Top of Funnel (TOFU) Marketing Strategies

**(00:00)**

David: Elaine, it’s a pleasure reconnecting with you live today. Excited to hear about your work with Tofu and how AI fits into go-to-market strategies, which I know is a hot topic here. I think it’s something we’re all grappling with in different ways. To start, could you introduce yourself? Then we’ll jump into the conversation.

Elaine: Thanks, David. Hi, everyone! Great to meet you all. I’m the co-founder and CRO of Tofu, a generative AI platform designed to automate top-of-funnel go-to-market strategies for B2B marketing teams.

I'll walk you through what that means, but a bit about me first. I actually started my career as a biomechanical engineer—which isn’t exactly related to marketing! I quickly realized the hardware world moves too slowly for me, so I pivoted into software.

I joined a startup as employee number two, where I spent five years and helped raise about $75 million.

I started in product, but no one was doing go-to-market, so I took it on and loved it. After that, I joined Slack to lead enterprise go-to-market when they launched their enterprise product.

I later moved into venture capital at SignalFire, where I focused on AI and software investments for five years. 

I've been following generative AI since 2018. In fact, my first product, Soundspot, turned corporate blogs into podcasts.

This was back in the days of GPT-1 and Google’s early Tacotron model for speech. Looking at how much these tools have advanced, it feels like magic now. Seeing how

AI is transforming industries, I wanted to dive back into the startup world, which led to Tofu. So that’s how I ended up here.

David: Fantastic. I can’t wait to dive in!

**(04:30)**

Elaine: Let’s jump into some areas where AI is making waves in go-to-market strategies. Text generation is already working well; the tools are mature, especially for shorter forms like blogs and social posts. Audio, too, has come a long way, especially with cloning tech like ElevenLabs that can replicate voices in just minutes. I’ll play an example so you can see what I mean.

[Elaine plays a voice synthesis example showing how AI can clone her voice accurately.]

It’s amazing but also a little frightening when you think about the implications. Image generation is advancing but not fully ready for B2B marketing. It’s still challenging to generate text within images accurately, and for some tasks, image models don’t yet meet quality standards for B2B.

**(08:37)**

Elaine: So, how is AI changing go-to-market? I recently came across an interesting perspective from Kieran Flanagan, CMO of Zapier. He said that in the past, there were "original thinkers," "copy-paste marketers," and "lazy marketers." With AI, the original thinkers’ group has shrunk while copy-paste marketers have grown, as AI makes it easier to replicate content. The winners in marketing will be those who leverage AI to enhance their unique ideas.

Content is still critical, but AI is changing how we create it. SEO is a huge area of transformation. If you’ve tried Google’s new Gemini-powered search, it now surfaces direct answers right at the top, so you don’t even need to scroll. That will impact content strategies significantly.

**(12:00)**

Elaine: Let’s talk about where AI is already proving effective. It’s great for brainstorming, outlining, and helping marketers overcome the blank-page problem. For personalization, it’s helping marketers understand their audience deeply, identifying pain points, and adjusting messaging accordingly. Repurposing content is another high-value use case, especially in a tight economy where teams need to do more with less. Automating outbound prospecting is gaining traction as well, with tools providing AI-driven SDR functions. 

Here’s where it’s making a difference today:

1. Personalization across channels,

2. Segmenting audiences by account and persona,

3. Repurposing content from events, webinars, and reports into blog posts, social media posts, or emails.

**(18:00)**

Elaine: Now, let’s do a quick demo of how Tofu handles personalization. We target specific accounts, do research on them, and generate content tailored to them. For instance, let’s say we’re creating a landing page for Rapid SOS, a company in emergency response. I select the elements to personalize, run the page template, and Tofu uses our research to adjust the content to speak directly to their industry and values. We can do the same for emails, social posts, and ads, keeping each piece consistent with the company’s tone and branding.

**(25:00)**

Elaine: Another strong use case for Tofu is content repurposing. For example, we can take a 45-minute AMA session, transcribe it, and turn that into multiple blog posts, social posts, and email follow-ups. This way, one high-value piece of content generates various assets across channels. Our customers love this feature because it saves hours, and teams don’t need to reinvent the wheel with each new campaign. We also use third-party content like reports from McKinsey or trending LinkedIn topics and repurpose them, giving credit to the original author, of course.

**(32:00)**

Elaine: There are a few key limitations to AI. The biggest challenge is its reliance on prediction without real understanding, which leads to hallucinations and inaccuracies. Especially in regulated industries like healthcare or finance, where the stakes are high, companies can’t afford these errors. At Tofu, we’ve built a rules engine that sits between the AI and the customer to check compliance with industry guidelines. But this technology isn’t flawless. For instance, the outputs are only as good as the inputs, and if you give it bad data, it will give you bad results.

**(40:00)**

Elaine: People often ask, "Will AI replace marketers?" My answer is that we still need humans, especially for the creative and strategic aspects of marketing. Humans bring the ideas, brand voice, and strategy that AI can’t replicate. What AI does well is amplify and accelerate human creativity. So, for things like messaging, content strategy, and channel optimization, human marketers are irreplaceable.

Eric: Elaine, this is an amazing presentation. I have two questions. First, what kind of models does Tofu use? Are you using open-source models or something like OpenAI? And second, with so many rights issues around AI training data, how are you handling that?

Elaine: Great questions. We use multiple models depending on the task. For short form, like social posts and ads, we find that GPT-4 works best. For long-form pieces, we prefer Anthropic’s Claude 3. We’re also careful with customer data; we don’t train on sensitive information, and our infrastructure is set up to keep customer data isolated. But you’re right; issues around data rights are complex, especially when models are trained on massive internet data sets.

**(50:00)**

David Cutler: What’s your most effective demo for new prospects?

Elaine: We always use the customer’s own data. I’ll demo Tofu using their target accounts and show how it can immediately generate landing pages, emails, or social content personalized for their audience. I also show our “webinar in a box” feature, where a single event generates multiple content assets for follow-ups and social promotion. It’s an eye-opener because they see the efficiency and time savings right away.

**(52:15)**

David: Thank you so much, Elaine. This was packed with insights, and I think we’ve all gained a lot from seeing how Tofu can simplify go-to-market. Everyone, please remember that Katherine Montgomery will be our guest next week, and we’ll have our in-person event at New York Tech Week soon. Have a wonderful weekend, and thank you all for joining!