Your AI mandate has no plan. Here is where to start.

Here is the situation, named plainly: leadership handed you an AI mandate. Maybe it came out of a board meeting, maybe a conference keynote made someone nervous, or maybe a competitor announced something. The budget is modest or nonexistent. Your CRM data is inconsistent enough that you do not fully trust it yourself. And you are expected to show progress by next quarter.

Nobody handed you a plan along with the mandate. That is normal. Mandates rarely come with execution plans attached, because the person issuing the mandate is reacting to pressure, not designing a rollout.

The direct answer: do not start by buying a tool. Start with a stack audit, then build a one-page roadmap. In that order. Here is why, and how.

Why “buy something to show progress” backfires

The instinct under this kind of pressure is to move fast and visibly. This might involve signing up for an AI SDR tool, rolling out a copilot, or announcing something in the next all-hands. It feels like progress because it is an action you can point to.

The problem shows up around week six to eight. The tool needs a workflow it fits into, and nobody defined one. It needs clean data to work from, and your CRM has gaps you already knew about but never fixed. It needs an owner accountable for adoption, and “everyone” is not an owner.

A mandate with no plan does not get fixed by adding a tool. It gets fixed by adding a plan.

By the 90-day mark, you often have a subscription with low usage. This leads to a harder conversation with leadership than the one you started with. Now you have to explain why the thing you bought did not work.

There is also a quieter risk worth naming: the mandate itself has a shelf life. If nothing visible happens for two quarters, the mandate does not stay open indefinitely so you can eventually get it right. It either gets reassigned to someone else, absorbed into a vaguer “digital transformation” initiative that loses urgency, or leadership concludes AI was not the priority they thought it was. This makes the next mandate, on any topic, harder to get resourced. Moving deliberately in week one is not just about avoiding a bad tool purchase. It is about keeping the mandate alive long enough to do something real with it.

The right first move: audit what you have

Before anything else, get a real inventory of your current stack. This is not a vibe check, but an actual list. It should include every tool, who owns it, what it costs, how many seats are bought versus actually active, when it renews, and what it integrates with.

This does two things. First, it tells you whether the problem the AI mandate is supposed to solve already has a tool sitting half-used somewhere in your stack. Second, it gives you a factual starting point instead of an opinion. This matters a lot in week one of a mandate when everyone has a different theory about what “AI in sales” should mean.

The full method for running this is in the stack audit guide. It is designed to be run in days, not weeks. It is the literal first deliverable most sales-ops-minded engagements start with, for good reason: everything downstream depends on knowing what you actually have.

What a one-page roadmap actually contains

Once the audit is done, turn it into a roadmap. This is not a strategy deck, nor a vision statement. A roadmap is a short, sequenced list of specific initiatives. Each one is concrete enough that someone could start on it Monday.

Each initiative on the page needs:

• The workflow it touches. Not “improve efficiency,” but “cut the time reps spend building account research before a first call.”
• An owner. A named person accountable for whether this actually gets adopted, not a department.
• A timeline. When it starts, when you check in, when it should be showing results.
• A kill criterion. The specific condition under which you stop and cut losses, decided in advance, not improvised after three months of quiet failure.

Three to five initiatives is enough. More than that and you have a wish list, not a roadmap. The full breakdown of what makes a roadmap real instead of a shopping list is in what is an AI roadmap for a sales team.

What if there really is no budget

A modest or nonexistent budget is common enough with these mandates that it is worth addressing directly. The audit and the roadmap itself cost you time, not money, so they are available regardless of what leadership approved for spend. That is not a workaround, it is the correct sequence anyway. You should know what you have and what you need before you ask for budget to buy something new.

When you do need to ask for money, a roadmap with named initiatives, owners, and kill criteria is a far stronger budget request than “we need an AI tool.” It shows leadership exactly what the money buys, how you will know if it worked, and when you will cut it if it does not. That specificity tends to get budget approved where a vague request would not, because it reads as a plan rather than a hope.

Report progress in week 2, not month 6

Here is the part that actually protects you politically. If you wait until you have results to report anything, you are silent for months while leadership wonders what happened to the mandate they handed down. That silence gets filled with the wrong assumptions.

Instead, report the audit findings in week two. This is not a finished initiative, just the factual state of the stack. Detail what you found, what surprised you, and what the roadmap will address first and why. This does three things. It shows movement immediately. It sets expectations honestly, so month three is not a surprise. And it puts you in the position of the person who is managing the mandate deliberately, instead of the person who got a mandate dropped on them and is scrambling.

Where a diagnostic conversation fits in week one

None of this requires an outside party. You can run the audit and build the roadmap yourself with the templates linked above. But if you want a second set of eyes on the audit findings, or want someone who has run this exact sequence before to sanity-check the roadmap before you take it upstairs, that is what a discovery call is for. It is a diagnostic conversation, not a pitch. The goal is to leave with a clearer roadmap, whether or not anything gets built together afterward.

For a self-check on whether your team is actually ready to run a first pilot once the roadmap is set, the AI readiness assessment walks through the four things that predict whether a pilot survives: ownership, process, data, and infrastructure.

What this looks like by the end of quarter one

If you run this sequence, here is roughly where you land by the time leadership checks in.

| Timeline | Deliverable