An AI policy that gives your staff clear answers.
Your people want to know which tools and data they can use, and who to ask before trying something new. Accel Comply puts those choices to leadership with a recommendation, so the policy can give plain answers rather than repeat a template.
The short answer
Accel Comply, founded by Behzad Motaghi, helps leadership teams in Denmark and the Nordics make the decisions a usable AI policy for employees rests on: which tools and data are allowed, who approves new uses, what is off limits and how use is followed up. Accel Comply then drafts the policy from those decisions in language staff can use.
Four decisions a usable AI policy has to rest on
A usable AI policy rests on four decisions. Which tools can be used, and with what data: can customer data, personal data or a draft bid go into a free tool, or only into tools you have a contract for? Who approves a new use, so staff know who to ask? What is off limits whatever the tool? And how is use followed up, so the policy lives somewhere other than the intranet? Once those four are settled, people can use AI without guessing, and leadership knows what it has said yes to.
Why a template can't make those decisions for you
A template gives you a structure, but it doesn't know your data, your customer contracts or the tools you already pay for. So it falls back on lines like ‘use AI responsibly’ and ‘take care with confidential information’, which don't help someone with a customer contract in one window and a chatbot in the other. Nor can it say who in your company has the mandate to approve.
From how AI is used today to a policy people can use
The work starts with how AI is being used in your company today, not with a document. The senior leading the engagement talks to the people using the tools, looks at which suppliers you have agreements with, and puts the four decisions to leadership with a recommendation and the consequences of each option. Once leadership has decided, the same senior lead writes the policy in language staff can use. After that, Accel Comply checks whether people can work with it and reviews it again when a new tool or a new use turns up.
When staff ask and nobody can answer
A salesperson asks whether she can paste a customer's tender into a free chatbot
Nobody in leadership has decided, so the answer depends on who she asks. With a policy built on a decision about data and approved tools, she can check what's allowed herself and knows who to ask when she's unsure.
The policy came from a template and was signed, but staff still ask
It says AI should be used responsibly, but not which tools are approved or who can say yes. Accel Comply finds the missing decisions so leadership can take them, and the policy becomes something people actually use, not something they once signed.
Developers want an AI coding assistant, and marketing already uses image generators
Each team has found its own tools, and nobody knows who approves them. With one route for new uses and an owner for each, teams can get on without waiting for leadership to take up every request.
What you get, and what you don't
What Accel Comply takes on
- A picture of which AI tools are already in use in your company, and for what
- The decisions the policy rests on, put to leadership with a recommendation and the consequences of each option
- A draft policy written from leadership's decisions, in language staff understand
- One route for approving new uses, and a date for when the policy comes back for review
What Accel Comply does not take on
- A template with no decisions behind it
- Technical set-up, blocking or monitoring of tools
Leadership decides what the policy says and is accountable for it.
Questions about AI policies for employees
What should an AI policy for employees cover?
Above all, four things: which tools and data are allowed, who approves new uses, what is off limits, and how use is followed up. Staff should also be able to see who to ask when they're unsure. If a draft policy can't answer those four questions, the decisions behind it haven't been made yet. Those are the decisions Accel Comply helps leadership make.
Can employees use ChatGPT at work?
Until leadership has decided, practice can vary from person to person, so staff should follow the rules you already have and ask a manager if they are unsure. The decision covers three things: whether staff must use a company account under a business agreement with the provider or may use a personal one, which data must never be typed in, and which tasks the tool may be used for. Once it is in the policy, the same rule applies to everyone.
Can't we just use an AI policy template?
A template can give you the structure, not the content. It doesn't know which tools you have contracts for, which data is sensitive in your business, or who has the mandate to approve. Use it as a checklist by all means, but let leadership make the decisions first, or you end up with a policy nobody can act on.
Will the policy be legally watertight?
Your lawyer needs to assess that. The policy rests on leadership's decisions about how AI is used, and Accel Comply sets out what your lawyer should check, from the AI Act to data protection rules and your contracts, so you aren't paying for an open-ended review.
Who should approve the AI policy?
Leadership, because the policy sets how much risk the company accepts with data, customers and suppliers. Whether that's the executive team or the board depends on how responsibilities are split in your company. What matters is that whoever approves it also has the mandate to say yes or no to new uses later.
What about Microsoft 365 Copilot?
The policy can say what staff may use Microsoft 365 Copilot for, but the tool also needs its own decisions on data, access and approval, which leadership should make before it becomes routine across the company. Those decisions are covered on the Microsoft 365 Copilot governance page.
Related pages
What are your staff asking that nobody can answer?
List the tools in use and who will approve the policy. You'll get a view on where to start.