Reasons AI SDRS Struggle in the Market

Oh boy. I'm getting worried for AI SDR companies. I'm hearing from more and more people churning off them after a couple weeks of results. I just read in a private Slack group about someone already on their 3rd AI SDR company...and was looking for the 4th (I thought 3 strikes was an out!). Clearly, there is something here people want, but is AI ready to deliver it? Not yet. Here are some of the pitfalls my network is talking about: 1. Hallucinations - when AI does produce an amazing insight, it's typically not accurate bc it was made up which kills any chances of meeting. 2. Lack of nuance - most products and outbound pitches require a level of nuance AI can't replicate bc it's not in the LLMs used to train it. They can do industry and persona level messaging fairly well, but poor company and person level personalization leads to more generic messaging. 3. Sameness - when lots of companies use the same LLMs to find people and message them, you inherently sound similar to everyone else using it which makes it harder to stand out. 4. Start fast, finish last - AI quickly grabs the low-hanging fruit...it does a good job at the obvious stuff that reps can miss which means pilots return results but once you get into the hard stuff results tank. 5. More is better - when the AI starts to deliver less results, almost all leaders will do what's obvious = they'll turn up the volume which means you'll have scortched earth of leads who mark you as spam. 6. Infrastructure - more volume and more spam results in deliverability issues (which creates the vicious cycle of turning up volume even more) and most teams are not ready to manage huge email deliverability infrastructures which means your emails won't even hit inboxes (another chance for leaders to turn up the volume even more). One way to find answers to these issues is to use AI SDRs and have those companies work out the kinks. There is probably huge upside there. But, I fundamentally believe AI SDRs are still missing the inputs needed to be anything other than an experiment right now.

The AI SDR is dying. I've spoken to dozens of companies who have used AI SDR tools, and every single one of them has churned or plans to churn. Why? Companies who use them book $0 in pipeline. AI SDR Tools: ➝ Pull lists of thousands of prospects ➝ Spray-and-pray outreach to all of them Spray-and-pray doesn't work. Human sellers know this, otherwise anyone with an Apollo/Zoominfo license would be putting up record outbound numbers. The best SDRs—the ones AI SDRs SHOULD be trying to replace: ➝ Qualify target accounts ➝ Spot compelling events that signal that the timing is right to engage ➝ Stay all over prospects via the phone, email and linkedin ➝ Follow up relentlessly (but respectfully) when someone is in a place to convert Can AI automate many pieces of the SDR workflow today? Sure. We built a product that does a lot of it. But if you're expecting a piece of software to show up and understand prospecting as well as you do out of the box, you're in for a big disappointment.

I’m hearing whispers that AI SDRs aren’t living up to the hype. Shocking right?! This is my read on what’s happening… 🤖 The AI SDR promise Conceptually, an AI SDR makes a ton of sense. - My SDR teams run a highly repeatable process (think Predictable Rev). - AI can run that same process at scale and for a fraction of the cost. - Let’s cut costs and overhead, and we’ll get the same outcome. And sure, if this is the playbook you’re running, by all means, use an AI SDR instead of an SDR team. AI SDRs, in fact, do the above pretty well. But pretty good isn’t good enough in today’s market. The difference between a pretty good SDR and a ~~10x~~, scratch that, 11x-SDR is vast. Let’s compare the two: 👍 Workflow of a pretty good SDR - Get handed an account list from RevOps. - Hit up Sales Nav or fav prospecting tool to grab the right titles from that list. - Do some surface-level research on each prospect (e.g., check out their LinkedIn profile, look at their website, ask ChatGPT to do some research). - Match the prospect to an outbound sequence (maybe personalize the first message a bit) and fire away. - Repeat. 🦾 Workflow of a top-performing SDR. - Get handed an account list from RevOps. - Prioritize that account list by conversion dimensions (AKA signals), like previous engagements, recent hires, product activity, website visits, etc. - Identify prospect(s) in said accounts and go beyond the surface level, looking into job history, social activity, previous interactions (think email ops) in the CRM, etc. - Connect the dots above to reach out with the right message to the right person at the right time. - Repeat. Yes, the top-performing workflow is a lot more tedious. But it converts a heck of a lot better. When your conversion rate is 0%, scale and speed just exacerbate the problem. So, what’s the solution? In my (biased) opinion, AI needs to go beyond the surface level to be effective at the SDR motion. That is, it needs to be built on your proprietary data infrastructure (think your CRM, data warehouse, external buying signals) and unify that data at the person and account levels to get a holistic view of each prospect that roles up to each account. From there, you can feed AI with the right inputs to get quality outputs that actually convert. That’s what we’re building at Common Room and it makes all the difference. Quality in = quality out. === One last thing… tomorrow, I’m going to be sharing a helpful resource that is a major step in the direction we’re headed. Stay tuned.