In May we relaunched Greenplaces around one idea: expert-reviewed work, across every framework. Since then the question I get most often has nothing to do with frameworks. It's usually some version of, isn't AI going to do all of this?
Some of it, yes. A lot of it already. But the question skips past a better one, which is what a professional services firm is really selling.
My answer, after a year of having this conversation: the next generation of professional services firms won't sell hours, software, or dashboards. They'll sell completed, expert-reviewed work. That's the model we're building at Greenplaces.
Three things firms sell today
What the customer wants is the finished thing. The completed questionnaire. The filed report.
For most of the history of this industry, the only way to buy the finished thing was to buy hours, because producing it was expensive and hours were how you paid for the production.
That second part is what changed.
What AI actually changed
Figuring out which of five systems holds the number. Mapping one access-control policy onto SOC 2, ISO 27001, and a customer's own questionnaire. Reformatting the same emissions figure for CDP, for EcoVadis, and for a supplier portal that only takes a PDF. Writing 200 security questionnaire answers when you already answered 190 of them last quarter, in a different order, for a different buyer.
Machines are very good at that. It's also where most of the hours in a traditional engagement go.
The expensive part of compliance work was never judgment. It was assembly.
Compress it and you don't end up with cheaper consulting. You end up selling something different. The unit stops being time and starts being output.
What AI didn't change
Someone still has to be right.
Take this year. CARB reset the first SB 253 reporting deadline to November 10, 2026, pending final administrative sign-off, and kept Scope 3 out of the first cycle, with a limited five-category version planned for 2027. SB 261 is sitting behind a Ninth Circuit injunction in a First Amendment challenge, though the same court let SB 253 go ahead. In the EU, the Digital Omnibus on AI has been in force since late July and pushed the AI Act's high-risk obligations from August 2026 to December 2027, while the transparency obligations largely landed on schedule.
A model working from last year's training data will give you the old dates, fluently, with complete confidence.
The dates matter less than the pattern. These requirements move, and they move for reasons that are political, litigated, and unfinished. Knowing which deadline is real this week isn't a retrieval problem.
Under the dates sit harder questions. Is your data good enough to assert? Which of two defensible methodologies holds up when the auditor asks a second question? When is "we don't know yet" the right thing to tell a buyer, and when does it lose you the contract?
Somebody has to own those calls, and owning one means a name on it and a person who answers for it if it's wrong. A model can't do that. Neither can a dashboard.
So we say expert-reviewed rather than AI-assisted.
The review isn't quality control at the end of an automated pipeline. It's the thing you're buying. Nothing goes out with our name on it that a senior practitioner hasn't read and put themselves behind.
What that looks like in practice
Those aren't gains you get from better software. Nobody's score moves because a platform filed their evidence more neatly. They move because a practitioner reads a scorecard, works out which specific gaps are worth closing before the reassessment, and then closes them.
What this means if you're the one buying
You buy completed work instead of effort. The deliverable is the deliverable, not a discovery phase you fund and hope converges.
And instead of a tech support associate, you're working directly with an expert who maintains decades of experience from Big 4 accounting to leading Sustainability consultancies.
Why we do this
The line at the top of our site is still the one we hear on almost every first call. Your customers keep adding requirements. Your team hasn't grown.
That gap isn't closing. Consulting can't close it at consulting prices. Software can't close it, because software hands the work back to the person who didn't have time for it to begin with. The only thing left is to change what gets sold.
I think the firms that matter in the next ten years of professional services will be the ones that deliver finished work: assembled by machines, owned by experts, priced on the outcome instead of the hour. That's less a forecast than a description of how we already run.
If requirements are showing up faster than your team is growing, book a strategy call. We'll start by telling you what actually applies to you.