How Holding Companies Can Automate Customer Conversations Across Their Brands

A group with several brands rarely decides how its brands talk to customers. It happens brand by brand. One brand hires an agency that sets up a basic chatbot, another answers WhatsApp from a manager's phone, a third bought a platform nobody fully configured. Each one pays for its own tools and its own integrations, and the group cannot see what any of them is hearing from customers.

The answer for most holding companies is to centralise the engine and the data, keep every brand distinct at the front, and roll out one brand at a time. One system behind all the brands avoids paying for the same integrations again and again, gives the group one view of its customers, and lets what works for one brand reach the rest. Each brand still keeps its own WhatsApp number, Instagram account and Facebook Page, its own voice, catalogue and rules. Letting each brand buy its own tool looks faster and costs more every year after.

Why this is a regional question

In the Gulf and the wider Middle East, the group is the normal shape of a business, not the exception. Strategy&, part of PwC, wrote in 2020 that family businesses constitute 60 percent of the GCC's non oil GDP and employ 80 percent of its workforce, citing PwC's Middle East Family Business Survey. Many of those businesses run several retail brands, franchises or distribution agencies under one owner.

That structure changes the automation question. A single brand asks "which tool should we buy?" A group has to ask who buys it, who runs it, what is shared, and what each brand keeps to itself. Get that wrong and you end up with several half working chatbots and no group view at all.

Three ways a group can buy it

We compare the three models on the criteria a group finance or operations lead would actually use. This is editorial judgment from plumcut, which sells the third option, so read the last column with that in mind.

Each brand buys its own is right when the brands genuinely share nothing: different customers, different systems, different markets. It is wrong the moment two brands share a warehouse, a courier, an ERP or a customer base, because every shared thing gets integrated twice.

One platform license run in house is the strongest option for a group with a real digital team. Several platforms already support multiple brands, such as respond.io with a workspace per brand or Zendesk with up to 300 brands on its Enterprise plans, both checked on 25 September 2026. The license is the easy part. The group still has to build the connections to stock, orders and couriers, write the rules for each brand and keep all of it working.

One managed system for the group is for the group that wants the outcome without building a team to produce it. The provider builds the engine once, connects it to the group's systems, sets up each brand's front door and runs it. The trade is dependence on that provider, which is why data ownership has to be written into the agreement.

What to centralise and what to leave to each brand

The rule of thumb: anything that costs money to build or is a risk to get wrong belongs to the group. Anything the customer can see belongs to the brand.

Groups rarely run one system across every brand. One brand runs an older ERP, another a newer one, a third runs its online store on Shopify. The automation layer should connect to each brand's system through its own connector, so that when a brand moves to a new ERP, one connector changes and the conversations carry on. If a provider's answer to an ERP migration is "we rebuild", that is a cost you will pay every time a brand upgrades.

The Meta setup decision most groups skip

Every WhatsApp number a group uses sits inside a business portfolio on Meta, and that choice has consequences most groups discover late. Checked against Meta's documentation on 25 September 2026:

• Numbers are capped. A new portfolio can register two business phone numbers, rising automatically to 20 once the business is verified or reaches a messaging limit of 2,000. A group with several brands needs verification early.
• Messaging limits are shared. Limits are set per portfolio and shared by every number in it. A new portfolio starts at 250 unique customers per rolling 24 hours for messages outside a customer service window, and higher tiers go up to unlimited. One brand's big campaign can use capacity another brand needed.
• Customer IDs follow the portfolio. WhatsApp's business scoped user IDs, in webhooks since early April 2026, are unique per portfolio and user. Brands in one portfolio see the same customer as the same ID. Brands in separate portfolios do not.

So one portfolio gives the group consistent customer identity and one verification, at the cost of shared capacity. Separate portfolios isolate each brand, at the cost of a joined up view. Brands that are separate legal entities, or that broadcast heavily, are good candidates for their own. Decide it deliberately, before the numbers are registered, because moving later is work.

Customer data across brands

The biggest prize in a group is also the easiest to misuse. A group can see that the customer who buys from one of its brands is exactly the customer another brand wants. That does not mean the second brand can message them.

WhatsApp's Business Messaging Policy requires opt in permission before a business contacts someone and requires every opt out to be honoured, and it says nothing to suggest that permission given to one business covers a related one. Our judgment: treat each brand's consent as belonging to that brand. Use the group view to understand customers, to decide where to invest and to spot what they keep asking for. Make cross brand offers only through a brand the customer already chose to hear from, or ask them first.

Data ownership matters as much as consent. Whatever the model, the group should own the conversations, the customer profiles and the insight built from them, and be able to take them if it changes provider. Put it in the contract.

The wholesale arm is a second customer

Many groups also distribute: they import brands and supply other retailers and supermarkets. Those buyers already order over WhatsApp, usually by messaging a sales rep. The same engine can serve them, with different rules.

Keeping these apart is non negotiable: a wholesale buyer should never see a retail promotion, and a retail customer should never be quoted a trade price. Sharing the language skills, the order connection and the reporting is what makes serving both affordable.

Roll out one brand at a time

A group rollout fails in two ways: everything at once, or a pilot too small to prove anything. The pattern that works:

1. Pick the brand with the most to gain. The most conversations and the clearest pain. A quiet brand produces too little data to judge the result, and a disappointing pilot on the wrong brand can stall the whole group.
2. Agree the measures before launch. Response time, share of conversations resolved without a person, orders placed in chat, handover rate. Write the current numbers down first. We explain how to judge whether a brand is ready in when a business is ready for AI customer service.
3. Cover a busy period. A pilot that only sees a quiet month has not been tested.
4. Copy the pattern, not the setup. The second brand reuses the engine, integrations and reporting, and only its front door, catalogue and rules are new. How that works across numbers, accounts and Pages is set out in how to run one AI across several stores.
5. Move the back office once the front works. Order handovers, stock checks between branches and courier updates are where the larger savings sit, as we cover in automation beyond customer service.

How plumcut works with a group

plumcut is the managed option in the table above. We build plum once for the group and give every brand its own front door on WhatsApp, Instagram and Messenger, in its own voice, with its own catalogue and rules. Behind them, plum connects to each brand's store, ERP, payments and couriers through their own connectors, reads Arabic, Arabizi and English, and feeds one dashboard with every conversation and customer profile, which the group owns. We run it and keep it working as brands change systems, as described in how it works.

Pricing follows the same shape as for any client: a setup fee for the build, then a fixed monthly fee set by what the solution does, with Meta messaging and AI usage billed at cost, as shown on pricing. For a group, that usually means scoping the first brand properly, then adding brands as a change in scope rather than a new project.

When we are not the right answer. If your group has a digital team that wants to own the build, license a multi brand platform and keep the difference. If your brands share no systems, customers or suppliers, letting each choose its own tool is reasonable. We fit the group that wants every brand answering and selling properly, and has nobody whose job it is to make that happen.

What to do this week

1. Count the doors. Every WhatsApp number, Instagram account and Facebook Page across the group, and who answers each.
2. List what the brands share. ERP, warehouse, couriers, payment provider, customers. Every shared item is an integration you should only pay for once.
3. Check the Meta setup. Which portfolios exist, who owns them, and whether they are verified.
4. Pick the first brand. The busiest one with the clearest problem, and write down its numbers today so you can measure the change.