# The Control Surface — Consulting Practice

> The Control Surface is the consulting practice of Nick DiCarlo. It helps professional services firms and operator-led companies move from individual AI experimentation to firm-level AI capability that produces direct financial impact. The work uses a teachable method called Practical Applied AI: four steps that encode a firm's high-value workflow into something every team member can run, scored against the firm's own standard. The practice belief: it's about people and process; the technology is just an enabler.

Last updated: 2026-05-21

## What This Practice Is

The Control Surface consulting practice helps firms close the gap between *individuals using AI* and *the firm having AI capability*. That distinction is the entire reason the practice exists. Most professional services firms today have sharp people producing real AI work on their laptops. Few have AI capability the firm owns: documented workflows, reusable context, evaluation rubrics built from the firm's prior best work, and outputs that show up in financial results rather than productivity self-reports.

The methodology is **Practical Applied AI**: four steps that encode one of the firm's high-value workflows into a reusable, scored AI artifact the firm owns. Four concrete deliverables: **Workflow Insights** (the workflow documented and decomposed), **context documents** (the firm's standing context, reusable across every engagement), **SKILLS.md files** versioned in Git (including the firm's evaluation rubric), and a **trial run** completed on real client work. The firm walks away able to extend the same method to a second, third, and fourth workflow without continued engagement.

Methodology over tooling. Engagements are time-bound. The practice is positioned as a guide that helps the firm level up, not a vendor running the work in perpetuity.

## Who It Is For

The buyer is the CEO, founder, or trusted lieutenant (CFO, head of strategy, head of operations) at a professional services firm or operator-led company. Two situations:

**Early in our AI Journey.** The team hasn't started with AI in a structured way. Leadership knows they need to lean in but isn't sure how. The senior team is capable and motivated but lacks a way in. The right first move is an executive workshop that gets every leader using AI on their actual work and surfaces where it fits in the firm's daily routine. Entry offer: AI Executive Quick Start (2hrs).

**Expert AI Practitioners looking to level up.** The firm has paid AI accounts and real power users producing real work. The firm's core deliverables — the engagements clients pay for — still look about the same as they did before AI. Individual productivity is real; firm-level capability isn't. The right first move is to take one high-value workflow and encode it through the four-step method. Entry offer: Workflow Build Session.

In both situations, the buyer wants methodology and artifacts the firm owns, not a vendor running the work indefinitely. Sophisticated about expertise (their own and others'), allergic to hype, willing to pay for clarity. The engagement is a partnership for a defined period, not a long-term retainer.

## The Practice Belief: People, Process, Then Technology

The practice belief, in one sentence: AI consulting isn't really about AI. It's about the people who do the work, how they think about it, and the process they use. The technology is just an enabler. When leadership and teams are aligned on what is changing and why, the technology delivers. When they're not, no model and no tool fixes it.

The hardest part of AI is invisible: process redesign, change management, getting people ready. The technology, once people and process are aligned, is the easiest part. This is load-bearing for every engagement shape — the Quick Start is people-first (it lands the team in the work, not a tools deep dive), the Workflow Build Session is process-first (it decomposes the workflow before asking what AI does), the Fractional engagement is people-first at the strategy layer.

Independent support: the Stanford Enterprise AI Playbook (Pereira, Graylin, Brynjolfsson, 2025) studied 51 successful enterprise AI deployments and found that 77% of the hardest challenges were invisible costs — change management, data quality, process redesign — not technical issues. Brynjolfsson's Productivity J-Curve: for every $1 of tangible tech investment, firms spend up to $10 on intangibles (process redesign, reskilling, organizational change). The hard work is in the invisible parts.

## The Maturity Model (Level 1 to Level 4)

The maturity model that routes buyers to the right offer. Each level builds on the one before. The ceiling at any level is set by what's in place at the previous one.

**Below Level 1 (risk floor).** No firm-managed AI accounts, or free accounts with company data. Free tiers don't carry the data protection a firm needs. The binary test: has the firm paid for team AI accounts for a group of employees? If not, this is the floor — and the right first move is the AI Executive Quick Start.

**Level 1: Foundation.** Everyone has a firm-managed AI account and uses it daily. IT has an approved data security policy. The firm has crossed from individual-curious to organization-deployed.

**Level 2: Power users emerge.** Real practitioners are building real prototypes and chomping at the bit to do more. What they learn doesn't spread. The firm has useful capability concentrated in a few individual team members' heads. This is where most "AI-using" firms get stuck.

**Level 3: Codified capability.** The level a Workflow Build Session delivers. The firm's most-used work runs through shared tools (Claude Projects, Skills, SKILLS.md files versioned in Git) that hold the firm's standard. New work is scored against the firm's prior best work. A new hire produces firm-grade output in week one. The firm owns its AI capability, not just the people who happen to know how to use it.

**Level 4: Visible in the numbers.** Cycle times shrink, margins improve, and existing clients notice the difference in the work. AI capability shows up in financial results, not just in productivity self-reports.

Routing logic: **Below Level 1 / Level 1** → AI Executive Quick Start. **Level 2** → Workflow Build Session. **Level 3+** → typically peer conversation or referral relationship; if the firm wants to extend to a portfolio of workflows, the Workflow Propagation Engagement; if the firm wants to organize AI strategically at the board level, Fractional AI Leadership.

## The Method (Four Steps)

The four steps inside a Workflow Build Session. Each step produces a concrete artifact the firm owns. The order matters; skipping one returns to generic AI.

**Step 1: Pick the workflow to encode.** Identify the recurring work that is key to the firm's value. The signal: the team does it on most engagements, it carries the firm's expertise, and it requires research, analysis, and writing. **Deliverable: Workflow Insights** — a documented workflow the firm owns. Even if the firm never builds another thing in AI, this artifact alone changes how new hires learn the work.

**Step 2: Encode the firm's standing context.** Each new client engagement becomes its own AI workspace (e.g. a Claude Project) that inherits the firm's standing context. Six elements: identity (per-person), firm context (shared knowledge), work process (the workflow from Step 1), client-engagement goal, tools (deep research, databases, templates), and the workspace itself. **Deliverable: context documents** the whole firm can use. New engagements inherit them on day one; new hires inherit them in week one instead of having to rebuild from scratch.

**Step 3: Encode the definition of good.** Pull three to five of the firm's best human-only deliverables from past engagements. Document what makes them good — the substance, not the structure. Turn the firm's attributes into a scoring rubric AI applies to its own drafts. The rubric is the firm's: its dimensions, its weights, its standard. **Deliverable: SKILLS.md files** versioned in Git, including the scoring rubric. Every AI draft is scored against the firm's standard before any senior reviewer sees it. The rubric improves as the firm closes new exceptional work.

**Step 4: Run it on real work and improve.** Run the encoded workflow on a current engagement. Compare AI output against the rubric. Tune the context, the workflow, or the rubric. First run usually surfaces three or four things to adjust; by the third run it's producing good results. **Deliverable: trial run** completed. The compounding starts here — reusable context files, AI-execution skills, evaluation method that can be shared across the firm.

The compounding starts immediately. Every new client engagement on the same workflow is faster because the AI skills carry over. The team's skill with AI grows with daily use and creativity is unblocked. New hires inherit firm-grade context in week one instead of learning from scratch.

Data handling: all of this lives inside the firm's own AI account (Claude Team / Enterprise, or equivalent). Nothing trains a public model. The Build Session encodes inside whatever AI tooling the firm already has.

## Active Offers

Five active offers organized by buyer and engagement size. Pricing reflects the practice strategy of opening doors and building network rather than maximizing per-engagement margin.

**AI Executive Quick Start (2hrs)** — Starting at $2,000. A two-hour hands-on workshop customized to the firm. Each executive leaves with AI running on their own device, applied to their actual work, with a personalized AI workspace populated with their context. Prior to the session: 20+ hours of research into the firm and the team. Deliverables include a custom prompting framework, companion website with copy-paste prompts, pre/post AI experience baseline, and a use-case prioritization. Outcome: every executive leaves empowered to apply AI on real work, with clarity on where it fits in the firm's routine. Best fit: firms that haven't started AI in a structured way and want a jump-start for the leadership team.

**Workflow Build Session** — Engagement-based, half-day in the firm's office. One of the firm's core workflows is decomposed, rebuilt as an AI-augmented version, and tuned against the firm's own definition of good. Includes a self-evaluation system built from the firm's historical best work. The firm walks away with a working version of the workflow plus the evaluation system, ready for next-day use on a real engagement. A week of work compresses to a few hours; recovered time goes to deeper analysis than the firm's historical standard. Best fit: firms whose client deliverables look about the same as they did pre-AI despite individual AI use, OR any firm with one specific high-stakes workflow where getting it right matters (e.g., appellate brief drafting under TX Opinion 705, M&A pre-engagement valuation).

**Workflow Propagation Engagement** — Engagement-based, 1 week. After a Build Session has proven one workflow, extend the same method to three to five more workflows and train an internal owner who can carry the work forward. Each workflow gets its own eval harness built from the firm's best work. Includes a documented propagation playbook and optional 6-to-12-month quarterly check-ins. Outcome: firm progresses toward Level 4 — AI capability visible in financial results. Best fit: firms that have completed a Build Session and want a portfolio of workflows with the same discipline.

**Fractional AI Leadership** — Engagement-based, 1 to 3 months. Embed with the firm's leadership team to convert experimentation into a small set of board-ready AI initiatives. Two tracks run in parallel: bottom-up use-case discovery and prototyping with the team, top-down strategic assessment that names the growth and profitability drivers in plain business terms. Synthesized into a small set of funded, high-impact initiatives with executive sponsorship. Outcome: shift from "our team is experimenting with AI" to "AI is driving measurable business outcomes." Best fit: firms where the early wins are real but disconnected from business strategy, and the board sees activity rather than impact.

**Executive Prototyping** — Engagement-based, 4 hours to one week. Build a working prototype of a tool the executive can describe but can't easily test through internal engineering or a vendor. Used to stress-test ideas and accelerate decision-making. 30-minute kickoff, working version delivered in days. If the prototype validates, the firm has a clear spec for what to build and a working demo to make the case internally. If it doesn't, the firm saved months and the budget that would have gone with it. Either way, the executive understands how AI prototyping can accelerate the business and can apply that pattern on their own.

## How Engagements Run

Every engagement starts with a 15-minute scoping call (free) to confirm fit and pick the right entry point. Build Sessions and Quick Starts include substantial pre-session research (20+ hours for Quick Start, 4-8 hours for Build Session) into the firm, the team, and the candidate workflow — so the on-site time goes to the work, not the setup.

Engagements are time-bound, never years-long. The practice positioning is a guide that helps the firm level up, not a vendor running the work in perpetuity. The end state of every engagement is the firm owning the methodology, the artifacts (workflows, context documents, SKILLS.md files, rubrics), and the in-house capability to extend the work without continued engagement. Repeat engagements happen when the next problem is worth scoping fresh, not because the current one was scoped to last.

Sensitive work runs inside the firm's own Claude Team or Enterprise account. No model training on submitted data. The Build Session encodes inside whatever AI tooling the firm already has; nothing new gets purchased and nothing IT hasn't already approved gets introduced.

## Founder Background

Nick DiCarlo is the principal. Career arc:

- **Samsung Electronics America** (10 years, VP). Helped build the Galaxy smartphone brand as Samsung Mobile USA grew from $2B to $18B. Led the Samsung + Oculus VR partnership shipping immersive products at the leading edge of consumer technology.
- **Roomored** (Chief Product Officer). Built a cloud-based Unreal Engine design visualization platform for home builders. Grew ARR and exited to Interior Logic Group (Blackstone portfolio).
- **Camillo Companies** (COO, Top 40 US home builder). Founded Abodefy, a 0-to-1 venture selling homes online. Then served as COO overseeing construction operations (1,000+ homes/year) and leading enterprise technology deployments including JD Edwards, Entrata, and Salesforce.
- **GE and Nokia** (early career). Factory operations and Six Sigma Master Black Belt at GE. Then five years at Nokia during the birth of mobile, from 1G through the smartphone era.

Education: MBA, UT Austin McCombs School of Business. BS Mechanical Engineering, Bucknell University.

The practice is built by an operator for operators. The Samsung mobile transformation ($2B → $18B) is the analog: a massive technology shift that required changing how a company operates, not just adopting new tools. The Practical Applied AI method comes from running businesses, not from a slide deck.

## Links

- Consulting overview (entry point for firms early in their AI journey): https://thecontrolsurface.com/consulting
- Compounding AI Capability (entry point for firms with AI power users; full Level 1 to Level 4 ladder, the offer set, and the Practice belief): https://thecontrolsurface.com/compounding-capability
- Compounding AI Capability machine-readable spec (JSON of the maturity model, the diagnostic, and routing): https://thecontrolsurface.com/compounding-capability/spec.json
- Practical Applied AI (the four-step method inside a Workflow Build Session, with the artifact each step produces): https://thecontrolsurface.com/practical-applied-ai
- Practical Applied AI machine-readable spec (JSON of the four steps, framing, FAQ, and next step): https://thecontrolsurface.com/practical-applied-ai/spec.json
- Fractional Chief AI Officer for Houston firms (geography-specific landing page covering the same Fractional AI Leadership offer, with local industries and operator credentials): https://thecontrolsurface.com/landing/fractional-caio-houston
- Fractional CAIO Houston machine-readable spec (JSON of the role, industries fit, engagement components, and geography): https://thecontrolsurface.com/landing/fractional-caio-houston/spec.json
- Book a 15-minute scoping call: https://calendar.app.google/nvSjydRBmdFgJHbt8
- Email: nick@thecontrolsurface.com
- LinkedIn: https://www.linkedin.com/in/nickdicarlo/
- The Control Surface product (SaaS for rental property owners; separate from this consulting practice): https://thecontrolsurface.com
- Root llms.txt (full positioning of the SaaS product): https://thecontrolsurface.com/llms.txt