I know you see the sales emails, LinkedIn posts and comments that are blatantly written by AI.

Not only is it becoming increasingly easy to spot (words like frameworks, gap, playbook, quietly, actually, sharpen, for example), but transparency requirements are becoming a part of AI platforms and it’s important for you and your team to know what that means for your business development strategy, the work you do for your clients, and use of AI overall for your business. 

 So below, I’m laying out, in clear, brief language, what you need to know and what it means. 

 What AI Transparency Requirements Are 

 While the bulk of the prospecting we do at RSW, and for many of you reading, is based in the States and in Canada, the EU AI Act, and specifically Article 50 of the act lays this out: 

Article 50 of the AI Act applies from 2 August 2026. It sets out transparency obligations for providers and deployers of certain AI systems, including generative and interactive AI systems and deepfakes. 

In short, Providers must: “ensure individuals are explicitly informed whenever they interact with an AI system directly” and “add machine-readable marks to enable the detection of AI-generated or manipulated content.” 

And Deployers of AI systems must inform individuals when they are exposed to,  

“Text publications on matters of public interest without human review or editorial control. 

What is Provenance and Watermarking? 

A basic definition of AI provenance: 

 Provenance in terms of AI-generated content refers to the verifiable history, origin, and chain of modifications of a piece of digital media (image, video, audio, or text). 

  Its primary goal is to restore trust in digital content by providing transparency about who created the content and what tools were used, especially whether Artificial Intelligence was involved in its creation or editing. 

And the definition of AI watermarking 

AI watermarking is the process of embedding a unique signal, marker, or set of metadata into AI-generated output—such as text, images, video, or audio—to identify its origin, trace its provenance, and verify authenticity. 

A separate post could cover how all the major social media platforms are (or aren’t) using watermarks right now, but instead I’ll provide a simple table (yes, created by AI), that shows the differences between Invisible / Machine-Readable Watermarks and Visible / Human-Readable Labels: 

Is There a Federal AI Transparency Standard in the U.S.? (Not Yet) 

 Without getting too bogged down here, it’s important to know where we currently stand, as far as industry standards. The highlights: 

1. C2PA / Content Credentials is The Coalition for Content Provenance and Authenticity (C2PA), a cross-industry group that publishes an open standard called Content Credentials. 

C2PA is an open standard for Content Credentials, attaching cryptographically signed, tamper-evident provenance to media, like a digital nutrition label, showing origins, edits, and helping combat misinformation. Sanity.io 

C2PA, however, has no enforcement power, it’s a standards body, with an open, royalty-free standard that Anthropic, Google, and OpenAI have all indicated they’ll adopt, or already have as of this writing. 

2. California’s AI Transparency Act (SB 942): As of January, 2026, it directs vendors toward open standards like C2PA but permits any interoperable protocol, and requires providers to store provenance logs including model version, prompt, and timestamp for at least three years. 

 Essentially, California functions as de facto national policy, because platforms don’t build state-specific versions of their products. 

 Worth noting that Canada doesn’t have federal oversight either at this point, instead leaning on voluntary disclosure and general consumer protection law.  

And before I inundate you with all kinds of other data, if you’d like to go down a rabbit hole, there’s also:

• Section 5 of the FTC Act-The FTC has issued guidance treating undisclosed AI-generated content in commercial contexts as a potential deceptive trade practice
• The TAKE IT DOWN ACT-signed May 19, 2025, which requires platforms to remove non-consensual intimate imagery, including AI-generated deepfakes, and empowers the FTC to enforce compliance, and,
• A pending bill, not law: in July 2026, Reps. Gottheimer, Kean, and Liccardo introduced bipartisan legislation that would require AI-generated content to carry a label embedded in its metadata.

Where AI Watermarking Stands Today

 OK, I know, that was all somewhat dry, but important.  If nothing else, you can give your clients an info drop to show them you’re on top of this. 

But, it is important to your business development process as well.

As I mentioned, major social media platforms handle visible labels using a specific approach. 

For example, you’ve probably seen Meta (Instagram/Facebook), TikTok, or YouTube Shorts, where you scroll through feeds and these platforms display visible text tags like “AI Info”, “AI-generated”, or “Altered or synthetic content”.

Look, at this point, we know it when we see it, but there are all kinds of studies backing it up as well.  

For example, this one from a USC & University of Florida Study: 

Researchers measuring audience perception found that 83% of recipients rate human-driven messaging as sincere, whereas sincerity scores drop to 40%–52% when messaging is perceived as heavily AI-generated.

 Now, this doesn’t mean every prospect is routinely scanning cold emails, but detection tools and “AI-pattern recognition” are now easily available, and AI summaries are becoming more prevalent. 

What Watermarks Can (and Can’t) Prove

 Of note: the most concrete U.S. watermark rule today is about images, video, and audio, not prospecting copy. 

And there’s not a single universal “AI-written” stamp across platforms: a missing watermark  doesn’t establish that content is human-created.

Conversely, a detected signal is evidence about origin, not a guarantee that the final claim, image, testimonial, or ad is accurate. 

Bottom line: adoption is uneven, technical approaches vary by vendor, and text watermarks are inherently vulnerable once a person substantially rewrites, translates, summarizes, or blends the output with original writing. 

So, moving from your biz dev process to your client work for a moment: the question isn’t whether an agency used AI, that’s expected to varying extents at this point, it’s whether the agency can explain its process, protect client information, and stand behind the work. 

What does Provenance and Watermarking Mean For Your Business Development Strategy? 

In business development, the first impression is increasingly not always cold call or an introductory meeting, but a comment, an email, a follow, a connection request, or an interaction with the prospect’s content.

Below is a table that accumulates all the ways blatant AI stands out and are probably harming your business development efforts if you’re engaging in them:

Where Your Agency New Business Outreach is Most Exposed

How To Think About Using AI In Your Prospecting Today

We’re at the stage where having AI completely write blog or LinkedIn posts, or LI comments, is counter-productive.

There are absolutely advanced users or just someone who’s found very solid prompts that can mimic your writing style and take out all the ways AI becomes readily apparent.

But now even that will come under scrutiny with provenance and watermarks.

Are there ways around that already?  Apparently, yes.  For example: Text AI watermarks will always be trivial to remove

AI provenance tools are already making some kinds of generated content easier to identify, but it’s early days.

It all begs the question though, how many hoops are you willing to jump through?

You already barely have time for new business as it is.

AI has made it easier to finally keep up on the demand gen and content side for agencies where they just didn’t have time before.

But continue to use it wholesale, and it will most likely harm your efforts, rather than help them.

How I Did This, For Example

I have absolutely had posts over the last year where I gave AI a thorough prompt, had Claude (or another platform) write it, and then I went through and changed 30-50% to make sure it fit my voice and said what I wanted it to say.

For this post, I did the opposite.  I had Claude do a whole lot of research (which I will always use it for, it’s fantastic for that), then I wrote the post, threw it into AI and asked for critical thoughts, and specifically how I could improve it for SEO/GEO.

It felt good to write it, even if it took twice as long.

Bottom Line

If you’re using AI to write all your business development content for you, think about changing that up, because you may ultimately have to anyway.

Marketing agencies overwhelmingly believe their own firm is well differentiated, even while agreeing that the agency industry as a whole is not.

That gap matters even more now that AI-powered search is reshaping how prospects find and evaluate agencies in the first place, a shift we break down in Generative Engine Optimization for Agencies.

In our recent survey of marketing agencies across North America, (The State of Agency Positioning and AI Search Readiness: What Our 2026 Survey Found) more than 80% said their agency is “somewhat” or “very” differentiated from the competitors they run into most.

Ask the same group about the industry overall, and over 75% call it “somewhat” or “very” undifferentiated.

Both answers came from the same agencies, in the same survey, which helps explain why so many agency websites end up sounding alike even when the people writing the copy for them are convinced otherwise.

Is Agency Positioning Really as Similar as It Sounds?

While I do think it’s gotten better, the answer is yes.

We’ve spent 21 years talking and working with agencies who are confident their positioning is unique.

“We’re more strategic.” “Only senior-level people work on your account.” “We’ve been on your side of the desk.”

All are usually true, but many other (most?) agency sites say these same things, often in nearly identical words.

The industry, and small and mid-sized firms, have a sort of default positioning vocabulary.

And when a prospect sees/reads phrases like this, there are eye-rolls, or worse, they’re ignored.

If ten agencies on a shortlist all describe themselves this way, the words lose meaning, regardless of how true they are for any one of them.

Why Do Agencies End Up With Positioning That Sounds Like Everyone Else’s?

The answer: most agencies build and test their positioning entirely in-house, without checking it against how the market sees the industry

Our survey found 85% of agencies developed their current positioning using only their own employees.

15% brought in a consultant.

Only 20% tested their positioning statements with actual prospects, and 22% tested them with current clients.

When your firm is testing positioning with current clients, I will give you credit, at least you’re moving outside your four walls.

However, that exercise tends to produce the same handful of compliments: smart, hardworking, strategic, a great partner.

Those are things a happy client should, and ideally does say, about any agency doing solid work.

It’s useful feedback about the relationship certainly, but not useful feedback about your actual differentiation.

What Should Small and Mid-Sized Agencies Do Differently?

You know it’s coming.  Yes, specialization is paramount here.

Per this recent Ad Age piece (What makes a small agency successful as clients demand more than independence):

True specialization is the biggest white space for small agencies to take advantage of to win more business, Davis said, adding that they must create a point of view that holding companies can’t credibly claim.

“The biggest gaps are not creative ideas because there are so many agencies that can come up with creative ideas … We’re more inclined to see an absence of the less sexy stuff,” Davis said. Agencies that can deliver exceptional cross-cultural marketing, outcomes-based creative or strategy backed by deep data and analytics can be really competitive.

But as I’ve stated many times in these posts, specialization doesn’t have to mean a one-vertical focus.

It can be an umbrella vertical, like “healthy living” for example, that can encompass many different verticals and sub-verticals, but gives your firms a point of view as an expert in the space.

And something else to think about as an aside, from the article:

Success doesn’t always mean chasing a high-spending advertiser and going up against the holding companies, several people interviewed stressed. “You can still chase massive share of a certain kind of client,” Wieser said, i.e. “the non-sexy, regional, mid-sized or small marketers.”

We do see a fair amount of this strategy with small and mid-sized firms and it’s a smart move.

They may not all be sexy, and your team will just need to accept that, if they’re a solid client.

How Should Agencies Test Their Positioning?

Against outside, not internal, eyes.

They’re not weighing your internal sense of what makes you special against the internal sense every other shortlisted agency has of itself.

They’re comparing sites, decks, and first impressions side by side, and from that vantage point, a lot of agencies blend together.

The differences between how an agency sees itself and how it reads to an outsider usually takes an outside perspective.

Not because the people inside don’t know their agency well, but because they know it too well

Which is a big reason why (shameless plug here, can’t help myself) we’ve built an AI Brand Compass we use for every new client program we kick off.

Combined with our experience working solely with agencies and PR firms, it gives us an outside, objective look, based on hundreds of other firms, at how an agency’s positioning comes across before we start representing that agency to prospects.

(Agencies can also get this if they’re not an RSW client, more on  the brand compass here.)

So, ultimately, it’s not whether your positioning feels true to you, but whether it would still feel distinct if you read it on ten other agency homepages.

undefined

AI summaries are just one dimension of how artificial intelligence is reshaping agency outreach. For a broader look at how to apply AI across your entire new business effort, see our guide to actionable AI business development strategies for ad, PR, and marketing firms.

One of the most frustrating challenges in agency new business?

When prospects go dark

Always tough, and agencies tend to get worried or even desperate when it’s been a few days or weeks, understandably so.

(I’ve written several posts arond it, here’s one that might be helpful: Prospect Ghosting You? Use this ghosting schedule to follow up.)

You can’t get anxious (easy for me to say I know) and when you do reach back out, find something of value to further the conversation so it’s not just a generic check-in.

Something interesting about going dark that’s not often talked about though:

Agencies go dark with their own prospects

Usually without meaning to, but excellent opportunities are lost this way.

For the same reasons agencies work with us (predominantly lack of time for a consistent process because clients have to come first) we realized we often needed to take the reins (where it made sense) after those first meetings.

And you can’t let the fact that you’re slammed cause a promising prospect to fall through the cracks.

It most often happens with those prospects who have a need, but it’s several weeks or months away.

How to Follow Up With Prospects Without the Generic Check-In

Don’t just send a generic “just checking in” or “touching base” email.

It tells a prospect that you need something from them, which gives them very little reason to reply.

The agencies that successfully stay top of mind in the interim between meetings bring something helpful.

That could be industry content, a quick breakdown of a campaign you just wrapped up, or an observation about their vertical and the market.

Doesn’t need to be a major production: a brief note with a link to a relevant case study or a trade article keeps the relationship warm without putting pressure on them.

It helps to look at this approach less as a standard follow-up and more as a way to keep a conversation going.

Tactical Steps to Keep Prospects Engaged

Before that first meeting ends, respectfully ask for a second, as well as a timeline if possible.

Immediately afterwards, set calendar reminders for yourself and take a few minutes to craft a meaningful email or send a thank you via snail mail.

Don’t pester them, but stay consistent in your reach-outs.

Especially if there’s new work to tout, awards won, or content around their vertical you’ve written, and the prospect will see what working with you looks like, making them feel better and better about you before you’ve started working together.

Everybody’s busy, and it stinks when a prospect goes dark-don’t do the same.