Buy-In Beats Mandates for AI Transformation

The hard part of AI transformation isn't the technology, it's the people. And there's a playbook we learned for winning them over.

Picture yourself as a CTO or engineering lead who just rolled out Copilot or Claude across the firm.

People are building their own shortcuts: getting cleaner status reports, faster RFI replies, and self-serviced reports. That’s a real win, but it’s also the easy part.

Leadership wants more than this “decentralized” transformation effort, though. They’re asking how this shows up in the numbers that matter this quarter: contracted revenue, project margin, delivery speed.

A few hundred personal automations don’t add up to that on their own. We covered the reason in The Future of Work: these are siloed quality-of-life improvements, but don’t fundamentally change the throughput of an end-to-end project.

So how do you steer a business toward the wins that leadership is actually seeking to justify their AI investments?

You need to invest in a “centralized” transformation effort: a rollout designed top-down, that redesigns a process from first principles, and gets mandated rather than discovered.

It’s how you get those wins. But it’s also where your employees get scared.

Decentralized

Task 1

Task 2

Task 3

Outcome 1

Outcome 2

Outcome 3

Outcome 4

Centralized

Task 1

Task 2

Task 3

Outcome 1

Outcome 2

Outcome 3

Outcome 4

Decentralized transformation is many small wins scattered across the org. Central transformation is one process, redesigned and pushed through end to end.

Carrots, not just sticks

Automating a task is not the same as replacing an entire job responsibility, but as AI gets introduced into more and more tasks, people understandably worry about their job.

The fix isn’t a bigger stick; it’s a carrot that incentives people to buy into uncomfortable workflow changes in the short-term which will free up their time from all the manual toil in the long-term.

Getting from a mandate that people tolerate to a rollout where people are clamouring to be next and use the cool technology you’re building requires a level of tact & strategic introduction. Skipping too fast to the end result is the most common way central transformation stalls.

Below is a common framework that brought us success in the past.

Start with a wedge, not the whole domain. Pick one process to redesign from first principles under the new technology, narrow enough to ship in weeks and show a result everyone can see. Pilot projects play this role in BIM adoption: a firm runs one project through the new workflow, irons out what breaks, then expands. Big enough to matter, small enough that a stumble doesn’t sour the whole initiative.
Find your champions inside the wedge. Every pilot surfaces one or two people who get genuinely excited rather than merely compliant, often the ones closest to the tedious part of the old process. AEC firms already have a name for this in BIM rollouts: the BIM champion, who mentors peers and troubleshoots on-site instead of from corporate. Bring them into the build itself: let them write the skills that encode how the work should run, since they understand the process better than any outside designer. Their credibility with peers is worth more than any announcement.
Only then reach for the high-ROI expansion. Once a wedge has shipped and champions are pulling colleagues in rather than a mandate pushing them, you’ve earned the right to run orchestrated AI systems across the rest of the domain. The effort shifts from “does anyone trust this” to “how much of the domain can we now cover.”

The value is in the skills, not the model

The best practices that make a workflow reliable need to be written down, kept current, and stay debuggable in plain English, so an operational partner, not just an engineer, can read one and know exactly what it does. Wiring those skills to a firm’s existing SaaS tools, on infrastructure not locked to one model vendor, is the layer we build at Sinusoidal.

/rfi-follow-up

Step 1: Pull open RFIs from the tracker each morning.

Step 2: Match each RFI to its assigned discipline lead.

Step 3: Flag any RFI missing a response after 5 business days.

Step 4: Post a summary to the #rfi-tracking channel.

Reviewed by: J. Alvarez

Vibe-Coded Automation

import { Pool } from "@internal/db-client";

const pool = new Pool(DB_CONFIG);

async function checkStaleRfis() {

  const { rows } = await pool.query(`

    SELECT r.id FROM tbl_x27_rfiq r

    JOIN cfg_bizcal c ON c.id = r.cal_id

    WHERE DATEDIFF_BIZ(r.last_resp, NOW()) > 5

    AND r.flag_stat NOT IN (2, 7, 9)

`);

  if (!rows.length) return [];

  return rows.map(normalizeRfiRow);

}

A skill that reads like an instruction sheet, not code, is one a champion can actually maintain.

We’ve watched this order, wedge, champions, then scale, play out in other knowledge-work settings.

What changes by industry is who’s willing to run the first wedge, and in civil engineering, that’s still mostly untapped ground.

If you’re weighing a top-down AI initiative and want a partner to help run the first wedge, find your champions, and get the skills written and governed right, we’d welcome that conversation.

Talk to Sinusoidal