AISalesReps

Crafting Effective Follow-Ups: My Reality Check on Automating Follow-ups with AI

Dan Hartman headshotDan HartmanEditor··6 min read

Learn how automating follow-ups with AI can go wrong (and right). I share my experience building an outbound sequence, the debugging pain, and the real costs.

Crafting Effective Follow-Ups: My Reality Check on Automating Follow-ups with AI

Last month, I found myself staring down a mountain of outbound leads. We needed to hit aggressive targets, but hiring more SDRs wasn’t in the budget, and our existing team was already stretched thin. The biggest bottleneck, as always, wasn’t the initial cold email – we’d gotten pretty good at writing those – it was the follow-up. Specifically, personalizing those subsequent touches in a timely manner without drowning in manual work. That’s where I thought, “Okay, this is it. This is the perfect use case for automating follow-ups with AI.”

I’ve shipped enough agents to know better than to blindly trust the hype, but the idea of an AI handling the nuanced dance of a sales follow-up was incredibly appealing. Imagine an agent that could read past interactions, detect sentiment, and craft a perfectly tailored, concise email to re-engage a prospect. No more generic “just checking in” emails. No more missed opportunities because a human got busy. The dream, right? Well, the reality, as always, was a bit more… textured.

The Promise vs. The Pain of AI Follow-Ups

My initial thought was to use a simple automation platform. Something like n8n for sales workflows or Zapier could trigger an email after a certain delay if there was no response. But that’s just glorified mail merge. It doesn’t actually understand context or personalize. For real smart follow-ups, you need an LLM in the loop, making decisions based on data. That’s where an agentic workflow comes in.

I started prototyping with a custom setup. The core idea wasn’t overly complex on paper: ingest email threads, identify non-responders, then have an LLM generate a follow-up. I used the Vercel AI SDK to quickly get an LLM hooked up, then began building out the orchestration. For more complex decision trees – like, “if they opened but didn’t click, send this; if they replied negatively, flag for human review” – I knew I’d eventually need something like LangGraph or CrewAI to manage the state and transitions. But even the simpler version was a headache.

Here’s my concrete gripe: the debugging pain of multi-step agents. When an LLM call goes sideways – it hallucinates a detail, misinterprets the last email, or just generates something off-brand – tracing that back through a chain of prompts, API calls, and conditional logic is brutal. You’re not just debugging code; you’re debugging emergent behavior. Tools like LangSmith or Langfuse help immensely with observability and tracing, but even with them, it’s still a slog. You’ll spend hours pouring over logs, trying to figure out why your agent decided to congratulate a prospect on a promotion they didn’t get. It’s a silent killer of productivity, and frankly, it’s why most people don’t get past the demo stage.

Where My Agent Broke Down (and How to Fix It)

The failures weren’t always obvious. Sometimes, the agent would just… stop. An API rate limit, a subtle change in the LLM’s response format, or a malformed input would halt the entire sequence. That’s a silent failure, and it costs money in missed opportunities. We quickly learned to build robust retry mechanisms and detailed error logging, but even then, you need constant monitoring. And forget about using something like AutoGen right out of the box for this kind of sensitive task; the risk of an unconstrained agent sending out bizarre emails is just too high.

Cost creep was another monster. Every token adds up, especially when you’re generating multiple drafts, refining prompts, or just debugging. You think, “Oh, it’s just a few cents per email,” but at scale, those cents become dollars, then hundreds, then thousands. Suddenly, your “cost-saving” automation is looking pretty expensive. We had to get incredibly strict about prompt engineering to minimize token usage, focusing on concise instructions and tight response formats. This isn’t just about efficiency; it’s about not burning through your budget for no good reason.

Then there’s the compliance nightmare. Handling PII (personally identifiable information) in outbound sequences is a minefield. GDPR, CCPA, and general data privacy are serious business. Our agent needed to be incredibly careful about what information it stored, processed, and, most critically, sent out. An agent hallucinating someone’s company or title is one thing; an agent accidentally leaking sensitive data is a whole different ballgame. We baked in strict data sanitization and anonymization steps, and every decision the agent made had to be auditable. You simply can’t deploy agents touching real user data or real money without a clear audit trail. This is where a system like Arize or Langfuse becomes non-negotiable for production deployments, not just development.

The Moment It Actually Worked: My Concrete Love

Despite the headaches, there were moments of pure magic. My concrete love? One specific instance where the agent truly shone. We had a prospect who had opened our initial email about a new analytics feature but hadn’t clicked through. Two weeks later, the agent noticed this, pulled up their company’s recent funding announcement (thanks to a quick data enrichment step via Clay.com — honestly, if you’re doing outbound, you need to be enriching your data), and crafted a two-sentence follow-up. It read something like: “Saw your recent funding round – huge congrats! Just thought our new analytics dashboard could be especially useful as you scale your team post-raise.”

That email got a reply within an hour. A booked meeting, directly attributable to the agent’s ability to synthesize context and respond with genuine relevance. That’s the power I was chasing. It wasn’t just automating; it was augmenting our sales process. For someone who isn’t keen on building all that infrastructure themselves, platforms like Lindy SDR agents or Bardeen are trying to abstract away some of this complexity, offering pre-built agentic workflows for sales and support. They’re not perfect, but they’re a hell of a lot faster to get running.

Build vs. Buy: What’s Worth Your Money?

So, should you build your own agent for automating follow-ups with AI, or buy a platform? It’s a classic dilemma. If you’ve got a dedicated engineering team, a complex, unique workflow, and you’re comfortable with the debugging and governance overhead, building with frameworks like LangGraph or even a custom Python script using the Vercel AI SDK is viable. The free tier of the Vercel AI SDK is enough to prototype, but costs scale quickly once you hit production volumes. You’ll need to account for your LLM costs, too, which can easily run into hundreds, even thousands, of dollars a month depending on volume and model choice. The free plan for some of these agent platforms, though, is often a joke; it’s just a demo. You’re not actually getting enough usage to truly test it at any meaningful scale.

Honestly, for automating follow-ups with AI at scale without dedicated engineering, a platform like Lindy is probably the only one I’d actually pay for right now if I wasn’t building it myself. Their focus on sales-specific tasks means they’ve already thought through some of the prompt engineering and error handling you’d otherwise be figuring out from scratch. $199/month for their pro tier, which gives you decent volume and features, is fair if it actually delivers on its promise of saving SDR time and booking more meetings. Otherwise, you’re just paying for another tool to manage. If you’re a solo operator or a small team, a more general automation tool like Bardeen could work for simpler tasks, but it won’t give you the deep personalization of a dedicated agent.

If you want the deep cut on this, AI agent platforms coverage.

Ultimately, the value isn’t in the automation itself, but in the quality of the interaction it enables. Don’t just automate; enhance.

— The Colophon

One AI tool. Tested. Reviewed.
In your inbox every Sunday.

~3 minute read. Real outcomes from operators, not marketers.

— More like this
Outbound Tools

The Reality of Best AI-Powered Sales Dialers in 2026

As a builder, I've deployed AI-powered sales dialers. Here's what actually works, what breaks, and if these tools are worth the cost for your sales team.

7 min · May 29
Outbound Tools

How to Train AI for Sales Scripts That Actually Convert

Stop wasting time with generic AI. Learn how to train AI for sales scripts using your own data, ensuring brand voice, compliance, and higher conversion rates.

8 min · May 29
Outbound Tools

Email vs LinkedIn Outreach Automation: What Actually Works in 2026

Comparing email vs LinkedIn outreach automation for B2B sales in 2026. Learn which channel delivers real results and avoids compliance headaches.

6 min · May 29