Guide · Choosing

Choosing a work messaging app.

Every comparison of work messaging apps reaches roughly the same conclusion, because for an office of desk-based staff the answer genuinely is Slack or Microsoft Teams and the rest is taste. This guide covers that quickly, then spends its time on the three things those comparisons leave out: what happens when the people who need to be in the conversation have no work account at all, what a free plan does to your message history, and the question an organisation has to answer if anyone might ever ask it to produce a conversation.

What is actually on the market

Team chat has consolidated. Nearly every organisation ends up with one of a short list, and the differences between the leaders are smaller than the comparisons suggest.

Microsoft Teams. Included with most Microsoft 365 business plans. Strongest where the organisation already runs Microsoft: identity, files, meetings and the messaging app are one system rather than four
Slack. The reference point for the category. Better at channels, search and third-party integration; a separate purchase and a separate identity to manage
Google Chat. Included with Google Workspace. Simpler than either, and sensible if your organisation already lives in Google
The independents. Products aimed at specific gaps: frontline and deskless teams, sectors with particular requirements, or organisations wanting to self-host

Where they genuinely differ is narrower than the marketing suggests, and it is worth knowing before a demonstration. Slack is better at channels: it assumes work is organised by topic, its search is the strongest in the category, and its directory of integrations is far the largest. Microsoft Teams is better at meetings and at files, because it is a front end to the rest of Microsoft 365 rather than a chat app with attachments, and for an organisation already on Microsoft 365 it costs nothing more. Google Chat is the simplest of the three and does least, which is a virtue if your team communication is mostly one-to-one and your documents are already in Google Workspace. Slack is the one people miss when they leave it; Teams is the one nobody has to be asked to install.

For general internal collaboration, the honest advice is to pick whichever team chat app matches the productivity suite you already pay for and move on. The integration you get for free by staying inside one vendor is worth more than any feature difference between them, and every one of these apps for business will let a team collaborate perfectly well.

That advice changes only when the messages themselves have to be producible later. That is a different requirement, and it is covered further down.

You probably already own one

The commonest mistake is buying a chat app that duplicates something already included in a licence the organisation pays for. If you have Microsoft 365, you have Teams. If you have Google Workspace, you have Google Chat. Neither is an upsell.

This matters beyond the licence cost. A second messaging app means a second directory of who works here, a second set of leavers to remember, a second place to search when someone asks a question, and a second supplier holding your data. Organisations that end up with both usually did not decide to; the second arrived with a department.

The exception is not a small one, and it is why the advice above is qualified rather than general. Teams and Slack both assume that everyone in the conversation has an account on your systems. A great many of the people organisations actually need to talk to do not.

Frontline and deskless staff are the first group: care workers, cleaners, caterers, drivers, teaching assistants, bank and agency staff, volunteers. Many have no work email address, and issuing a full productivity licence to each of them so they can be told that a shift has changed is an expensive way round a simple problem. The second group is larger still and is usually forgotten at this stage entirely: the people the organisation exists to serve. Residents and their families, tenants, clients, parents. Guest access in the mainstream products was built for a contractor joining a project, not for a daughter asking about her mother, and it requires an account either way.

Where that describes your situation, the deciding feature is what the product accepts as an identity. One that can treat a verified mobile number as the identity lets a bank carer, a volunteer or a family member take part without an account being created for them on your systems. That single design choice determines whether half your intended participants can join at all, and no feature comparison lists it.

Before you compare products, write down who has to be in the conversation, including the people outside your organisation. If a third of them have no work email address, that single fact will decide more than any feature list.

What to compare, once you have a shortlist

Assume the basics work. Every serious collaboration software product handles business communication the same way: group chats, direct messages, the ability to share files, search, mobile notification, voice and video calls, video conferencing and screen sharing, with the better ones adding task management and some workflow automation. Those are table stakes and they will not separate a shortlist. The differences that actually affect daily use are narrower.

How it handles notifications. The single largest driver of whether people keep it open. Look at how granular the controls are, and whether quiet hours are respected on mobile
Identity and joiners and leavers. Does it use the identity you already run, so that removing someone removes them everywhere?
Guests and external people. Almost every organisation needs to talk to someone outside it. Check what a guest can see, and what happens to that access when the project ends
Search. Whether you can actually find a decision from eight months ago, which is a question about retention as much as about search quality
Integration. Both with your productivity suite and with the handful of tools your organisation genuinely depends on. Do not weight a directory of hundreds of integrations you will never connect
Voice channels and presence. Useful for in-office teams who want an always-open room to drop into, and largely irrelevant to a distributed shift-based workforce
Where it stops. Team chat is not project management, and products that try to be both tend to be a weaker version of each

Weight all of that against your own use case rather than against a review's scoring. The best app for a software team in one office is rarely the best instant messaging app for a charity with sixty volunteers, and roundups of the best work chat apps are written for the first. Run the shortlist with a real team for a fortnight rather than in a demonstration. Chat products fail on habit rather than on features, and habit does not show up in a feature grid.

What a free plan does to your message history

This is the criterion buried in the pricing table, and it surprises organisations later.

Free plans are generally limited on message history, and the administrative features that matter most tend to sit behind an extra cost: older chat history becomes inaccessible after a period or a message count, depending on the product. Paid tiers usually advertise unlimited message history and unlimited users, and it is worth confirming that both apply to the specific plan rather than to the product. The messages are not necessarily deleted, but you cannot reach them, which for every practical purpose is the same thing. Organisations discover this at the moment they go looking for something old.

Two consequences follow. The first is legal rather than commercial: if the organisation might have to produce a conversation, whether for a subject access request, an employment matter or a regulator, a plan that hides history at ninety days has made that impossible while looking like a saving. The second is institutional memory: the reasoning behind a decision is usually in the thread, and a rolling window quietly deletes your own explanation of why you did things.

The mirror image is also a problem. A paid plan that keeps everything for ever has no retention policy, it has an accumulating liability, because personal data has to be kept no longer than necessary. What you want is a retention period you chose, applied automatically, with the ability to hold specific material when something is live.

Ask the supplier two questions: what is the default retention on the plan we are actually buying, and can we set it ourselves. A surprising number of answers to the second are no.

The question the comparisons never ask

Everything above is about workplace communication as a productivity problem. For a large minority of organisations there is a second requirement, and no comparison of the best work chat apps covers it: if someone asked, could we produce this conversation?

That question arrives from three directions. A regulator asking what was decided and when. A subject access request, where a person is entitled to the personal data in messages about them, with a one month clock. Or litigation and employment tribunals, where the other party usually has their own copy already.

The organisations that need to answer it are more numerous than the phrase regulated industries suggests: schools and trusts, charities, care providers, housing associations, professional bodies, and anyone whose staff discuss individuals in the course of their work. If your people talk about named children, residents, tenants or clients in a chat app, you are in this group whether or not you think of yourself as regulated.

For those organisations the criteria change. The record has to be made by the service rather than assembled from handsets. It has to survive the person leaving. It has to be searchable by the organisation, not only by the participants. Retention has to be yours. And where the record ends up matters as much as that it exists: a record held only in a supplier's cloud is one you have to ask a third party for at exactly the moment you are short of time.

The related question is transparency. If conversations are kept, everyone in them should be told plainly and at the outset. Organisations worry that staff will object; in practice people object to discovering it, not to being told.

A short way to decide

Three questions, in order, will settle most of it.

Who has to be in the conversation? If every one of them has a work account on a suite you already pay for, use what that suite includes and stop. If frontline staff, volunteers, service users or families are on the list, ask what the product accepts as an identity before you ask anything else
Might we ever have to produce one of these conversations? If no, buy on usability and price. If yes, the requirements in section 05 apply and most general chat products do not meet them
What happens to the history? Check the retention on the plan you are buying, not the plan on the marketing page

It is entirely reasonable to answer the second question with no. Most teams, most of the time, are coordinating work rather than making decisions about people, and a good instant messaging app bought on usability is the right purchase. The failure worth avoiding is answering no by not asking.