Seven Questions for Your IT Partner – Ask Them Before You Sign

Most IT partnerships do not fail on technology. They fail on expectations nobody ever put into words — and on questions nobody asked while the mood was still good.

In a sales meeting, everyone agrees. The provider is competent, likeable, and has a solution for every problem. That is precisely the problem: a conversation in which nothing hurts tells you nothing about how things go when they go wrong. You only find out how a partner really works once something breaks — and by then the contract is long signed.

The following seven questions are deliberately uncomfortable. We are not listing them because we answer them particularly elegantly, but because they surface the information that never appears in a polished proposal. Ask them of every provider — including us.

1. "Who owns what at the end of this?"

When a project finishes, things exist that did not exist before: source code, configurations, documentation, automations, a Microsoft tenant, credentials. Who owns all of that sounds like a trivial question — right up until you want to change providers.

Access rights are the delicate part. It happens more often than you would think that a company holds no administrator rights to its own tenant, because the service provider set it up in their own name years ago. Technically that is quickly fixed. In negotiating terms, it is not.

• Good answer: The provider names precisely what belongs to you — code, data, configuration, licences, tenant — and what they keep as their own tooling. And it is in the contract, not just in the conversation.
• Warning sign: "We do not need to worry about that right now." Or: the tenant, domain, or licences are registered to the provider rather than to your company.

2. "How do we get out again?"

Raising the exit question in a first meeting feels like discussing divorce before the engagement. That is exactly why you should: it is the only moment at which you still hold negotiating power.

What matters is not whether an exit is possible — it always is — but what it costs, how long it takes, and in what format you get your data back. An export nobody but the outgoing provider can read is not an exit.

• Good answer: Clear notice periods, a described offboarding process, data export in common formats, handover documentation — and a price for it that is fixed in advance.
• Warning sign: Evasion, a wounded reaction ("You are already thinking about leaving?"), or exit costs that are only quantified once you terminate.

3. "Who is actually going to be at the keyboard?"

In the sales meeting you are often sitting across from the sharpest mind in the company. The question is whether that mind will also deliver your project — or whether, after signing, a team you have never met takes over.

This is not an accusation, it is normal operations. No provider puts their most experienced people on every task, and they should not have to. You simply need to know how the work is distributed: names, roles, the share handled by subcontractors — and what happens if the one person who knows your system leaves.

• Good answer: Specific names and roles, an honest account of subcontractors, and at least two people who know your setup.
• Warning sign: "We have a large team" — without a single name. Or: all knowledge of your environment sits with one individual.

4. "Do you earn money on what you are recommending to us?"

This is the most awkward question on the list, and the most important. Many IT providers earn on licences they resell, on hardware, or through vendor partner programmes. That is entirely legitimate — as long as you know about it.

It only becomes a problem when a recommendation stops being driven by your requirements and starts being driven by margin. Anyone recommending a solution they earn on should say so unprompted, before you have to ask.

• Good answer: Disclosure. "Yes, we are a Microsoft partner and we earn on the licences" is a good answer. Better still: an explanation of why the recommendation would be the same without that revenue.
• Warning sign: A proposal in which everything comes from a single vendor, with no explanation of why that vendor in particular.

5. "What would you advise us against?"

A partner who answers yes to every idea is not an advisor, they are a supplier. Advice means someone contradicts you even when the contradiction shrinks the size of the engagement.

This question tests two things at once: whether the provider has understood your business — without that understanding they cannot advise against anything — and whether they are willing to trade short-term revenue for long-term credibility.

• Good answer: A specific "I would leave that alone," with reasoning. Ideally about something they could have sold you without difficulty.
• Warning sign: "Anything is possible." Which is even true. It just does not answer the question.

6. "Tell me about a project that went wrong."

Every provider has references. References are curated successes and tell you correspondingly little. Far more revealing is what happened when something did not work — because at some point, that will happen to you too.

Pay attention to where the blame lands. Anyone who talks exclusively about the difficult client, the unrealistic budget, and the incompetent predecessor will talk about you the same way after your project.

• Good answer: A specific story that includes their own share of the failure — and a consequence that can be described: something they have done differently ever since.
• Warning sign: "Nothing has ever gone wrong for us." That means either too little experience or too little honesty.

7. "How will we measure in twelve months whether this worked?"

Most IT projects have a budget and a deadline, but no success criterion. So they end in a grey zone: the system runs, the invoice is paid, and nobody knows precisely whether it achieved anything.

Define two or three measurable figures before you start. Not "the IT is more modern," but things you can count: downtime per month, processing time for a given workflow, number of support cases, cost per workstation. A provider who engages with metrics is agreeing to be measured by them.

• Good answer: The provider proposes metrics themselves, including a baseline — and states what lies outside their control.
• Warning sign: Success criteria defined at the end of the project. At that point the outcome simply gets declared to be the criterion.

What the answers add up to

None of these questions has exactly one right answer. A provider can be weak on one of them and still be the right partner for you. What matters is the pattern across all seven.

A provider who answers clearly on ownership, exit, and conflicts of interest has had these conversations before and has the conflicts behind them. One who evades has either never thought about it — or is hoping you will not either.

And watch how the person across the table reacts to the questions, not only what they answer. A good partner likes these questions. They have long since asked them of themselves.