOW creates unified sub-brand system for Pets (11 November 2026; dated public project coverage).