Using AI tools for sales has a specific tension worth naming upfront: sales is fundamentally a human relationship discipline, and the most successful salespeople I’ve observed using AI tools are the ones who use them to have more and better human conversations, not fewer. The misuse of AI in sales — automated outreach at volume without genuine personalisation, AI-generated scripts that make conversations feel scripted, reliance on AI analysis over listening to what a prospect actually needs — produces worse results than not using AI at all. For a broader walkthrough, our AI Tools for Every Industry is a good next read.
The productive use removes the mechanical overhead — research, admin, first-draft writing, CRM data entry — so more time goes into the human work that actually closes deals. The specific tasks where AI tools deliver measurable value in sales fall into three categories: pre-call research and preparation, outreach and communication drafting, and post-call administration and follow-up. In all three, AI handles the mechanical work. The selling — understanding the prospect’s real situation, building genuine rapport, identifying the right solution, navigating the decision process — remains entirely human.
Prospect research — where AI saves the most time
Thorough research on a prospect before a conversation measurably improves the quality of that conversation, and AI tools make thorough research faster than it has ever been. The workflow that produces the best results:
Company research with Perplexity AI (free tier) — search the company name and review recent news, funding announcements, product launches, challenges mentioned in press coverage, and executive changes. This surfaces current context that generic CRM data doesn’t capture and that takes 30+ minutes to find manually through multiple searches.
Individual research with Claude or ChatGPT — paste the prospect’s LinkedIn profile summary into an AI tool and ask it to identify what this person likely cares about professionally, what their priorities probably are given their role, and what questions might be most relevant to ask them. This gives a starting hypothesis about the prospect’s perspective before the conversation begins — not a script, but an informed frame.
Synthesising into a call brief: “Based on this company and prospect research, what are the three most relevant angles for a first conversation and what questions should I ask to understand their current situation?” produces a focused call preparation brief in minutes rather than a scattered set of research notes. For a 45-minute first meeting, spending 10 minutes on AI-assisted research produces noticeably better conversation quality than walking in cold or with only basic CRM information.
Outreach and prospecting — personalisation at a manageable pace
For cold email and outreach drafting, AI tools are most useful for generating personalised first-draft messages that a sales professional reviews, adjusts, and sends — not for generating and sending at volume without review. The distinction matters enormously for effectiveness: personalised messages that reference something specific about the prospect have measurably higher response rates than templated messages, and the personalisation requires human judgment about what is genuinely relevant versus what sounds like it was scraped from a LinkedIn profile.
The effective AI outreach workflow:
- Research the prospect and identify one specific, genuinely relevant angle — a company announcement, a challenge mentioned in a podcast, a recent hire that signals a strategic shift
- Give Claude or ChatGPT that specific angle and ask for a short (5–7 sentence) outreach message that opens with that reference and connects it to a potential area of value
- Review the draft, adjust the opening to sound like you rather than AI, verify the specific details are accurate
- Send
This takes 5–7 minutes per message rather than 15–20 minutes from scratch, with output quality comparable to what you would write with more time. The human contribution is the research judgment about which angle is genuinely relevant and the voice adjustment that makes it sound personal. Without both, the AI draft is a template with a LinkedIn fact bolted on — visible to experienced buyers and less effective than a shorter but genuinely personalised message.
Apollo.io ($49+/month) and Outreach (enterprise pricing) are the sales engagement platforms with AI features for outreach at scale — AI-generated personalisation at the sequence level, AI-suggested follow-up timing, and AI analysis of which message types are getting responses from which prospect profiles. For sales teams doing significant outbound volume, these platforms reduce the manual work of managing sequences without sacrificing the personalisation that makes outreach effective.
CRM and sales intelligence
Salesforce Einstein AI (within Salesforce, enterprise pricing) and HubSpot AI (within HubSpot, various tiers) provide AI-powered sales intelligence within the CRM — lead scoring that prioritises the prospects most likely to convert, deal health scoring that identifies at-risk opportunities before they go dark, AI-generated call summaries, and conversation intelligence that captures key information from calls and emails automatically rather than requiring manual CRM updates.
The CRM data entry elimination is one of the highest-value AI sales applications for individual salespeople. Manual CRM data entry after every call and email consumes 15–30 minutes daily and directly competes with time in actual sales conversations. AI tools that automatically capture conversation data, update CRM fields, and suggest next actions remove this overhead and improve CRM data quality simultaneously — because manual entry is prone to both omission and error, and AI capture is neither.
Gong (enterprise pricing) is the conversation intelligence platform most widely used by sales teams — recording, transcribing, and analysing sales calls to identify what distinguishes winning from losing conversations, what objection handling approaches are most effective, and which topics correlate with deal advancement. For sales leaders coaching teams and for individual salespeople who want to understand their own performance patterns, Gong’s AI analysis of conversation data provides insights that are genuinely difficult to obtain from any other source. The pattern identification across hundreds of calls reveals what actually works versus what salespeople believe works — often different things.
Proposals and presentations
Proposal creation is another high-value AI sales application — the mechanical work of taking discovery conversation information and structuring it into a compelling proposal consumes significant time that AI tools compress substantially.
After a discovery call, write up the key findings — the prospect’s stated challenges, their goals, their current situation, and what success would look like — and give this to Claude or ChatGPT with a brief about your product or service. Ask it to draft a proposal structure that maps your solution to their specific stated needs. The draft proposal takes 10 minutes to generate and 30–45 minutes to review and personalise with specific pricing, timelines, and detailed product/service specifications. The total is substantially faster than building the proposal from scratch while producing an output that is genuinely customised to the specific prospect’s situation.
Sales AI tools reference
| Sales task | Best AI tool | Time saving | Still human work |
| Pre-call research | Perplexity + Claude | High — 30 min to 5 min | Interpreting relevance to specific prospect |
| Cold outreach drafting | Claude or ChatGPT | Medium — 15 min to 5 min | Personalisation judgment; voice adjustment |
| CRM data capture | Gong or Salesforce Einstein | Very high — eliminates manual entry | Review and correction of AI capture |
| Lead prioritisation | Salesforce Einstein or HubSpot AI | High — focuses time on best opportunities | Context not captured in CRM data |
| Proposal drafting | Claude or ChatGPT | High — 2 hrs to 45 min | Pricing, specific terms, relationship judgment |
| Conversation intelligence | Gong | Indirect — improves future calls | Applying insights to specific relationships |
| The actual selling | Not appropriate for AI tools | N/A | Everything — relationships, judgment, closing |
Follow-up and long-cycle nurture
The follow-up consistency that distinguishes top-performing salespeople from average ones — the discipline to send the right message at the right time through a long sales cycle — is exactly the kind of systematic work that AI tools support well.
AI-generated follow-up drafts based on the previous conversation, reminders based on CRM signals, and suggested content to share based on prospect interest profile reduce the cognitive load of maintaining consistent follow-up through long sales cycles with large prospect pipelines. Apollo.io’s sequence intelligence suggests follow-up timing and content based on what’s working across similar prospects. Gong surfaces the moments in past conversations that should inform the next touchpoint. HubSpot’s AI drafts follow-up emails from conversation summaries and CRM context.
The consistent pattern across all of these: AI identifies the opportunity and drafts the message; the salesperson adds the personal judgment and authentic voice that makes the follow-up land. Taking an AI-drafted follow-up, editing it to sound exactly like something you would write, and sending it takes 3 minutes. Writing the follow-up from scratch takes 10. Across a pipeline of 50 active opportunities requiring weekly touches, that 7-minute difference per touch compounds into hours of recovered time per week.
Building an AI-assisted sales practice
The salespeople who get the most from AI tools have usually gone through a similar adoption sequence: starting with pre-call research (immediate value, low risk), then outreach drafting (time saving on a high-frequency task), then CRM automation (eliminating the most disliked administrative work), then conversation intelligence (the most powerful but also the most involved adoption). Each step delivers value that makes the next step feel worth the investment.
The trap to avoid: using AI tools to scale outreach volume without maintaining personalisation quality. Sending 200 AI-generated “personalised” emails per day that are identifiably templated produces worse results than sending 30 genuinely personalised messages. Buyers have become adept at recognising AI-assisted outreach that wasn’t sufficiently reviewed — and they discount it accordingly. The effectiveness of AI-assisted sales outreach depends entirely on the quality of the human review and personalisation layer that comes between the AI draft and the sent message. Our guide on writing better prompts covers the specific prompting techniques that produce the best AI-generated outreach and proposal drafts. Our guide on best AI tools for small business covers the sales and CRM tools most appropriate for smaller operations where enterprise platforms aren’t cost-justified.
Sales coaching and performance improvement
One of the most compelling but least-used applications of AI in sales is for coaching and skill development. Gong’s conversation intelligence doesn’t just summarise calls — it enables systematic analysis of what distinguishes high-performing salespeople’s conversation patterns from lower-performing ones, and it makes those patterns visible in ways that anecdotal coaching from memory cannot.
Specific coaching applications:
- Talk ratio analysis: Gong surfaces the percentage of each call spent talking versus listening. High performers in discovery-oriented sales tend to listen more than they talk; AI analysis of talk ratio across the team identifies who needs to develop their listening practice.
- Question patterns: the types of questions salespeople ask correlate with deal outcomes. AI analysis of conversation transcripts identifies which question types appear more frequently in won deals and which are more common in lost deals.
- Objection handling patterns: what responses to common objections actually advance the conversation versus what responses tend to stall it. AI analysis across many calls reveals patterns that individual memory from specific deals cannot.
- Topic timing: when in the conversation certain topics (pricing, competition, implementation) are introduced correlates with outcomes. AI analysis surfaces whether raising certain topics too early or too late is hurting deal progression.
For sales leaders, this AI-generated insight transforms coaching from “I listened to three of your calls and here’s my impression” to “across your last 20 calls, here are the specific patterns that differ from top performers and here are examples.” The specificity makes the coaching more actionable and the development more targeted.
The limits of AI in sales — worth being explicit about
Several sales capabilities remain firmly human territory regardless of how AI tools improve:
Reading the room. The non-verbal cues, the subtle shift in energy, the moment when a prospect’s engagement changes — these are signals that experienced salespeople read and respond to in real time. AI tools can analyse recorded calls after the fact; they cannot read the room during the conversation.
Building genuine trust. Trust in a sales relationship is built through consistent follow-through, genuine interest in the prospect’s success, and the kind of honest conversation that requires one human choosing to be vulnerable with another. AI tools can help you prepare for trust-building conversations; they cannot build the trust itself.
Creative deal structuring. Complex enterprise deals often require creative thinking about how to structure an agreement that works for both parties given unusual constraints or requirements. This problem-solving draws on accumulated deal experience and contextual judgment that AI tools can inform but not replicate.
Navigating internal politics. Understanding who really makes decisions in a prospect organisation, who the internal champion is and what they care about, and how to navigate competing stakeholder priorities are skills that require human intelligence about human organisations in ways that AI analysis of publicly available information cannot adequately approximate.
The most effective AI-assisted salespeople use AI to be more prepared for every conversation, more organised in their follow-up, and more efficient with their administrative work — so they have more time and cognitive capacity for the genuinely human work that these tools cannot touch. That division of labour, consistently applied, is where the measurable performance improvement actually comes from. Our guide on How to Use AI Tools for Presentations covers an adjacent issue.
