NewYour 2027 technology budget is going up. Here's where you still have choices.Read it →

AI Strategy & Implementation

From deciding where AI fits to rolling it out across your organization. Independent advice and hands-on help with tools, policy, training, and the information AI depends on.

Tell us what's going on

Photo by Chris Lawton / Unsplash

Most organizations are past asking whether to use AI. Staff are already using it, often on personal accounts. Vendors are adding it to software you already pay for. Your board, funders, or clients are asking what your plan is.

The real questions now are more practical: which tools, for what, under what rules, and how will you know it’s helping? We help leadership answer those questions, and then we help make it happen.

We usually hear from organizations when:

  • Staff are using ChatGPT, Claude, Gemini, or Copilot on their own, and leadership wants a coordinated approach before something goes wrong.
  • You’re choosing between Copilot, Gemini, Claude, ChatGPT, or something else, and every vendor says theirs is the obvious answer.
  • You rolled out an AI tool, and usage dropped off after the first month.
  • Software you already use just added AI features, and nobody has looked at what they do with your data.
  • Your board or funders are asking about AI, and leadership wants a realistic picture of what it can do today, and what it can’t do yet.
  • You’d like AI to take on real work, not just help write emails.

Tell us where you are with AI →

Strategy: deciding what to do

  • Straight talk for leaders and boards. Briefings and workshops for executive teams, boards, and funders on what AI can do well today, what it can’t do yet, and where the risks are, without the hype or the doom.
  • Finding out what’s already happening. Most organizations have more AI use than they realize. We look at which tools staff are using, on which accounts, and with what information, so you’re starting from reality. More on shadow AI.
  • Use cases and priorities. Where AI would make a real difference for your organization, and what to try first.
  • Tool and platform decisions. Claude, ChatGPT, Copilot, Gemini, or a model that runs on your own hardware, chosen to fit your work, your data, and your budget. That includes the AI features quietly appearing in software you already use.
  • Policy and governance. Short, readable acceptable-use policies, rules for what information can go into which tools, who approves new tools, and what changes when AI starts taking actions on its own instead of just answering questions.
  • Getting your information in shape. AI tools draw on whatever they can reach. Sorting out sources of truth, outdated content, and permissions first makes the answers far more reliable.

Implementation: making it work

  • Organization-wide rollouts. Configuration, security and privacy settings (including whether your data can be used to train models), testing, and phased rollout of Claude, ChatGPT, Gemini, or Microsoft 365 Copilot.
  • Training and playbooks. Practical, role-based training, plus prompt libraries and internal playbooks your staff can keep using after the training ends.
  • Monitoring and adjusting. Regular reviews of who’s using what, what’s working, and where people are stuck, so adoption keeps growing after launch.
  • Connecting AI to your systems. Integrations and agents that let AI work with your other tools, custom dashboards and internal tools, and local models when information can’t leave your environment.
  • Putting AI to work in operations. Rethinking specific workflows with your operations team. See our process optimization work.

What this has looked like

Claude for an entire organization. For one of our fractional CIO clients, we rolled out Claude to the whole team this year. We handled the enterprise configuration, testing, and training, then monitored usage weekly after launch to see who needed more support and where.

Copilot, more than once. We’ve implemented Microsoft 365 Copilot for several organizations, from configuration and testing through training and rollout.

Knowledge first, then AI. An affordable housing nonprofit wanted to know whether their information was ready for AI before investing in it. We assessed their knowledge management practices and built a phased roadmap focused on ownership, content lifecycle, and searchability. Read more client stories →

Why FireOak

We’re independent. We don’t resell AI tools or take commissions, so we can tell you honestly which tool fits, or whether you need one at all.

We use these tools every day. In our own work and alongside clients. Here’s how we use them.

We start from your information. Our background is in knowledge and information management, which is exactly what determines whether AI gives good answers or confident wrong ones.

We see the whole picture. As fractional CIOs, we look across your systems, contracts, security, and people, so AI decisions fit with everything else.

We’re part of the conversation. Abby is a member of Senior Executive’s AI Think Tank and speaks and writes regularly about AI and knowledge management.

Common questions

Should we wait until AI settles down?

Waiting usually means staff use AI anyway, without guidance. A small, well-governed start is safer than no plan, and it gives you real experience to base bigger decisions on.

Which AI tool should we choose?

It depends on your work, your existing systems, and your data. Copilot often fits Microsoft 365 organizations, and Gemini often fits Google Workspace ones, but many get more from Claude or ChatGPT. Plenty end up with more than one. We’ll help you decide based on your needs, not a vendor’s pitch.

Do we need an AI policy?

Yes, and it can be short. Staff need to know which tools are approved, what information they can and can’t use, and who to ask. We can help you write one people will actually read.

Do we need to clean up our information first?

Not all of it, and not before you start. But AI tools draw on whatever they can reach, so the information they’ll use most needs to be current and properly permissioned. We help you focus cleanup where it matters. See how we approach knowledge management.

Related reading: Your First 5 Steps Toward AI Readiness, Should You Let ChatGPT, Claude, or Perplexity Use Your Data for Training?, What Happens When Your Approved Software Adds AI?, and AI That Acts: Why Agentic AI Changes the Risk Equation.

Tell us where you are with AI

A few sentences is plenty: what you’re using now, what you’re considering, and what’s worrying you. We read every message ourselves and reply within one to two business days.

Prefer email? Write to us at [email protected].

We only use this to reply to you. No mailing lists unless you ask.

Let's keep in touch

Signal, the FireOak newsletter: practical tips and thoughtful takes on technology, knowledge, and AI. No spam. Unsubscribe anytime.