How We Use AI To Run A 90-Day Growth Audit

Most growth audits are a performance. Someone shows up with a slide deck, interviews a few stakeholders, and delivers a 40-page PDF that lives in a drawer. The team feels busy for three weeks, and nothing changes. I’ve been on both sides of that transaction, and I got tired of it.

At my growth consultancy, we run 90-day growth sprints for venture-backed and private equity (PE)-backed companies. The audit is the first phase. It used to take two to three weeks of manual work just to get a clear picture of what was happening inside a company’s marketing organization. Now, with AI woven into every step, we compress that discovery into days and spend the remaining time actually fixing things.

Here’s exactly how we do it.

Why Traditional Growth Audits Fail

The classic consulting audit has a structural problem. The people conducting it are incentivized to find complexity because complexity justifies a bigger engagement. So the deliverable becomes a laundry list of everything that could be improved, ranked by nothing in particular, with no connection to what the business actually needs in the next quarter.

I ran marketing at companies ranging from Fortune 200 to early-stage startups before starting my own firm. At one company, a 30-minute meeting with the CEO required two or three pre-meetings just to polish the deck. The decision was made in minutes. The deck went into a drawer. All those hours, gone.

That experience shaped how I think about audits. The output has to be a working document that becomes the blueprint for what happens next. Not a souvenir.

The AI-Assisted Audit Framework

Our audit covers three areas: the marketing org itself, the tech stack, and what I call AI readiness. That last one didn’t exist two years ago. Now it’s arguably the most important piece, because it determines how much of the roadmap a company can actually execute without hiring five more people.

Each area follows a specific process, and AI shows up differently in each one.

Phase 1: Intake And Context Building

Before we talk to anyone on the client’s team, we feed everything we can get our hands on into Claude. Investor decks. Board presentations. The company’s public marketing. Competitor creative. Job postings from the last six months. Glassdoor reviews. Product screenshots. Pricing pages.

Two years ago, synthesizing all of that required a senior strategist spending a full week reading, annotating, and building a briefing document. Now, we build a comprehensive context package in a day. Claude processes the raw material and produces a structured brief that includes the company’s positioning gaps, messaging inconsistencies across channels, competitive white space, and the questions we should be asking in stakeholder interviews.

The output isn’t a summary. It’s a diagnostic framework tailored to that specific company. We review it, challenge it, add our own operator instincts, and walk into discovery calls with a point of view instead of a blank notepad. That changes the conversation immediately. Clients notice when you’ve done the homework.

Phase 2: Tech Stack And Workflow Mapping

This is where things get specific. We pull a full inventory of all of the tools the marketing team uses. Customer relationship management (CRM). Email platform. Analytics. Attribution. Ad platforms. Content management. Design tools. Project management. The average mid-stage startup has between 15 and 30 marketing tools, and in almost every audit, at least a third of them overlap or go mostly unused.

We document every workflow: how a campaign goes from idea to live, how leads get routed, how reporting happens, who touches what, and when. Then we map each workflow against what’s now possible with AI-native alternatives.

A real example: One client had three people spending a combined 40 hours per week on creative production for paid social. Briefing a designer. Waiting for rounds of revisions. Resizing for different placements. Exporting. Uploading. We replaced that workflow with a combination of AI creative tools and a custom automation that handled asset generation, versioning, and platform-specific formatting. The same volume of creative now takes roughly eight hours of human time per week, and most of that is strategic review rather than production.

Tools like HeyGen and ElevenLabs handle video and audio production that used to require a studio. Custom AI agents built on open-source AI harnesses like OpenClaw and Hermes automate research, competitive monitoring, and content drafts. The point isn’t to name-drop software. It’s that the landscape of what can be automated has expanded dramatically in the last 18 months, and most marketing teams haven’t caught up.

Phase 3: AI Readiness Assessment

This phase is the one that surprises clients the most, because it’s less about technology and more about people.

We evaluate three things. First, does the team have the curiosity and willingness to adopt AI tools? Some teams are eager. Some are terrified. Knowing where people stand before you start pushing new workflows prevents the kind of resistance that kills transformation projects. I spoke about AI readiness to a group of senior marketers at a hyper-growth consumer app, and the first question asked was: “Isn’t the magic in our human work and interactions?” They were afraid.

Second, does the company’s data infrastructure actually support AI-driven optimization? If your CRM is a mess, your attribution is broken, and your analytics are built on vanity metrics, no AI tool is going to save you. Garbage in, garbage out still applies. We flag the data hygiene issues that need to be fixed before any AI implementation will produce reliable results. And the audit acknowledges the data gaps and how (and why) to fix them.

Third, where are the highest-leverage automation opportunities? Not everything should be automated. Creative strategy still requires human judgment. Brand decisions still need a human with taste and context. The audit identifies which workflows will benefit most from AI and which ones need a human firmly in the loop. AI readiness is not about replacing all humans with AI tools and agents.

What The Deliverable Actually Looks Like

We don’t hand over a deck. We produce a shared document with four sections: current state diagnosis, prioritized opportunity map, 90-day implementation roadmap, and a tool-by-tool recommendation list with estimated time and cost savings.

The roadmap breaks the 90 days into three phases. The first month focuses on quick wins, the workflows where AI can be plugged in with minimal disruption and immediate impact. Month two tackles the structural changes, things like rebuilding attribution models or redesigning the content production pipeline. Month three is about training and handoffs, ensuring the team can run the new systems independently.

The document is collaborative. Clients can comment, push back, and reprioritize. It becomes the working blueprint for the engagement, not a PDF that gets emailed and forgotten.

Where The Real Savings Show Up

The savings are rarely where people expect them. Most founders assume AI will cut their ad spend or reduce their agency fees. Sometimes it does. But the bigger wins tend to be in time recaptured.

A marketing team that was spending 60% of its week on production and reporting and 40% on strategy gets those numbers flipped. Humans focus on the work that actually requires taste, judgment, and relationship-building. The AI handles the repetitive execution that was eating their calendars.

One engagement reduced a client’s creative production cycle from three weeks to four days. Another automated their weekly reporting entirely, freeing up a senior analyst to focus on actual analysis instead of pulling numbers into slides. A third rebuilt their email lifecycle from scratch using AI-generated segmentation and content, which cut their cost per acquisition by 30% in the first 60 days.

None of those outcomes required firing anyone. They required moving people from low-leverage tasks to high-leverage tasks. That’s the part of the AI conversation that gets lost in the layoff headlines.

What I’d Tell Any Marketing Leader Reading This

You don’t need to hire a firm to start. Pick one workflow on your team that is repetitive, time-consuming, and doesn’t require deep creative judgment. Map it out step by step. Then ask whether an AI tool could handle any of those steps today.

Begin by tackling reporting. Next, focus on competitive research. Consider first-draft content production as an early win. Finally, initiate the process wherever the pain is loudest and the risk is lowest. Get a win. Show the team what’s possible. Then expand.

The companies that will struggle are the ones waiting for someone to hand them a playbook. The companies that will win are the ones running their own experiments right now, even clumsy ones, and learning what works inside their specific context.

The audit is just a structured way to do what every marketing team should already be doing: looking honestly at how time gets spent and asking whether there’s a better way. AI just made “better” a lot more accessible than it was 18 months ago.

More Resources:

Featured Image: Tetiana Yurchenko/Shutterstock