When the Board Asks If You’re Efficient
The note usually arrives on a Sunday.
It’s short, it’s friendly, and it’s from a board member. Sometimes there’s a link attached, a company they’d like you to take a look at. And somewhere in it is a version of the same question:
Now that we’re past the first wave of AI, are we as efficient as we could be?
It reads like curiosity. It isn’t. It’s a request for a number, and the person asking already suspects you don’t have it.
What the first wave actually bought
Be honest about what phase one was, because almost every revenue organisation ran the same play.
You bought tools, and you bought them per person. Call intelligence for the team. A sequencer or an AI SDR. Note-takers in every meeting. Copilot or Agentforce on top of the CRM. A pile of individual licences for whichever assistant your reps liked best.
Adoption happened. People are genuinely faster at things. There are dashboards, and the dashboards are green.
What almost nobody did was redraw the work around the tools. The motion stayed exactly where it was, and AI got laid on top of it like a second coat of paint. Every step still exists. Each one just happens sooner.
That’s the whole problem, and it’s why the efficiency question is so hard to answer. Speed is visible. Structure is not. And the board is asking about structure.
The question underneath the question
Boards don’t ask about efficiency because they’re curious about your tech stack. They ask because they can see the cost line and they can’t see the return.
So when that note arrives, what’s actually being requested is a defensible story in three parts: here’s what we spent, here’s what it changed, here’s what we do next and what it’s worth. Most revenue leaders can deliver the first part instantly, gesture at the second, and haven’t built the third.
And when the second part is missing, the first part starts to look like a bill.
There’s a trap waiting right here, and it’s a comfortable one: answering a measurement question by buying a measurement tool. There are good platforms that will unify your revenue data and make your forecasting trustworthy, genuinely useful, and worth owning. But a visibility layer shows you the numbers you already have. It won’t tell you which parts of your motion got more expensive when you automated them. That’s not a data problem. It’s an operating-model problem, and no licence closes it.
Where the hours actually hide
In practice, the cost sits in three places, and none of them show up in a tooling review, because no tool owns them.
- Supervision. Every AI output someone reviews carries a cost nobody budgeted. When the review takes longer than the task did, you’ve bought a very expensive first draft.
- Hand-carried work. Deal desk. Quote and proposal assembly. CRM hygiene. Board-deck and QBR prep. The work that lives between the systems, done by people, invisible to every dashboard you own.
- The expansion tax. What happens when you take a motion that works in one market into the next one. If the playbook doesn’t transfer, if your most senior people rebuild it by hand each time, you don’t have a system. You have talented people doing it again.
The last one is the expensive one, and it’s the one that stays hidden longest, because it looks like your best people being excellent rather than your operating model being absent.
Answer it before you’re asked
The advantage in this situation belongs entirely to whoever goes first. A revenue leader who walks into the meeting with a baseline, a diagnosis and a plan is running the conversation. One who waits for the follow-up question is defending a budget.
So run the baseline yourself, quietly, before anyone asks for it.
I’ve put the seven questions I use into a single page, The GTM AI Efficiency Baseline. Each one has to be answered with a number or a name. Not a story, not “the team says it’s helping.” If the honest answer is I’d have to check, that counts as unanswered, and knowing which ones you can’t answer is most of the value.
It takes about ten minutes. Do it alone before you socialise any of it.
Open the GTM AI Efficiency Baseline →
Most people score in the middle: some tools, some wins, no system underneath. That’s not a failure. That’s what phase one produces. It only becomes a problem if you respond to it by buying more software.
One more thing
I run my own firm on the stack I’m describing, agents, governance, an operating model I had to write down before it worked. I’ll tell you the part that surprised me: the first real efficiency I found didn’t come from adding anything. It came from finding two workflows where the AI had quietly added a step and taking those steps back out.
That’s usually where it starts. Not with the next platform. With an honest look at what the last one actually changed.
What would your board see if they asked today?
Love Hudson-Maggio is a fractional AI executive. She builds the operating model behind the tools, shipped agents, signed governance, and human-plus-agent team design, for organisations from a $9B health system to a national nonprofit.