Doug Camplejohn
(00:02)
Hello, this is Doug Camplejohn and welcome to Revenue Renegades. My guest today is Anis Bennaceur, the founder and CEO of Attention.com. Anis, welcome to the show.
Anis Bennaceur
(00:15)
Thank you. I’m very excited to be here. Great to connect with you and hello everyone.
Doug Camplejohn
(00:21)
To start off, for everyone listening, please give us a bio on what Attention.com is and how you came up with the idea. What was the founding moment?
Anis Bennaceur
(00:34)
Absolutely. Attention.com is your system of AI agents for sales. We automate work from customer conversations. While many have talked about top-funnel AI agents and AI SDRs, little has been built around sales bottom-of-funnel automation. I developed the idea to solve business challenges I faced with my previous startup.
I’ve been a founder for 10 years in New York City. My last startup, Mixer, was in the creator economy, selling job posting subscriptions to companies like Square and Capital One. It was very enterprise-focused, and as a bootstrapped startup, the challenges were immense. It was extremely painful to survive solely on our own profits. We spent money on Salesforce, which required extensive ramp-up over calls to avoid mistakes, and that process was burdensome.
Anis Bennaceur
(02:20)
Nothing groundbreaking existed at the time—until the July 19, 2020 announcement of GPT-3. That announcement unlocked my thinking. I began experimenting with the API and realized you could feed text to an LLM. The first application I built was for writing emails with AI—sending cold emails well before many others were doing so in 2020–2021.
Then I met my co-founder, Matias, who had once been my biggest competitor. Over coffee, he explained his work in computer vision and fusion models. We discussed what NLP could do and decided to start Attention.com together. We identified several pain points, including the ability to feed an entire call transcript (by breaking it up because of context limits) into an LLM to obtain structured outputs. That was a breakthrough for us.
Anis Bennaceur
(04:05)
We realized we could auto-fill the CRM based on what was said during conversations—and also provide coaching feedback to sales reps. We understood early on that while LLMs weren’t yet advanced enough for coaching, each company had its own coaching style. The real pain point was filling the CRM. Auto-filling every field and property based on conversation data was a game changer.
Doug Camplejohn
(04:40)
Got it. So you were pretty early—before the broader discovery of ChatGPT 3.5. When did you actually deliver the first product and start charging?
Anis Bennaceur
(05:03)
We started in September 2021 and had our MVP in the hands of design partners by summer 2022. We gave it away for free to five companies to iterate. Our aha moment came in early November 2022 when DaVinci 3 emerged as a significant improvement over DaVinci 2—almost as capable as ChatGPT 3.5. With the world waking up to ChatGPT, we moved quickly and came out of stealth. By the end of January 2023, following a TechCrunch article on our seed round and product, we began selling to clients and grew rapidly that year.
Doug Camplejohn
(06:30)
CRMs are notoriously messy. I assume it was challenging to take unstructured transcript data, parse it into a structured format, and map it into the CRM fields. Can you explain how your AI managed that and some challenges you faced?
Anis Bennaceur
(06:57)
Great question. We discovered that most clients don’t know how to prompt properly. So we built a system within our product that allowed us to manually input specific prompts. For a while, we super-served each client—building custom prompts and integrations behind the scenes. This was acceptable at first because clients trusted us with access to their CRMs.
Doug Camplejohn
(07:45)
Wow.
Anis Bennaceur
(07:58)
I even had large clients give us access to their Salesforce, HubSpot, or even PipeDrive accounts. We manually mapped data, sometimes syncing not only field-level details but also summaries and call recording links. This flexibility allowed us to distribute our solution quickly and explore various growth opportunities.
Doug Camplejohn
(08:49)
Really interesting. In your bio, you mention growing revenue 10× in the first year. Without disclosing exact numbers, can you share some of the go-to-market techniques you used and what worked or didn’t?
Anis Bennaceur
(09:22)
Certainly. We went from zero to one million ARR in about 10 months—an incredible achievement. We prioritized selling mostly to inbound clients who sought us out rather than chasing outbound leads. During development, I even ran a bot that added our entire ideal customer profile on LinkedIn and posted regularly about AI’s potential for sales—from auto-filling CRMs to drafting automated emails. This built trust and a strong brand.
Referrals played a huge role; about 35% of our new organic leads came from client referrals without any formal referral program. We relied on direct conversations and the classic NPS question to ask for specific referrals, such as connecting with a VP of Sales at a target company. The key was not chasing prospects until they were in a buying cycle, which saved everyone’s time.
Doug Camplejohn
(11:02)
That’s great.
Anis Bennaceur
(11:17)
I was handling roles from CEO to BDR, AE, and CSM until after closing our first $200K in ARR. Once we reached that point, I began hiring AEs and systematizing client outreach and feedback.
Anis Bennaceur
(11:47)
After closing deals, we focused on asking for specific referrals. For example, instead of a general “Can you refer someone?”, we would say, “I noticed you used to work at Company XYZ—can you introduce me to their VP of Sales?” Often, people would either refer a qualified contact or suggest reconnecting in a few months. The trick was to keep them engaged on LinkedIn so they would eventually reach out when ready.
Doug Camplejohn
(12:40)
Super simple and straightforward—letting the product speak for itself. How did you approach pricing? How did you decide on the optimal price for your solution?
Anis Bennaceur
(13:07)
I could have developed an elaborate model, but the truth is I looked at what Gong was charging and decided to charge $20 more, not less. We positioned ourselves as a premium tool with a better offering.
Doug Camplejohn
(13:17)
That makes sense.
Anis Bennaceur
(13:28)
I wanted to position us as a premium product rather than undercut the competition. Fast forward to 2025, we’ve rolled out AI agents that automate work for dozens of people. We layered in a credit system—each thousand dollars spent consumes credits, and once they run out, users can top up. This approach allowed us to maintain our base pricing while upselling to clients who wanted to deploy more agents.
Doug Camplejohn
(14:43)
There’s a lot of talk these days about different pricing models, like outcome-based pricing. What are your thoughts on that for sales?
Anis Bennaceur
(14:53)
For sales, attribution is challenging because success is a mix of AI and human efforts. Improved conversion or win rates might be due to better coaching, product-market fit, or even new hires. While our product can automate about 10–20% of the workload, pricing based solely on outcomes isn’t feasible. In customer support, where an AI agent handles end-to-end work, outcome-based pricing is easier.
Doug Camplejohn
(16:32)
I’ve been scratching my head about that too. In support, models like Intercom’s per-case pricing or Brett Taylor’s Sierra’s outcome-based approach seem promising. Even Salesforce is exploring similar methods.
Anis Bennaceur
(16:34)
Exactly.
Doug Camplejohn
(17:00)
Beyond AI SDRs that claim to replace a full-time hire for a lower cost, I’m curious to see how the business model will evolve for sales.
Doug Camplejohn
(17:55)
What do you think people are getting wrong about AI right now? Everyone’s talking about it, whether it’s agents or multimodal capabilities. What’s being misunderstood?
Anis Bennaceur
(18:07)
Many use the term “agent” without understanding its true meaning. In my view, we did crack it in a way, but most haven’t considered the full definition. There’s a difference between an AI-enabled workflow—which follows a deterministic, pre-configured set of steps—and an AI agent, which has the autonomy to create its own decision tree based on varying outcomes. Many claim to have built an AI agent when in fact they’ve only built an AI-enabled workflow. Furthermore, people often confuse AI agents for sales with AI agents for growth. True AI agents for sales automate tasks from lead qualification to nurturing and closing deals.
Doug Camplejohn
(20:34)
I actually think many so-called agents are just simple if-this-then-that setups. I believe you announced some agent-related features for Attention.com. Can you explain what those are?
Anis Bennaceur
(20:52)
Certainly. Over the past several months, after hiring an excellent VP of Growth and a Head of Sales, I’ve returned to the lab to build AI agents internally. We’ve built over 50 agents, and we plan to release one or two each week over the next year. These agents automate end-to-end tasks. For instance, one agent automatically builds your sales enablement knowledge base by capturing and analyzing call transcripts and emails tied to specific opportunities in your CRM.
Another agent automates action items by coordinating with other agents. For example, one agent might build a business case for a prospect by collating all relevant conversations and emails. We’re also releasing AI voice agents—an inbound IVR system that qualifies prospects based on criteria such as ACV. These voice agents can leverage the full context of past conversations, technical documentation, and deal history to provide the right answers during a call. I’m not saying these will replace sales engineers, but they can be deployed for lower-ACV clients with technical needs.
Anis Bennaceur
(23:41)
Engineers can also feed their documentation into the system, so the AI agent has full context for the deal. This creates a dual link between Attention.com and the CRM, enabling tasks like building business cases and providing rapid analytical insights—what we call AI RevOps. One agent can analyze thousands of conversations in minutes, automating the work of an analyst.
Doug Camplejohn
(25:53)
Very cool. It sounds like you’ve got a year’s worth of LinkedIn posts queued up, too. What about interoperability? How do you view frameworks like Crew AI or marketplaces like Agent AI?
Anis Bennaceur
(26:22)
It’s exciting because it broadens people’s understanding of what agents can do. In my opinion, specialization is key. General note-taking apps or generalized AI assistants, like Zoom’s AI companion, may struggle against specialized solutions built for specific functions—whether for customer success, sales, HR, or even for doctors. Specialized agents that focus on narrow verticals can offer more targeted, function-centric workflows.
Doug Camplejohn
(28:09)
I think it will be interesting to see whether we end up with many silos or a unified marketplace—perhaps an app store or AppExchange for AI tools.
Anis Bennaceur
(28:36)
I believe if you focus on your vertical, you’ll never hit a ceiling because work is so varied. OpenAI might even roll out their own AI agent app store since they already have a GPT app store and operator that interacts with any interface. I wouldn’t be surprised if they release specialized agents soon, which will push other players to focus on niche markets.
Doug Camplejohn
(28:49)
Understood. What are some of the biggest challenges revenue organizations face today—beyond what Attention.com solves?
Anis Bennaceur
(28:55)
For sales, the main challenges include poor CRM hygiene and the difficulty of increasing conversion rates. Sales leaders focus on conversion rates, while enablement teams concentrate on market insights and RevOps teams need to understand close-loss trends. With AI, we’re not replacing software but future headcount. Another common challenge is generating top-funnel leads. Cold outbound is increasingly difficult—most people aren’t creative enough to stay ahead of the curve. By the time they try a new method, the response rates are extremely low.
Anis Bennaceur
(32:16)
The companies that scale fastest are those with a broad market fit.
Doug Camplejohn
(32:32)
That makes sense. It all comes back to having the right product.
One last set of questions: You’ve had an interesting career—banking, marketing at Tinder, running a startup like a creative version of LinkedIn, and now Attention.com. Have any lessons from those industries influenced how you run Attention.com?
Anis Bennaceur
(33:13)
Investment banking made me a hardcore worker. It instilled a work-hard culture—you show up every day (some even on Sundays)—and taught me to triple-check every email and decision. I never tolerate a failure of execution. Tinder didn’t impact my operations at Attention.com directly, but it led me to found Mixer, where I honed my growth hacking skills and developed a scrappy mindset. I still bring that founder mentality to every growth meeting.
Doug Camplejohn
(35:30)
Got it. As we wrap up, I have a few standard questions. AI touches every part of our lives. If you could automate one task in your life forever using AI, what would it be?
Anis Bennaceur
(35:54)
If I could have an assistant that handles any task—admin or personal—that would be amazing. Instead of searching on Google or Perplexity and doing tasks manually, I’d simply text a command to update my USPS redirect address, for example, and have it done automatically.
Doug Camplejohn
(36:52)
It’s funny—as a startup founder, when my last company got acquired by LinkedIn, I never had a full-time assistant. I had to manage expenses and travel on my own. If you had reliable agents handling these tasks, it would be phenomenal.
Doug Camplejohn
(39:12)
Maybe you need an AI bot that listens to your conversations and fills you in on what you missed.
Anis Bennaceur
(39:21)
Perhaps Attention.com could develop that too.
Doug Camplejohn
(39:26)
And one more: What’s one thing you can’t live without in your daily routine?
Anis Bennaceur
(39:37)
I’d say Le Croix strawberry peach coffee—I drink it every day. Also, 8 Sleep has completely transformed my sleep quality. When I travel and don’t have a proper sleep setup, I don’t sleep well. A good night’s sleep is essential for staying efficient during the day.
Doug Camplejohn
(40:39)
I’ve heard the same from friends. I’m finally going to get an 8 Sleep, though hotels need to catch up!
Anis Bennaceur
(40:59)
Absolutely. In fact, a hotel offering 8 Sleep beds would be a game changer.
Doug Camplejohn
(41:06)
That’s awesome. Well, Anis, it’s been a pleasure talking with you. How can listeners connect with you?
Anis Bennaceur
(41:23)
Connect with me on LinkedIn—I accept everyone. Please don’t include a message with your connection request because it generates multiple notifications.
Doug Camplejohn
(41:44)
Fantastic. Thank you so much for your time, and thanks to everyone for listening.
Anis Bennaceur
(41:51)
Thank you. This was awesome. Bye.
