Simple AI Business Growth Strategy You Didn't Know Existed

Full video transcript

Ashley Gross joined the session as a featured speaker, introduced as a recognized leader in delivering generative AI solutions for mid-market to enterprise companies. She opened by explaining her background: she started using AI in 2020 after becoming a new mother and needing to condense a 40-hour work week down to 15 hours, which led her to generative AI tools like Jasper. In 2022, while working as a marketer at an enterprise company, she was asked to roll out AI across a marketing organization of over 100 people — months before ChatGPT existed publicly. Her approach was to automate the tasks people hated doing most. Three months in, the team overachieved its $90 million pipeline goal, hitting $115 million, giving her a clear revenue-tied ROI for AI.

From there she built her own AI-first business, with the legal setup complete in 10 days. She now teaches a ten-module course on building a business with AI, and built a custom GPT with embedded prompts covering the first four modules: identifying and validating a business idea, mapping personal passions and skills against market opportunity, and building a business plan with SMART goals. She stressed that a solid business plan — including understanding how much money you need to break even in a given time frame — is the foundation before any scaling effort.

To grow the business itself, she took inventory of tasks she didn’t want to do (manual data entry, calendar syncing, back-and-forth scheduling for speaking engagements) and automated those first. Instead, she spent her time on free 15-minute coffee chats with people in her industry on LinkedIn. She put a paywall on her scheduling tool so that even discovery calls were paid, which let her generate revenue from day one and reach $10,000 in monthly recurring revenue within 30 days.

She uses Read AI as her notetaker for both networking chats and client calls. Beyond transcription, it sends recap agendas with action items after each call and audits her own communication — talking pace, filler words, interruptions — as well as signals like whether the other person seemed engaged. She credits free PR from podcasts and speaking engagements (which led to coverage in Forbes, Fast Company, and MSN) as a low-cost way to build visibility without paid advertising.

On her website, a cookie-consent icon signals that visitor data is being collected through RB2B, which is configured with her ideal-customer parameters (title, tech stack, etc.). When a matching visitor arrives, she gets a real-time Slack alert and text message. One minute later, that visitor receives a LinkedIn connection request from her; 24 hours after that, a personal video DM (explicitly not an AI avatar or voice clone) thanking them for connecting. All of the resulting engagement data — chatbot conversations, form fills, downloads — flows into HubSpot (which she runs on the free tier), so that by the time a prospect books a call, she already has full context on their journey rather than opening with generic assumptions about their role. Make.com orchestrates the automation between these tools, and she runs a self-sabotage AI agent that deliberately tries to break her Make workflows twice a month so she can fix issues before they affect customers. She reports roughly 70-80% conversion from these tools individually, and 99.2% when used together.

She closed the main talk by emphasizing that the point of all this automation is not to replace humans or make the experience feel like AI, but to make every touchpoint feel personal and consistent with her brand. Her core metrics are customer retention and high-intent prospects visiting her site and converting — not vanity metrics like MQLs or SQLs.

In the Q&A, she elaborated on several points. Her ICP took five to six days to define and has been iterated three times over seven months, using AI tools (including Jasper against Google Analytics data) to analyze which geographies and messages were converting. On tooling choices, she said she prefers deep mastery of a small, well-understood tech stack over constantly adopting new tools, in part so her team can be onboarded quickly and so she knows where to troubleshoot when tools break. She’s on HubSpot’s free tier and used Read AI’s free tier for about a year and a half before upgrading. She doesn’t have strong recommendations for AI/automation communities, since she prioritized learning tools herself before delegating or outsourcing.

On retention, she described a screening process before accepting discovery calls: prospects must pay to book, then answer 15-20 short-answer questions about their business and the specific problem they want AI to solve — no multiple choice, so they engage seriously and arrive self-aware about their gaps. She said this filtering, combined with prioritizing relationship-building over pitching, is why she reports no customer churn and why all of her leads have been inbound.

On content, she said she doesn’t generate her copy with AI. Instead, she uses Canva templates for LinkedIn posts and a custom GPT with different “personas” (strategist, creator, assistant) that reformat her own ideas into a consistent structure — a punchy opening line, two to three summary sentences, and a clear call to action — rather than generating content from scratch. She’s less enthusiastic about TikTok, citing privacy concerns as a parent, but continues investing in long-form YouTube content despite the platform favoring shorter videos.

Asked about her direction for 2025, she said she wants to expand beyond offering AI implementation services to businesses and start helping other consultants — people with deep domain expertise in other fields — get their first clients and build confidence using AI in their own areas.

Her closing advice for aspiring AI consultants: don’t try to compete as a generalist AI expert. Instead, integrate AI into the industry expertise and workflow understanding you already have, since understanding a business’s actual friction points is a skill set most purely AI-focused competitors lack.