Sources and provenance

#Which lessons come from which research pass

From the original deep-research brief (the same one Income Playbooks cites for its own queued "AI consulting/agency" module): The mechanism, Three paths compared, Market and regulatory reality, Capital and timeline, Feasibility for a non-resident operator (generalized from an originally individual-operator-specific framing — the underlying facts are unchanged, the "you" is now general rather than naming a specific person or city), Launch sequence, Tools and vendor stack, KPIs and kill switches, Failure modes, and Path to scale.

This build's own supplementary research (web search, August 2026, added specifically to cover the social/marketing-consulting angle the original brief only touches in passing): AI social and marketing consulting in full, plus the social/marketing cross-references added into the lessons above.

#Sources — original research brief

Established (disclosed operator/primary data):

• FTC, "FTC Announces Crackdown on Deceptive AI Claims and Schemes" (Operation AI Comply, Sept 2024) — ftc.gov
• FTC, "FTC Finalizes Order with DoNotPay..." (Feb 2025) — ftc.gov
• PwC, "2026 AI Jobs Barometer" (UAE analysis)
• Glassdoor, live "prior authorization" remote job-posting count (Aug 2026)
• CloudZero, "Claude pricing in 2026" (Anthropic API rate breakdown)

Directional (consistent secondary-source pattern, no single primary disclosure):

• Fast Company Middle East, CXO Insight Middle East, and Khaleej Times, all citing the same Qlik-commissioned UAE/Saudi AI-adoption survey (2026) — the specific percentages are real and disclosed, and independently reported by multiple named trade outlets, but they all trace to one underlying report commissioned by a data-integration/AI vendor with a commercial stake in the "data and governance gaps are the obstacle" framing. Downgraded from an earlier Established tagging specifically for that reason — see Market and regulatory reality for the full caveat.
• Indie Hackers, an operator account of using an owned agency as both testing ground and distribution channel — a single self-reported, non-independently-verified anecdote, not disclosed operator figures a third party could check. Downgraded from an earlier Established tagging for the same reason — see Capital and timeline.
• ainora.lt, AI voice-agent cost-per-minute comparison (2026)
• lowcode.agency, n8n pricing (2026)
• arnifi.com, freezone license cost guide (2026)
• open-entity.com, Stripe non-US-entity blocking explainer (2026)
• Truckstop and Okoone, freight-broker AI workflow trade coverage (2026)
• Accounting Today and CPA Trendlines, CPA-firm agentic-AI trade coverage (2026)
• Onix, healthcare revenue-cycle-management AI writeup (vendor-authored — performance-percentage claims in it are treated as unverified)
• mldeep.io, AI-readiness-assessment scope and deliverables reference

Speculative / marketing-tier (used only to triangulate ranges, never as standalone evidence): assorted 2026 "AI consulting cost guide" SEO content, cross-checked only against sources above; YouTube commentary on the "95% of AI agencies dead by 2026" framing, treated as rhetorical rather than a named study.

Explicitly flagged as unverified by the original research: a "~64% renewal / ~18-month" fractional-AI-engineer retention statistic that circulates in some marketing content could not be traced to any primary disclosed source. It does not appear as a claim anywhere in this course's lessons; it's named here only so you recognize and discount it if you see it elsewhere.

#Sources — this build's supplementary research (social/marketing angle)

Established (confirmed directly against a primary source):

• FTC, Endorsement Guides — the 2023 revision expanding "endorser" to cover AI-generated and synthetic personas — ftc.gov/business-guidance/advertising-marketing/endorsements-influencers-reviews
• FTC, Consumer Reviews and Testimonials Rule (16 CFR Part 465), effective September 2025, covering AI-generated fake reviews — ftc.gov (final-rule text)

Directional (consistent pattern across multiple vendor/trade sources, no single primary disclosure):

• Pricing and market-positioning writeups: digitalagencynetwork.com, outreachz.com, taskip.net, apaya.com (AI marketing/social-agency pricing benchmarks, 2026). Two of these four are themselves SaaS vendors selling into this exact market — taskip.net (agency-management software) and apaya.com (AI social-media-management platform) — publishing "what to charge" content aimed at the same buyer their own tool is marketed to. Their numbers are used only because they're corroborated by the other two, non-vendor sources; see the caveat inline in The social and marketing angle.
• Tooling landscape: sintra.ai, analyticsinsight.net, monday.com, enrichlabs.ai, blotato.com (AI social-media-management and automation tool roundups, 2026)
• Positioning and niche strategy: lilachbullock.com, mayankunfiltered.com (how to start an AI/social-media marketing agency, 2026). Lilach Bullock markets herself as an "AI Implementation Consultant" selling her own consulting services — this course draws on her content only for the generic, low-stakes "specialize rather than generalize" argument, not for any of her specific unsourced pricing or hours-saved figures, which are not used anywhere in this course and should be treated as consultant content marketing if you encounter them directly on her site.
• AI UGC/avatar video: influencermarketinghub.com, cometly.com, greyjournal.net, shhots.ai (AI UGC ad-generator and avatar-tool roundups and creator-economy pricing, 2026). shhots.ai is itself an AI-UGC tool vendor selling subscriptions and a done-for-you package into the same "AI-UGC operator" opportunity its content describes — the $4,000–15,000/mo full-time-operator figure in The social and marketing angle is flagged there as a vendor-interested, unverified ceiling for exactly this reason.

Speculative / unverified against a primary source: several 2026 marketing-compliance sites (including auditsocials.com and promise.legal) describe a specific "double disclosure" requirement and a further FTC update in May 2026 operationalizing synthetic-endorser disclosure. This could not be confirmed against ftc.gov directly in this research pass — see The social and marketing angle for exactly what's confirmed versus not, and verify current FTC guidance directly before building a client's compliance program around the specific disclosure format described in that secondary content.