AI-Amplified or AI-Free: two delivery modes, one standard

Most professional services firms have made a decision about AI in delivery. Most aren't telling clients about it.

The decisions tend to be one of three things. The firm has fully embraced AI in delivery and is bragging about it without quite saying what that means. The firm is using AI in delivery and is staying quiet because the trust questions feel hard. The firm is rejecting AI in delivery for cultural or practical reasons and isn't quite sure when to bring that up.

None of those is what we wanted to do. So we did something different. We use AI deliberately in delivery, we govern it strictly, we're explicit about it on every engagement, and we offer clients an explicit choice about how AI shows up in their work.

This post explains the choice, why we structured it the way we did, and what we've learned from the engagements that have run under each option.

The two modes

Every PluggedIn engagement runs in one of two delivery modes. The mode is named in the engagement letter at the start. The mode can be changed mid-engagement when the work calls for it.

AI-Amplified Delivery is the default. Senior practitioners use enterprise AI tools internally throughout the engagement. The tools handle document review, scenario modeling, first-draft analysis, code analysis, and research synthesis. The senior practitioner reviews every output before it leaves our shop. AI tooling is used only under approved enterprise data controls, with retention disabled where the tool allows and client data excluded from model training where contractually covered. Final decisions, recommendations, and deliverables are signed off by the practitioner — not the model.

AI-Free Delivery is available on request. The same senior practitioners deliver the engagement, but named AI tooling is disabled in the practitioner's working environment for the duration. Documentation of the systems used to produce the work is available on request. AI-Free Delivery requires additional senior hours, and pricing is set per engagement during scoping.

The choice is the client's. Not ours.

Why we offer the choice

We could have picked one mode. Most firms do, even if they don't tell clients about it. Three reasons we structured it as a choice instead.

The first reason is that the right answer isn't the same for every engagement. A growth-stage business preparing for a Series B doesn't have the same AI governance needs as a regulated financial services firm preparing for a regulatory examination. A foundation grantee getting technical assistance has different priorities than a public-sector consulting engagement under federal data residency requirements. Pretending one mode fits every situation is the same kind of one-size-fits-all framing we'd reject in every other part of our delivery.

The second reason is that the AI-Free option is meaningful only if it's real. Some firms claim to offer "human-only delivery" while continuing to use AI tools internally for things they've decided don't count — drafting emails, summarizing notes, transcribing meetings. We didn't want a soft line. AI-Free Delivery means exactly that. The practitioner's working environment is configured without AI tooling for the duration of the engagement, and we can produce an audit trail confirming it.

The third reason is the most honest one. Some clients don't want to be the test case for an AI-amplified delivery model, even if the model is well-governed. The trust hasn't been earned yet. Clients who feel that way deserve a clear path that respects the concern without forcing them to look for a different firm.

What we've learned from each mode

A few patterns have emerged from running engagements under both options.

AI-Amplified engagements move faster than fully human engagements by a meaningful margin. The practitioner uplift is real — particularly on document-heavy work, scenario modeling, and structured analysis. We've measured it informally across our delivery, and the productivity gain is consistent enough that we feel comfortable building it into pricing rather than offering it as a discount.

AI-Free engagements feel different to the practitioner. Several of our practitioners have delivered engagements under both modes. Their reflection: AI-Free delivery is slower, more time-consuming, and in some cases produces narrower analysis simply because the practitioner can't process as many documents or model as many scenarios in the available time. The senior judgment is the same. The breadth of analysis is genuinely different.

Mode-switching happens. Several engagements have started in AI-Amplified mode and switched to AI-Free for a specific phase — typically a regulatory submission, an audit support window, or a sensitive transaction. Other engagements have started AI-Free and shifted to AI-Amplified once the client became comfortable with our governance posture. Both directions are common. The flexibility matters.

Clients who choose AI-Free often have specific governance reasons. Most of the AI-Free engagements we've delivered have been for clients with internal AI policies, regulatory constraints, sensitive data classifications, or board-level positions on the use of AI in advisory work. AI-Free Delivery isn't typically a preference — it's usually a structural requirement.

Why this matters for the AI question your firm hasn't asked yet

Most professional services buyers — particularly in regulated industries, in mid-market businesses, and in publicly-funded portfolios — are starting to ask harder questions about AI in delivery. The questions are reasonable. They're not always being answered well.

If you're evaluating advisory firms, we'd suggest asking three things explicitly:

• Are you using AI in delivery on this engagement, and how? A non-answer is an answer.
• What governance applies to my data when AI tooling is involved? Ask which tools are used, what retention controls are available, whether training is disabled contractually, and how exceptions are handled. If a firm can't articulate the configuration, that's information.
• What's the option if my organization needs human-only delivery? Many firms can't offer one. Those that can should be able to articulate the cost, timeline, and audit-trail differences honestly.

Our positioning is that we'd rather have these conversations explicitly at the start of every engagement than have them surface later as a trust question. The two-mode model is how we made that explicit.

What's next

We expect this question to get harder, not easier. AI capabilities are moving fast; client governance frameworks are catching up; regulatory positions are still being formed. The firms that will earn trust over the next several years are the ones that can articulate their delivery posture clearly and adapt as the standards evolve.

We'll update the AI Principles page when our delivery posture changes. The change history will be visible there.

If you'd like to talk through how this fits a specific engagement, start a conversation — every client conversation includes a discussion of which mode fits the work.