Frequently Asked Questions About Greg Isenberg Tiny AI Agent Business Builder
23 answers covering everything from basics to advanced usage.
// Basics
What exactly is the one-liner in Greg Isenberg's AI agent framework?
The one-liner is a single compressed sentence that defines your entire business brief for the AI agent. It names the data feed, asset type, scoring criteria, delivery channel, and end buyer. For example: 'Monitor expired domain drops for domains with DR 20+, clean backlink profile, under $200 — deliver 10 picks to Slack each morning — flip to newsletter operators.' Writing this before touching any tool is the mandatory first step. It forces clarity and prevents the agent from producing generic, unfocused output.
What does 'agents are the new SaaS' mean in Greg Isenberg's framework?
It means shifting from a per-seat software subscription model to an outcome-based model. Instead of building a dashboard that customers must interpret themselves, you sell an AI agent that delivers a specific result — like a daily deal card or competitive intelligence brief — directly to the buyer. The customer pays for the outcome, not for access to a tool. This lowers the barrier to entry for the builder and increases perceived value for the buyer.
What is an obvious buyer and why does it matter?
An obvious buyer is a specific, reachable person or business with money who will predictably pay for the asset or intelligence the agent surfaces. Examples: newsletter operators for premium domains, new restaurant owners for used kitchen equipment, marketing agencies for hiring-signal leads. Without an obvious buyer, the Feed → Asset → Trigger chain is worthless. Identifying who pays first — before committing time or capital — is the single most important validation step in the framework.
Is this framework only for tech-savvy people?
No. The framework is designed for people with no technical background. You interact with the AI agent in plain conversational language — exactly as you would instruct a human assistant. The one-liner is written in natural language, not code. Setting up a Slack webhook is the most technical step, and the agent can walk you through it. If you can write a sentence describing what you want and answer clarifying questions, you can build a tiny AI agent business.
How much does it cost to run an AI agent for this framework?
Costs vary by agent platform and usage. With heartbeat off and daily scheduled runs, token costs are minimal — typically a few dollars per month. Heartbeat mode (checking every 30 seconds) increases costs significantly and should only be enabled once the business generates revenue. The main costs are the AI platform subscription, any API access fees for data sources, and your acquisition budget if you're flipping assets. Many broker-model businesses have near-zero ongoing costs.
// How To
How do I write a good one-liner for my AI agent?
Follow this template: 'Monitor [feed source] for [asset type] that meet [scoring criteria], flag when [trigger event], deliver ranked list to [channel] so I can [monetization method] to [buyer type].' Be specific about the data source and scoring but don't over-specify. If you're unsure about criteria thresholds, let the agent suggest defaults. The one-liner should fit in one sentence and immediately tell the agent what to do, who to serve, and how you make money.
How do I set up a Slack webhook for my AI agent?
Create a dedicated Slack channel for your business idea — never mix multiple agents into one channel. Go to api.slack.com/apps, create a new app, and add an Incoming Webhook. Copy the webhook URL and paste it into your AI agent's configuration. If you're unfamiliar with the process, ask your AI agent to walk you through it step by step — it will produce detailed instructions. Test by sending one deal card before enabling scheduled runs.
How do I choose between flip, broker, retainer, and relaunch monetization?
Flip if you want to buy low and sell at market value — requires capital but highest margins. Broker if you want zero inventory risk — connect seller and buyer for a 15–30% fee. Retainer if you want recurring revenue — sell the daily intelligence brief as a subscription. Relaunch if you want to acquire dead assets, improve them, and grow over time. Pick exactly one per agent. Mixing methods creates confusion and dilutes focus.
How do I validate my tiny AI agent business idea before building?
Run it through three screening questions: (1) Is there urgency? — will the asset be claimed by others if you don't act quickly? (2) Is there spread? — is the gap between acquisition cost and resale value large enough to be worth acting on? (3) Who pays first? — can you identify a specific, reachable buyer before committing capital? If any answer is no, refine or discard the idea. Then check the five-node chain: Feed → Asset → Trigger → Buyer → Monetization must all be present.
// Troubleshooting
My AI agent stopped producing results — is it broken?
Not necessarily. Agents can go quiet when they are autonomously reconfiguring or rebuilding themselves. Check the app status and logs before troubleshooting. If the agent is truly stuck, restart it and paste your one-liner again. Also verify that your data sources haven't changed their structure (e.g., a website redesign breaking the scraper). Common causes of silence include expired API keys, changed feed URLs, or the agent filtering out all results due to overly strict exclusion criteria.
My deal cards have HTML entities and broken links — how do I fix this?
This is a common first-batch bug. Tell the agent conversationally: 'The links in the cold emails have HTML entities bleeding through — fix the formatting so plain URLs appear cleanly.' The agent updates its own logic without you rewriting the original prompt. Always review the first batch of outputs manually before enabling automated outreach. Catching formatting bugs at low volume prevents reputation damage at scale.
What if I can't identify an obvious buyer for my niche?
If you cannot name a specific, reachable buyer with money before building, the idea is not ready. Go back to the three-lens filter: find where constant change meets neglected assets meets a liquid buyer. Try asking the AI agent to suggest buyer personas for your asset type. If no buyer emerges, switch niches. A brilliant feed-to-asset pipeline is worthless without someone who will pay on receipt of the deal card.
How do I prevent my AI agent from surfacing low-quality or irrelevant assets?
Set clear scoring criteria in your one-liner and use exclusion filters. For domains: minimum DR threshold, clean backlink profile, no adult or gambling history. For equipment: minimum spread percentage between acquisition and resale value. For hiring signals: specific role titles that indicate budget allocation. Review the first 10–15 results manually and tell the agent conversationally what to fix: 'These are too low-quality — raise the DR threshold to 30.' The agent refines its logic with each correction.
// Comparisons
How is this different from just using ChatGPT to do market research?
ChatGPT gives you a one-time research output. The Tiny AI Agent Business Builder creates an always-on, recurring system that monitors live data feeds, scores assets against your criteria, and delivers actionable deal cards on a schedule. The key difference is automation and recurrence — the agent runs while you sleep and produces a daily monetizable brief. A single ChatGPT session is a research tool; a configured agent running on a schedule is a business.
How does this compare to buying a franchise or an existing small business?
A franchise requires $50K–$500K+ upfront, a lease, employees, and months of setup. The Tiny AI Agent Business Builder targets hours to first revenue with near-zero capital. You deploy an AI agent to do the monitoring work instead of hiring staff. The tradeoff: franchises offer proven playbooks and brand recognition; tiny AI agent businesses offer speed, flexibility, and the ability to test multiple ideas simultaneously at negligible cost. They complement rather than replace each other.
How does this framework compare to traditional affiliate marketing?
Affiliate marketing requires building an audience, creating content, and waiting for commissions on other people's products. The Tiny AI Agent Business Builder cuts out the audience-building step entirely — you go directly to an obvious buyer with a specific deal card. You earn through flips, broker fees, or retainers rather than referral commissions. The margin is typically higher (15–30% broker fee vs. 5–15% affiliate commission) and the time to first revenue is hours, not months.
// Advanced
Can I run multiple tiny AI agent businesses simultaneously?
Yes, but keep each one in its own dedicated delivery channel. Never mix multiple agent outputs into a single Slack channel — it creates noise and kills your ability to act quickly on individual deal cards. Each agent should have its own one-liner, its own channel, and its own locked-in monetization method. Start with one, validate quality, then clone the process for additional niches. Think of each agent as a separate micro-employee with a single job.
When should I turn on the heartbeat feature for my agent?
Only after the business generates revenue. Heartbeat checks for pending events every 30 seconds, which consumes tokens continuously. Before revenue, it's a cost sink. Start with scheduled daily runs — the agent delivers a ranked brief at a fixed time each morning. Once you've validated that deal cards convert to cash and you need near-real-time monitoring (e.g., time-sensitive domain auctions), enable heartbeat. Think of it as upgrading from daily newspaper delivery to a live news ticker.
Can I productize my AI agent's output and sell it as a subscription?
Absolutely — this is the 'agents are the new SaaS' model. Once your agent produces validated, high-quality daily briefs, you can sell access to those briefs as a subscription. For example, a competitive intelligence brief for a specific vertical at $9.99/month. Build a simple landing page, use the deal card as your sales artifact for inbound prospects, and deliver the brief to subscribers' Slack or email. You're selling an outcome, not a software seat.
What data feeds work best for tiny AI agent businesses?
The best feeds are public, machine-readable, and update frequently. Proven sources include: expired domain drop lists, GoDaddy auctions, business-for-sale marketplaces, restaurant closure listings, bankruptcy court filings, job board postings, app store rankings, Product Hunt launches, competitor pricing pages, changelogs, and founder social media posts. The feed must exhibit constant change — if it updates rarely, the agent has nothing to monitor and the business can't recur.
What's the difference between a deal card and a landing page in this framework?
A deal card is your outbound sales artifact — a structured record showing acquisition price, resale value, spread, broker fee, contact info, and draft outreach message. You send it directly to an obvious buyer. A landing page is your inbound marketing asset — it attracts prospects who want the agent's intelligence delivered to them regularly as a subscription. The deal card converts individual transactions; the landing page converts recurring subscribers. Build the deal card first, the landing page second.
What happens if someone else is running the same AI agent business as me?
Competition is less of a concern than you'd think because the framework emphasizes niche specificity. Your scoring criteria, exclusion filters, geographic focus, and buyer relationships create differentiation. Two people running expired domain agents will have different DR thresholds, different buyer networks, and different keyword niches. Speed also matters — the agent that surfaces the deal first and has the buyer relationship wins. Focus on a specific micro-niche rather than broad markets.
Can I use this framework to build a business that eventually runs without me?
Yes, that's the goal. Once you've validated deal card quality, enabled scheduled runs, and locked in a delivery channel, the agent runs autonomously. For broker and flip models, you still need to close transactions manually. For retainer or subscription models, the agent can deliver briefs to paying subscribers with minimal intervention. The more you refine the agent conversationally, the less manual oversight it needs. Full automation is the end state, not the starting point.