Tagline: Innovators don't need more AI. They need the skills to unlock its value.
NOVA trains founders and innovation teams with the workflows and systems that produce real results. Open-access workflows for founders. Keynotes and training for innovation hubs and universities.
Audience badge: For founders & innovation ecosystems.
Headline: Innovators don't need more AI. They need skills to unlock its value.
Subheadline: NOVA trains founders and innovation teams with the workflows and systems that produce real results.
Supporting line: Open-access workflows for founders. Keynotes and training for innovation hubs and universities.
Primary calls to action: Book a Call · Explore AI Workflows.
Section heading: Three ways to build AI capability.
The AI Workflow Library covers every stage of the founder journey. Exact prompts included. Start anywhere.
Available free at aibynova.com/workflows.
Section eyebrow: Our Story. Subhead: Access was never the hard part.
I built and licensed a digital health AI company, then spent years inside McMaster's health and life science incubator — working with founders, clinicians, and institutional leaders trying to make AI actually work.
The pattern was consistent. AI literacy without structure produces noise. With structure, it produces results.
We are on a mission to train 1,000 entrepreneurs by 2027. Every workflow, workshop, and keynote moves that number.
Closing line: You are building the future. Build the systems that make it last. — Shaaf Farooq, Founder.
St. Joseph's Healthcare Hamilton, Founder Institute, University Health Network, McMaster University, e-Health Conference, University of Toronto, University of Calgary.
NOVA AI is a training and advisory practice for AI literacy. The "product" is a system: open-access workflows, structured workshops, and ecosystem advisory engagements that move people and organizations from "AI curious" to "AI native."
Most AI training delivers excitement, not capability. People leave a webinar with screenshots and slogans, then revert to their old workflow within a week. NOVA exists because the gap between "having access to AI tools" and "using AI to produce real results" is now the bottleneck for innovation teams.
The NOVA approach treats AI literacy as a system, not a skill. That means: shared mental models, repeatable workflows, the right tool for the right task, and follow-through after the workshop.
Section heading: Access is not the problem. Knowing what to do with it is.
Your team has the tools. Without a shared framework, AI stays a productivity trick.
One session produces excitement. Without structure and follow-through, teams revert within a week.
Teams chasing new models without foundational judgment fall further behind with every update.
The organizations winning with AI have a system — not just access.
Headline: An open-access third space for founders.
We're building a space where founders can responsibly experiment with AI, learning by doing, together.
Goal: Train 1,000 entrepreneurs by 2027.
Progress to date: 75 founders trained.
One session. Your team goes from curious to capable and leaves with workflows they use the next day.
Built for hubs, conferences, and leadership teams. Clear, honest, and immediately actionable.
For hubs that want AI literacy built into their programming — not bolted on after the fact.
The flagship NOVA resource is a step-by-step library of AI workflows for founders. Each workflow specifies the exact tools, the order, sample prompts, pro tips, and the expected output. All workflows are free and open-access at aibynova.com/workflows.
Outcome: By the end of this workflow, you will have rehearsed a realistic stakeholder conversation and captured actionable insights from the real interview — without scrambling for notes.
Before you simulate anything, you need context. Open Perplexity and research the person you're about to meet. Look for their role, specialty, published work, recent talks, and the language they use in their field. The goal is to understand how they think and what they care about so your simulation is realistic.
Example prompt:
Search for [stakeholder name / role / organization]. I need to understand: their professional background, their specialty area, recent publications or talks, the key terminology used in their field, and the top 3 challenges professionals in their role typically face. Summarize everything in a brief I can use to prepare for an interview.
Pro tip: Spend at least 10 minutes here. The quality of your simulation in the next steps depends entirely on how much context you gather now. Don't skip this.
Take everything you found in Step 1 and feed it into Claude. Ask Claude to become this person — not a simplified version, but a realistic, challenging version that uses real terminology and pushes back the way a real expert would. This is how you stop showing up to meetings sounding like a student who needs everything explained.
Example prompt:
You are a [stakeholder role, e.g., orthopedic surgeon / physiotherapist / health tech investor] with [X years] of experience in [specialty/field]. Here is background information about you: [paste research from Step 1]. From now on, respond as this person in a realistic way. Use industry-specific jargon. Do not simplify your language. If I ask a vague or naive question, push back the way a real [stakeholder role] would. If I demonstrate knowledge of your field, engage more deeply. I'm a founder building [your product/idea description] and I want to have a peer-level conversation with you.
Pro tip: Feed Claude their actual LinkedIn profile, any published papers, or conference talks you found. The more specific the persona, the more useful the simulation. Generic personas give generic answers.
Open ChatGPT, switch to voice mode, and paste in the persona context you built in Claude. Run through your interview questions as if this is the real meeting. This is where you catch bad habits — filler words, hesitation, weak questions — before the real call.
Example prompt:
I'm about to practice a stakeholder interview. You are going to play the role of [stakeholder role] with the following background: [paste persona from Step 2]. I will ask you questions as if this is a real meeting. Stay in character the entire time. Respond naturally and realistically. If my questions are too vague, tell me. If I use incorrect terminology, correct me. If I ask something insightful, engage deeper. After we finish, break character and give me feedback on: the quality of my questions, areas where I seemed unprepared, and suggestions for how to improve.
Pro tip: Do this at least twice. The first run is always rough. By the second run, you'll notice a dramatic difference in your confidence and fluency with the terminology.
Open Granola before the call starts and let it run silently in the background. It captures the full conversation automatically. Your only job is to be fully present — listen, ask great follow-ups, and have a real conversation. Don't take notes. Granola handles it.
Pre-meeting note to set in Granola:
Meeting with [stakeholder name], [role] at [organization]. Purpose: customer discovery interview for [your product/idea]. Key topics to cover: [list 3-5 topics]. After this meeting, I need to extract: key pain points mentioned, direct quotes, objections or concerns, and recommended next steps.
Pro tip: After the call, pull exact quotes from Granola's transcript. Direct stakeholder quotes are 10× more powerful than your paraphrasing — use them in your pitch deck, PRD, and investor conversations.
What you walk away with: a detailed stakeholder persona you can reuse for future prep; practice handling tough questions and industry terminology; a full transcript of your real interview with zero note-taking during the call; extracted quotes, pain points, and next steps ready to use in your pitch, PRD, or next meeting.
Outcome: By the end of this workflow, you will have pressure-tested your startup idea from every angle and have a clear, honest assessment of its strengths, weaknesses, and unknowns — before you write a single line of code.
Before you can stress-test anything, you need to articulate your idea clearly. Open ChatGPT and describe your idea, then ask it to sharpen your description into a single paragraph covering what it is, who it's for, and why it matters. This becomes the anchor for everything that follows.
Example prompt:
I have a startup idea and I need to articulate it clearly before I start validating it. Here's my rough description: [describe your idea in your own words, even if it's messy]. Rewrite this into one clear paragraph that covers: (1) what the product is, (2) who it's for, (3) what problem it solves, and (4) why now — what's changed that makes this possible or necessary today. Keep it under 100 words. Use simple language. No buzzwords.
Pro tip: If you can't describe your idea clearly in one paragraph, you're not ready to validate it. This step alone saves founders weeks of confusion.
Ask ChatGPT to generate every critical question you should answer before investing time and money into building. These questions become your validation research checklist.
Example prompt:
Here's my startup idea: [paste your one-paragraph description from Step 1]. Generate 20 critical questions I must answer before I build anything. Cover these categories: market size and demand, existing competition and alternatives, technical feasibility, regulatory or compliance risks, customer willingness to pay, go-to-market strategy, and key assumptions I might be making without realizing it. Number each question and organize by category.
Pro tip: Don't edit the questions to make yourself feel better. If a question makes you uncomfortable, that's the one you need to answer most.
Take each question from Step 2 and run it through Perplexity one by one. Perplexity searches live sources and gives you cited answers. Copy each answer into a running document so you build up a comprehensive research file.
Example prompt:
[Paste question from Step 2] — specifically in the context of [your industry/market]. Provide data, statistics, named competitors, and cite your sources.
Pro tip: Use Perplexity for this, not Claude or ChatGPT. Perplexity cites real sources and searches live data. Claude and ChatGPT will give you confident-sounding answers that may be outdated or fabricated.
Take everything you've gathered and feed it into Claude. Ask Claude to be your harshest critic. The goal is to find every weak spot, hidden assumption, and blind spot before a real investor or customer does.
Example prompt:
You are a seasoned venture capitalist who has reviewed 1,000+ startup pitches and invested in fewer than 50. You are known for being brutally honest and finding the weaknesses founders can't see. Here is a startup idea and the research gathered so far: [paste your one-paragraph idea description] [paste all research from Step 3]. Now do the following: (1) List every hidden assumption this founder is making. (2) Identify the top 5 risks that could kill this startup. (3) Point out what's missing from the research — what questions weren't asked that should have been. (4) Give a brutally honest assessment: on a scale of 1-10, how validated is this idea right now? What needs to happen to move that number up? Do not be encouraging. Be honest.
Pro tip: If Claude agrees with everything and says your idea is great, your prompt wasn't challenging enough. Add: "Do not give me encouragement. I need you to find problems. If this idea has a fatal flaw, I need to know now."
Upload everything — your idea description, research questions, Perplexity findings, and Claude's critique — into NotebookLM. This becomes your living validation document you can query anytime instead of digging through scattered notes.
Example prompt:
Based on everything uploaded, create a structured validation summary with these sections: (1) Idea overview, (2) Market evidence — what data supports demand, (3) Competition landscape — who else is doing this, (4) Key risks and assumptions flagged, (5) Open questions still unanswered, (6) Recommended next steps to further validate. Keep it concise and reference specific findings from the uploaded documents.
Pro tip: NotebookLM lets you ask follow-up questions across all your uploaded docs. Anytime a mentor or investor challenges your idea, search it in NotebookLM first.
What you walk away with: a crystal-clear one-paragraph description of your idea; a comprehensive research file covering market, competition, feasibility, and risks; a brutally honest critique identifying your blind spots and hidden assumptions; a structured validation summary you can reference, share with co-founders, or bring to a mentor meeting; a living NotebookLM workspace you can query anytime someone challenges your idea.
Outcome: By the end of this workflow, you will have a live, branded website that matches your pitch deck — something you can send to investors, customers, or partners instead of a static PDF.
Upload your pitch deck to Claude and ask it to extract everything needed for a website — your value proposition, messaging, section structure, and brand identity. This ensures the website Lovable builds actually matches your deck instead of looking like a generic template.
Example prompt:
Here is my startup pitch deck: [upload PDF or paste slide content]. Extract the following into a structured website brief: (1) Company name and tagline, (2) One-sentence value proposition, (3) Problem statement in customer-facing language, (4) Solution description — what the product does, (5) Key features or benefits as bullet points, (6) Target audience, (7) Team section — names and roles, (8) Brand colors mentioned or visible in the deck, (9) Suggested website sections in order, (10) Call-to-action — what should a visitor do. Format this as a clean brief I can hand to a website builder.
Pro tip: The better this brief is, the better Lovable's output will be on the first try. Spend a minute reviewing it and fixing anything Claude got wrong before moving to Step 2.
Take the brief from Step 1 and feed it directly into Lovable. Lovable will generate a fully functional, branded landing page based on your instructions. You describe what you want in plain English and it builds it.
Example prompt:
Build me a startup landing page with the following specifications: [paste entire brief from Step 1]. The design should feel clean, modern, and professional. Use the brand colors specified in the brief. Include these sections in order: (1) Hero with company name, tagline, value prop, and CTA button, (2) Problem section, (3) Solution section with key features, (4) How It Works — 3-step breakdown, (5) Team section, (6) Footer with contact and social links. Make it responsive for mobile.
Pro tip: Don't try to make it perfect on the first generation. Get the structure right first, then iterate on design details in Step 3.
Look at what Lovable built and tell it what to change in plain English. Adjust colors, swap sections, rewrite copy, add images. You can do as many rounds as you need.
Example prompt:
Change the hero background to [color]. Make the CTA button larger and [color]. Rewrite the problem section to be shorter — max 2 sentences. Add a testimonial section after the solution section with placeholder quotes. Move the team section above the footer. Add a subtle animation when the page loads.
Pro tip: Show the website to one person before you finalize it. Ask them: "What does this company do?" If they can't answer in 5 seconds, your hero section needs work.
Ask Lovable to add a functional form. This turns your website from a brochure into an actual lead capture tool.
Example prompt:
Add an email signup form in the hero section and the footer. When someone enters their email and clicks [Join Waitlist / Get Early Access], store their email in a database and show a confirmation message: "You're in! We'll be in touch soon." Also add a small counter above the form that says "Join [X]+ founders on the waitlist."
Pro tip: Send this link to 5 people today. A live website with a waitlist is 10× more impressive than a pitch deck PDF.
What you walk away with: a live, branded website that matches your pitch deck's messaging and visual identity; a functional landing page with hero, problem, solution, team, and CTA sections; a working email signup form connected to a database; a shareable link you can send to investors, customers, and partners instead of a static PDF.
Outcome: By the end of this workflow, you will have a clear system for when to use Claude vs. other tools — so you stop wasting Claude credits on tasks other tools do better for free.
Any question where you need factual information, market data, competitor analysis, or current trends should go to Perplexity — never Claude. Perplexity is built for research, cites real sources, and has no meaningful usage limits. Every research question you send to Claude instead of Perplexity is a wasted credit.
Example prompt:
[Your research question] — provide data, statistics, named sources, and recent information. Cite everything.
Pro tip: If your question starts with "What is...", "How many...", "Who are the competitors in...", or "What does the market look like for..." — that's a Perplexity question. Not a Claude question.
Brainstorming names, generating lists, drafting rough outlines, quick summaries — send it to ChatGPT. ChatGPT has virtually no usage limits on paid plans and handles these tasks well. Don't use Claude for drafts. Use Claude for finals.
Example prompt:
I need to brainstorm [topic]. Give me [number] options. Don't overthink it — I want quantity and variety right now, not perfection. I'll refine later.
Pro tip: ChatGPT is your brainstorming partner you can talk to all day. Claude is your senior consultant you bring in for the final deliverable. You don't waste consultant hours on brainstorming.
Before you use a single Claude credit, have everything ready. All your research from Perplexity. Rough drafts from ChatGPT. Meeting notes from Granola. Organize it all into one clean paste-ready block. One well-structured Claude prompt with full context should get you 90% of the way there on the first try.
Input template:
TASK: [What I need Claude to produce] CONTEXT: [Background on my startup / project / situation] RESEARCH: [Paste Perplexity findings] ROUGH DRAFT: [Paste ChatGPT brainstorm or outline if applicable] REFERENCE MATERIALS: [Paste any relevant docs, notes, transcripts] OUTPUT FORMAT: [Exactly how I want the final product structured] CONSTRAINTS: [Word count, tone, audience, specific requirements]
Pro tip: One comprehensive Claude prompt beats 15 back-and-forth messages. Every follow-up message costs credits. Front-load everything.
Now — and only now — open Claude. Paste your organized input block from Step 3 and ask for the final deliverable. This is where Claude shines: polished writing, strategic analysis, code, PRDs, investor-ready documents. Because you've done all the prep work, Claude should nail it in one or two prompts.
Example prompt:
You are a [relevant expert role — e.g., senior product strategist / startup advisor / technical architect]. I need you to produce [specific deliverable]. Here is everything you need: [paste your organized input block from Step 3]. The output should be [format — e.g., a 2-page PRD / a 500-word investor email / a full technical spec]. Constraints: [any specific requirements]. Quality standard: this needs to be production-ready — something I can send directly to [investor / customer / team] without editing. Do not include filler or generic advice. Every sentence should be specific to my situation.
Pro tip: After Claude gives you the output, be surgical with revisions: "In paragraph 2, change the tone from formal to conversational" beats "make it better." Specific revision prompts cost fewer credits than vague ones.
What you walk away with: a clear personal system for routing tasks to the right tool; Perplexity handles your research (free, unlimited, cited); ChatGPT handles your brainstorming and drafts (virtually unlimited); Claude handles only your final, high-stakes deliverables; an organized input template you can reuse every time you go to Claude; dramatically fewer wasted Claude credits and higher quality output per credit spent.
Outcome: By the end of this workflow, you will have conducted a well-prepared discovery call and walked away with organized, quotable insights you can immediately use in your pitch, PRD, or next team meeting — without having taken a single note during the call.
Open Perplexity and research the person you're about to speak with and their organization. Understand their role, what they care about, their industry's pain points, and any recent news. The goal is to walk into the call informed enough to skip surface-level questions and get to the real stuff faster.
Example prompt:
I'm about to have a customer discovery call with a [stakeholder role] at [organization/company type]. Research and tell me: (1) What are the top 3-5 pain points professionals in this role typically face? (2) What tools or processes do they currently use? (3) What recent trends or changes are affecting their work? (4) What would make them skeptical of a new product in this space? Cite your sources.
Pro tip: Even 10 minutes of research completely changes the quality of your questions. Stakeholders can tell immediately whether you've done your homework.
Feed Claude your research from Step 1 plus context about your startup. Ask it to generate a customized interview script with open-ended questions designed to surface pain points, current workflows, and willingness to change. This isn't a rigid script — it's a guide that makes sure you cover what matters.
Example prompt:
I'm a founder building [product description] for [target market]. I'm about to conduct a customer discovery interview with a [stakeholder role] at [organization type]. Here's what I know about them: [paste research from Step 1]. Generate a discovery interview script with: (1) A 30-second opening explaining who I am and why I'm talking to them, (2) 10-12 open-ended questions covering their daily workflow, biggest frustrations, current tools, what they've tried before, and what would make them adopt something new, (3) 3 follow-up probes for surface-level answers, (4) A closing question. No yes/no questions. Order from broad to specific.
Pro tip: Always end with: "Is there anything I should have asked you that I didn't?" This question reliably surfaces the most useful insight of the entire call.
Your interview script from Step 2 is thorough but too long to glance at during a live call. Ask ChatGPT to compress it into a one-page cheat sheet you can keep open on a second screen. This keeps you present in the conversation instead of reading from a document.
Example prompt:
Take this full discovery script: [paste Step 2 output]. Compress it into a one-page cheat sheet I can glance at during the call. Format: short bullet points only, 3-4 words per question, grouped by theme (e.g., "Workflow", "Pain Points", "Current Tools", "Adoption"). Add a tiny section at the bottom labeled "If they say X, ask Y" for the most likely answer threads.
Pro tip: Print this or open it on a second monitor. Don't share your screen with it visible — it should look like you remembered everything.
Open Granola before the call starts. Let it run silently. Your only job is to be fully present — listen, ask great follow-ups, and have a real conversation. Don't take notes. Granola handles the entire transcript and structured summary automatically.
Pre-meeting note in Granola:
Customer discovery call with [name], [role] at [organization]. Goal: understand their workflow, pain points, current tools, and openness to a new solution for [problem area]. After the call, extract: top 3 pain points, exact quotes, current tools they mentioned, objections, and recommended next steps.
Pro tip: If the call goes off-script in an interesting direction, follow it. The script is a safety net, not a cage. Granola captures everything regardless.
Take Granola's transcript and feed it into Claude with a clear extraction prompt. You'll get back a tight, organized summary with direct quotes you can drop straight into your pitch deck, PRD, or investor update.
Example prompt:
Below is a transcript from a customer discovery call: [paste Granola transcript]. Extract the following: (1) Top 3 pain points the interviewee described — in their own words where possible. (2) 3-5 direct quotes I could use in my pitch deck or investor update. (3) Tools or workarounds they currently use. (4) Objections or concerns they raised about a new solution. (5) The single most important thing I learned. (6) Recommended follow-up actions. Be concise. Use bullet points. Quote directly when possible.
Pro tip: Direct customer quotes are the single most powerful asset you can put in front of an investor. Build a running document of quotes from every interview — that's your social proof bank.
What you walk away with: a researched, tailored interview script for the specific person you're meeting; a one-page cheat sheet you can glance at during the call; a full transcript captured automatically by Granola; a structured insight summary with direct quotes ready to use in your pitch, PRD, or investor update; a growing library of customer quotes you can reference for life as a founder.
Tagline: Turn your team into confident AI operators.
A structured, hands-on training curriculum built for real organizational needs. Not another AI webinar — a transformation program.
Build a shared understanding of what AI can (and can't) do. Reframe how your team thinks about automation, workflows, and leverage.
Move from theory to execution. Your team gets direct experience building with AI tools tailored to your actual workflows and challenges.
Turn skills into systems. We help you map AI across your operations and create a sustainable playbook your team owns.
Headline: Shaaf unlocks AI potential.
Custom keynotes for hubs, conferences, and leadership teams. National AI coach & keynote speaker · Practical AI upskilling · Senior AI advisor.
Inquiries: shaaf@aibynova.com or hello@aibynova.com.
Shaaf Farooq — Founder of NOVA AI. National AI coach and keynote speaker.
Shaaf built and licensed a digital health AI company, then spent years inside McMaster's health and life science incubator working with founders, clinicians, and institutional leaders trying to make AI actually work.
LinkedIn: linkedin.com/in/shaaf-farooq.
Email: hello@aibynova.com
Keynote inquiries: shaaf@aibynova.com
Web: https://aibynova.com
This page is a static, server-delivered HTML document maintained for AI assistants, search-engine crawlers, and link previewers. All content above is hardcoded — no JavaScript is required to render it, and no client-side data fetching is performed. For the interactive site, visit aibynova.com.