Most agencies pitching AI have not built anything. They have subscriptions, ChatGPT, a scheduling tool, maybe a wrapper someone called proprietary, and a new vocabulary for the same deliverables. The harder part is that distinguishing this from real integration requires asking questions agencies are not used to answering. This checklist is those questions.
What AI Washing Actually Means in an Agency Context
AI washing is presenting AI involvement as proof of value when the involvement is either superficial, misleading, or entirely unrelated to the results you’re buying. The SEC charged two investment firms in 2025 for falsely claiming AI-driven strategies, establishing legal precedent that AI washing isn’t just hype, it’s fraud when money changes hands on the basis of it.
In agency marketing, it’s rarely that explicit. It’s subtler: a repositioned proposal, a rebranded workflow, a testimonial about “faster turnaround” with no mention of whether the output performed.
The Difference Between Using AI Tools and Building AI Systems
An agency that uses ChatGPT to draft copy is using an AI tool. An agency that has built a retrieval system that pulls your brand guidelines, past campaign performance data, and competitor positioning into every brief generation, that’s closer to a system. One is a subscription. The other takes months and real engineering.
Most agencies pitching “AI integration” are in the first category and charging for the second.
When “AI-Powered” Is Just a Rebrand of Old Automation
Watch for proposals where “AI agents” are handling email sequences, social scheduling, or basic A/B testing. These workflows existed in 2018. Mailchimp’s automation, Hootsuite’s scheduler, and Optimizely’s testing engine predate the current AI wave by years. Renaming them “AI-powered workflows” in 2026 is not an upgrade, it’s a vocabulary change.
The tell: ask what specifically changed in the workflow after the agency “added AI.” If the answer is vague or describes a tool you can buy yourself for $50/month, that’s your answer.
The 7 Red Flags in Agency Proposals and Pitches
These aren’t hypothetical warning signs. Each one maps to a real pattern in agency sales decks.
Vague Efficiency Claims With No Baseline Data
“AI lets us work 40% faster” is meaningless without knowing the baseline. Faster than what? If the agency can’t tell you what the previous timeline was, what the current timeline is, and how that speed improvement translates to a deliverable you receive, it’s marketing copy, not a metric.
Legitimate efficiency gains have units. “We reduced first-draft turnaround from 5 days to 2 days” is a claim you can verify.
”Human Oversight” Mentioned Once, Never Defined
Every agency with an AI workflow should be able to explain exactly where a human reviews AI output before it reaches you. Not “we always review everything”, that’s a reassurance, not a process.
Ask: Who reviews it? Against what criteria? What percentage of AI drafts are materially revised before delivery? If those answers aren’t immediate, oversight is aspirational.
No Explanation of What Happens When AI Is Wrong
AI systems produce confident, plausible errors. A 2025 WordStream study found 20% of AI responses to PPC questions contained inaccurate information, 26% specifically from Google AI Overviews. Any agency using AI in a client-facing workflow has encountered this. The question is whether they have a documented response to it.
If the proposal doesn’t address failure modes, the agency hasn’t thought about them seriously.
Testimonials About Speed, Not Results
“They turned around our content in half the time” is a testimonial about the agency’s operations. It tells you nothing about whether the content worked. Scan every case study and testimonial for outcome metrics: traffic, conversion rate, revenue, rankings, cost-per-lead.
Speed is a feature. Results are the product. If testimonials only reference features, ask why.
Ethics Language With No Enforcement Mechanism
“We’re committed to responsible AI use” appearing once in a proposal is decoration. What does responsible use mean in this agency’s specific context? What happens if an AI tool they use introduces bias into targeting? Who on their team is accountable?
Ethics principles without enforcement mechanisms are brand positioning, not policy.
Proprietary AI With No Technical Specifics
“Our proprietary AI system” is a phrase worth interrogating hard. What model underlies it? What data was it trained or fine-tuned on? What does it do that a commercial API (OpenAI, Anthropic, Google) doesn’t? What’s the data retention policy on anything you share with it?
A legitimate proprietary tool survives these questions. If the answer is deflection or NDAs that prevent you from understanding what you’re using, that’s a red flag.
AI-Generated Case Studies Dressed as Proof
Look at the specificity of an agency’s case studies. Generic phrases, round numbers, missing client names, and claims without screenshots or third-party attribution are common signals that the case study was assembled with AI assistance, or fabricated outright. Cross-reference. Ask for the client contact. Request the raw analytics screenshot. Verifiable results survive scrutiny.
The Five Questions to Ask Before You Sign
These questions work in a first meeting. You don’t need technical expertise to ask them, just a willingness to wait for a real answer.
Map the Claim to a Workflow Step
Ask the agency to walk you through one specific deliverable, say, a campaign report or a content brief, and explain exactly where AI is involved. Not “across our workflow.” One deliverable, step by step.
This forces specificity. If they can’t do it for one example, the AI integration is a positioning layer, not an operational one.
Ask for the Failure Mode
“When your AI tool gives you bad output, what does that look like and what do you do?” is a question every legitimate AI workflow has an answer to. A good answer includes a specific example, a specific correction process, and what safeguard now exists to prevent it from reaching you.
A non-answer, or a claim that it hasn’t happened, tells you they’re not measuring for it.
Demand Measurement of Incrementality, Not Activity
Activity metrics (content pieces produced, posts scheduled, emails sent) can be inflated by AI with no improvement in outcomes. Ask what specific outcomes they’re contracted to improve, how they’ll be measured, and what the baseline was before the engagement started.
Incrementality means: what changed because of this engagement, measured against what would have happened anyway? If that framing confuses them, they’re not measuring results.
Ask Who Owns the Output and the Audit Trail
If an agency uses an AI tool to generate copy, designs, or strategy documents for you, who owns those outputs? What happens to your data inside their toolchain? Can you audit what AI generated versus what a human wrote?
This matters for intellectual property, for compliance in regulated industries, and for your ability to replicate results after the engagement ends. Get it in the contract.
Find Out When They Last Updated Their AI Knowledge Sources
AI tools trained on data from 2023 will give you confident, outdated answers on fast-moving topics like ad platform algorithms, search ranking signals, or competitor pricing. Ask the agency what their process is for keeping their AI tools current, and whether they’ve verified that their tools’ knowledge cutoffs match the timelines that matter for your campaigns.
What Legitimate AI Integration Actually Looks Like
Knowing the red flags is half of it. The other half is knowing what you’re comparing them against, so you’re not just pattern-matching for absence of trust signals.
Clear Boundary Between Automation and Human Judgment
Legitimate integration defines what AI handles and what humans handle, and why. At Designodin, we’re explicit about this: AI assists with image optimization, draft scaffolding, and code review. Final copy, design decisions, and client strategy are human. That boundary exists because those are the places where AI errors compound quietly, wrong brand tone, flawed strategy logic, compliance issues, and where a human catching them early is cheaper than a client catching them late.
An agency that can’t articulate that boundary with the same specificity is either not thinking about it or doesn’t want you thinking about it.
Measurable Baselines and Defined Improvement Targets
Before any AI-assisted campaign starts, there should be a documented baseline: where you are now, what metric you’re trying to move, and what constitutes a success at 3 months, 6 months, and 12 months. This is how you know whether AI involvement helped or hurt. Without it, any claim of “AI-driven results” is unfalsifiable.
88% of organizations use AI regularly, but only 33% have scaled AI programs beyond pilots, according to McKinsey’s 2025 State of AI report. The difference between those groups almost always comes down to measurement discipline.
Transparent Tooling (Not Always “Proprietary”)
A trustworthy agency can tell you what tools they use, what those tools do with your data, and what happens to your outputs if you end the engagement. “Proprietary” is sometimes legitimate; but it should never mean “we can’t explain what it does or how.”
Our hand-coded WordPress sites use explicit, auditable tooling at every layer. We can show you the code. That’s the bar.
Frequently Asked Questions
Is it AI washing if an agency uses ChatGPT for content but doesn’t disclose it?
It depends on what they claimed and what you paid for. If the agency sold you on “expert human-written content” and delivered AI-generated drafts with minimal revision, that’s deceptive regardless of the tool. If they were transparent that AI assists with drafting and humans do final review, it’s a workflow choice, not washing. The issue is disclosure and accuracy of the claim, not the tool itself.
What’s the difference between AI washing and just exaggerating in a sales pitch?
AI washing is specifically the practice of attributing outcomes or capabilities to AI involvement when that involvement doesn’t substantively contribute to results, or doesn’t exist as described. Regular sales exaggeration is about scope or timeline. AI washing is about crediting a mechanism that isn’t doing the work. The SEC has treated it as a material misrepresentation when investment decisions follow from it.
Can I ask an agency to prove their AI claims before signing a contract?
Yes, and you should. Ask for a live walkthrough of a deliverable that uses their AI workflow. Ask for anonymized case study data with verifiable baselines. Ask to speak with a current client whose results were driven by the AI integration specifically. A legitimate agency with real AI integration will welcome this. One that deflects or generalizes is protecting a claim they can’t back up.
Are there legal consequences for agencies that make false AI claims?
Increasingly, yes. The SEC established precedent in 2025 by charging investment firms for misleadingly claiming AI-driven strategies. While most marketing agency claims haven’t been tested in court yet, false claims made in a contract that induces a buyer to spend money are subject to fraud and false advertising law in both the US and EU. The legal exposure is real and growing as regulators focus on AI marketing claims.
How do I evaluate an AI integration proposal without a technical background?
Ask the agency to explain one deliverable step by step without using the words “AI-powered,” “intelligent,” or “automated.” Ask for the failure example and the fix. Ask for the metric they’ll be held accountable to. Technical credibility shows up as precision, a practitioner who has built something can describe it concretely. Vagueness is a signal regardless of jargon. If the explanation only makes sense at a high level, that’s because it only exists at a high level.
What should a legitimate AI proposal actually include?
It should include: the specific tools being used and their data policies, the workflow steps where AI is and isn’t involved, the human review process and who’s responsible for it, the baseline metrics before engagement and the improvement targets, and ownership terms for AI-generated outputs. If any of those are missing, ask why before signing.
If an agency can’t walk you through one deliverable in plain language and explain exactly where AI helped, that’s your answer. The phrase “AI-powered” has become a price justification, not a capability description. Protect your budget accordingly.
If you want to talk through what this looks like for your operation, start a conversation. We’ll be direct about whether what you’re being sold is real. See how we scope and build this at designodin.com/ai.