AI in sales: scalable but not a substitute

ryanseo

i've been testing AI tools for sales this year, and the honest take? They're useful, but only if you treat them as leverage-not a replacement for actual selling. We use a mix of LLMs and custom frameworks, but full automation still feels like shouting into the void. AI-written outbound gets ignored fast, it's obvious when there's no human touch.

Where it actually helps:

Research cleanup and prioritisation. Summarising accounts, spotting buying signals, cleaning CRM junk. That stuff saves hours.
Warm lead follow-ups, but only after we've picked up interest signals manually. we use Sales Navigator for prospect research and scraping, then Expandii for automating LinkedIn messaging-but capped at 20 per day, and only on strict conditions.

The key is setting up the workflow to do exactly what you want, tuned to send the specific follow-ups on the right triggers. Leave any agentic framework alone too long, and it'll go rusty. We keep everything human-run at the core.

TLDR: AI is a necessity for scaling sales, but it's just leverage. if you don't know how to sell effectively, no tool will fix that.

sophiasaas

I've been using AI tools in our sales process too, but I treat them like a sharpened chisel - useless if the sculptor doesn't know what they're doing. For us, they're purely supplementary: they speed up delivery without sacrificing quality, but we keep strict boundaries on what they can and cannot touch. Everything still runs through human oversight because if you leave any AI framework alone for too long - especially a custom agentic one - it starts going rusty.

On the tools side, Sales Navigator is essential for prospect research and scraping, and I use Expandii for automating LinkedIn messaging. I cap it at 20 per day, and ideally only fire messages after we've spotted real interest signals. Works well if you configure it down to the exact follow-ups and conditions you want. Same goes for any automation in our outbound - the result is determined entirely by how you set up the workflow.

In my corner of SaaS growth, this all ties back to LTV:CAC and churn reduction. If you use AI to blast out 200 messages a day but your onboarding funnel leaks like a sieve, you're just burning money. The tools are leverage, not a replacement for actually understanding how to sell - or in my case, how to build a product-led experience that closes the loop.

funnelhacker

Straight to ROI: AI for research cleanup and prioritisation gives us the biggest time save per dollar spent. Summarising accounts, flagging buying signals, scrubbing CRM junk-those are low-effort wins that actually scale. Full AI-written outbound? Reads like boilerplate and gets binned before the second sentence. No metric on earth fixes that.

davidsearch

Yeah, using them heavily here too. Mostly for intent scoring, call summaries, and drafting initial outreach sequences. The whole "AI SDR replaces the team" narrative still feels like VC-funded clickbait though - most fully automated outbound I've seen tanks CTR fast because it reads like generic slop.

The real wins came from tools that prioritise intent signals over volume. I've been testing Sendio AI for LinkedIn prospecting, and it's one of the few that keeps the conversation path feeling human rather than robotic. Timing and context consistently beat blasting 10k lists, same as with PPC - CPA drops when you match the right message to the right search moment.

liamcopy

Prospecting AI is a mixed bag.
depends entirely on the use case.

For cold outbound at scale?
Sales Co beats any DIY LLM setup.
Hands down.

olivia-roas

Yeah, using AI for sales now. But it's not doing the selling - it's doing the soul-crushing admin that eats up your morning. Stuff like scraping intent signals, writing up the cold message drafts, sorting replies by category so I don't waste time on the ones who clearly want to unsubscribe. That alone saves us around 4-5 hours per head per week.

Where it falls flat? Reading the room. AI can't tell when a prospect is ready to get a bit of pressure, can't sense the tone in a quick Slack-style reply, and absolutely crumbles on the weird edge-case objections that come out of nowhere.

Treat it like a hyper-fast junior assistant, not another rep. You still have to do the actual sale.

noahcontent

Absolutely. For prospecting, qualification, and research, AI handles the repetitive touches beautifully and surfaces warmer leads faster. The hybrid approach is key though - human judgment still needs to take over once buying intent or complex objections appear. I've seen frameworks using tools like 11x for exactly that kind of workflow.

miaaffiliate

Same here. Built a roleplay agent specifically for training new affiliates. Went from a three-month ramp down to under two weeks. Cuts straight through the bullshit-no wasted time on theory, they get live scenarios from day one. Margins thank me later.

olivia-roas

I run marketing and pipeline's one of my KPIs. Inbound's been fluctuating like crazy, so outreach falls on my plate. honestly, I don't have the bandwidth to hand-write every sequence, and i don't think I should have to.

AI handles the heavy lifting. Agents analyse sales calls, retention data, CRM-digs out the actual pain points. Then i feed my short blurbs into our own tool: it builds the ICP, spits out messaging angles and value props, designs the campaign with touchpoints and channels, and runs the sequences. I review and approve before anything goes live, that's it.

Works for me because I'm measured on pipeline delivered, not copy crafted. if your day job is writing the copy, different story. but for anyone responsible for the outcome, not the doing, AI's the only way to get it done at scale.

ethan-pr

i've found AI genuinely useful for streamlining the repetitive side of sales-meeting summaries, first-draft outreach. the relationship building remains human work. my team pairs those tools with Highnote for deck sharing and tracking post-call engagement, that combination actually holds up in daily use.

justauser

AI definitely takes the edge off the repetitive stuff 🙌 but that human connection-trust and reading the room-still makes or breaks the deal

avadata

We're running a few for prospecting too. The consistency is real-nobody drops the ball on follow-ups at 2am. But if you're not layering in intent signals and firmographic filters, you're just blasting noise. I've seen teams claim 3x reply rates because they spammed 10k accounts with a LLM-generated sequence. That's not scale, that's burning pipeline. The trick is treating it like a multiplier, not a replacement for strategy. You still need real ICP definitions and message iteration. Otherwise it's just faster garbage.

lucasvid

AI for meeting notes? Same here. instead of splitting focus between talking and scribbling, i just stay locked into the conversation. the transcript gets fed into a summariser that spits out a follow-up email and CRM notes. feels like having an editor handle the boring bits while you focus on the actual performance - retention on the call goes way up when you're not distracted by admin