AI moves grunt work to agents and shifts humans to persuasion and risk-reduction in-meeting; winners optimise time-to-answer and cut no-decision losses.
The Past: Sales Has Always Adapted to New Tools
Sales is often called the world’s second-oldest profession, and it has survived countless transformations. In the past, selling was 100% human-driven, a craft honed through face-to-face meetings, phone calls, and personal charm. Every time a new technology emerged, people predicted the end of the salesperson. For example, when e-commerce and digital channels rose in the 2010s, some analysts forecast a mass extinction of B2B sales jobs. In 2015, Forrester famously predicted that one million B2B salespeople would be displaced by 2020 Forrester.com. Dire headlines warned that websites and automation would make human reps obsolete.
Yet those dire predictions didn’t fully come true. Instead, sales teams adapted. Early CRM systems in the 1990s didn’t replace salespeople, they empowered them to track customers better. Email and LinkedIn didn’t eliminate relationship-building – they added new channels to reach buyers. The past teaches us a clear lesson: sales as a profession has a knack for absorbing new tools rather than being obliterated by them. Each technological leap has augmented the salesperson’s abilities, not rendered them useless. The fundamental human elements of sales - building trust, understanding customer needs, and persuading through storytelling remained as vital as ever. As one sales veteran might put it: “Tools change, but people buy from people.” In fact, history is rich with examples of technology creating new opportunities for sellers. When James Watt coined “horsepower” to sell steam engines, it was a creative marketing tactic that helped customers grasp an unfamiliar technology. This pattern repeats: the best salespeople leverage new tech in creative ways to connect with customers, rather than letting the tech replace genuine human connection.
If our ancestors had a CRM, it was a cave wall. We weren’t the strongest species—we were the ape with accessories. Fire extended our day, spears extended our reach, writing outsourced memory, the telephone collapsed distance, and the web collapsed search costs. Every great tool moved some job from “inside the skull” to “outside the skull.” AI is simply the next move in that long march: it collapses figuring it out slowly into figuring it out now. The question isn’t “Will tools replace us?” It’s “Which parts of selling do we want to offload so the human parts can matter more?
So, looking at the past, we see a narrative of adaptation… Is AI just the next tool sales teams will adapt to, or something fundamentally different? To answer that, we must examine what’s happening right now.
The Present. AI Enters the Sales Arena
Fast forward to today, and artificial intelligence has stormed into the sales world. What was once hype is now reality: four in five sales teams are either experimenting with or have fully implemented AI, according to Salesforce’s 2024 State of Sales (6th ed.). At the enterprise level, McKinsey’s 2025 global survey finds 78% of organizations use AI in at least one business function, with marketing and sales among the top adopters—so AI in sales is no longer a niche experiment. Assets McKinsey & Company
Why the momentum? Early results are material. Salesforce reports that 83% of sales teams using AI saw revenue growth versus 66% of teams without AI. In practice, that spans AI “copilots” drafting personalized outreach, models that prioritize and forecast, and assistants that handle initial customer queries—freeing reps to spend more time on judgment, negotiation, and relationships. Salesforce
It’s telling that sales and marketing saw the greatest spike in generative AI adoption of any business function from 2023 to 2024 mckinsey.com. Only about one-fifth of companies have fully deployed AI in their sales process so far (another fifth are running pilots) mckinsey.com, but those who have are very excited by what they’re seeing. Over 85% of sales leaders who’ve started using generative AI say they’re “very excited” about its potential*, citing improved efficiency, higher sales growth, and better customer experiences as key benefits mckinsey.com. This optimism is palpable across the industry. It feels like we’re at the dawn of a new era – one where AI is a standard part of the sales toolkit.
Across the enterprise, 78% of organizations now use AI in at least one business function, with marketing and sales among the most active adopters. McKinsey & Company
To understand the present state of AI in sales, consider a few concrete examples of how it’s being used today:
- Lead Qualification and Research: Remember when junior staff spent hours researching prospects or qualifying cold leads? Now, AI can handle much of that groundwork. Modern sales teams deploy AI algorithms to scour databases and the web for ideal prospects, or to automatically qualify inbound inquiries. Repetitive tasks like gathering company info, finding contacts, and logging activities can be offloaded to AI.
- Personalized Outreach at Scale: Crafting a tailored email for a single prospect is easy; doing it for 1,000 prospects is hard. Generative AI is tackling this by writing surprisingly human-like emails and messages. For example, AI-driven outreach platforms analyze a prospect’s industry, role, and even social media activity to generate emails that feel personal. This hyper-personalization at scale was once nearly impossible. Sales reps still review and tweak these AI-written drafts to add their personal touch, but the heavy lifting of first-draft writing is handled by the machine.
- CRM and Data Entry Automation: Ask any salesperson what they dislike about their CRM, and you’ll hear one answer: the data entry. Updating contacts, logging call notes, setting follow-ups, these eat into selling time. Today’s AI assistants transcribe calls, summarize key points, and push structured updates into the CRM automatically. Two things are true at once: sellers still spend only ~28–30% of the week “actually selling,” and early adopters report more time with customers and less back-office drag as they automate the grunt work. Broad workforce studies also show >50% of Gen-AI users save 5+ hours/week. Treat this as real reallocation of time, not an overnight 2× jump. Looking ahead, Gartner expects 95% of seller research workflows to start with AI by 2027 (from <20% in 2024), a strong signal of where admin time is going. Salesforce McKinsey & Companybcg.com Gartner
- Forecasting and Analytics: AI is lifting forecasting and pipeline analytics: models scour historicals, comms signals, and market context to surface risk and next-best actions. Reality check: industry-wide, only 7% of orgs hit ≥90% forecast accuracy; the median is 70–79%. That’s why “95% accuracy” should be framed as case-study outcomes under strong process + data hygiene, not a baseline. For example, Upwork (Gong) and Carbon Black/Tungsten (Clari) report landing within ~5% of actuals after revamping process and tooling. Use such figures as an upper bound and aim to beat your own baseline. Gartner gong.io Clari
- Training and Coaching: AI is also changing how reps are trained and coached. Conversation-intelligence systems flag talk ratios, missed value points, and coachable moments; simulators role-play tough customer scenarios for low-stakes practice. Case studies report meaningful ramp gains, e.g., Addepar cites ~60% faster onboarding with Gong. Outside sales, immersive/AI-based learning is repeatedly shown to compress time-to-competency (Deloitte and others), but treat those as contextual evidence, not a sales-specific benchmark. AI can shorten ramp and improve coaching quality when it’s embedded into real workflows and measured. gong.io Training Industry Deloitte
Human + agent is the new default: assistants handle the repetitive mechanics; humans handle persuasion, framing, and trust.
The day-to-day is already hybrid: an assistant suggests who to call, drafts the first pass, logs the meeting, and highlights risks, while the rep uses judgment, empathy, and narrative skill to move the deal. Microsoft’s 2025 Work Trend Index even formalizes the idea of human-agent teams as the new operating model. That’s the arc: automate the repetitive, amplify the human. Microsoft
Friction still exists and matters
Integration is messy: data quality limits model value, change management is non-trivial, and vendor claims vary. The fear of falling behind is real. Salesforce reports ~78% of sales leaders worry about missing out on Gen-AI’s benefits, but rushing in without a plan backfires. Pilot with clear KPIs (time-to-answer in calls, win-rate lift vs. “no-decision,” forecast error), then scale what works. Salesforce
In summary, the present state of sales is human - AI collaboration. The average day of a salesperson in 2025 might involve an AI assistant suggesting who to call next, drafting the first pass of outreach, updating the CRM after the call, and even surfacing coaching cues, while the rep applies judgment, empathy, and narrative skill. As Microsoft’s 2025 Work Trend Index puts it, orgs are moving toward human-agent teams, where agents take on specific tasks under human direction.
However, not everything is rosy. Many organizations are still figuring out integration: data quality, change management, and tool selection remain hard problems. And the fear of falling behind is real. 78% of sales professionals worry about missing out on generative AI’s benefits, underscoring the pressure to adopt while staying thoughtful about trust and control. Salesforce
The present is a balancing act: use AI enough to stay competitive, but don’t lose the human touch and strategic oversight. That sets the stage for the real question: what’s next, and how do the competing visions of the future play out?
The Future. Two Visions of AI’s Impact on Sales
Peering into the future, we find two contrasting visions for how AI will shape sales. On one side are the automation enthusiasts, those who believe AI will fundamentally reinvent the sales profession, perhaps eliminating many sales roles. On the other side are the human-centric champions, those who insist that no matter how smart machines get, sales will remain a deeply human endeavor at its core. The likely reality is somewhere in between, but exploring these extreme viewpoints helps illuminate the road ahead.
Vision 1: The AI-Driven Sales Machine (Automation Enthusiast)
In this vision, the sales organization of the future is heavily automated. The mantra here is “If it’s repetitive, a machine should do it”.. Advances in AI could handle not just data entry or email drafting, but end-to-end sales processes for simpler transactions. Imagine an AI Sales Development Representative (SDR) that can initiate contact with a prospect, educate them about a product, and schedule a demo, all via natural-sounding, AI-generated communications. We’re already seeing early signs of this: some companies have AI “bots” reaching out to leads or following up after a webinar. As the technology improves, these AI agents will only get more capable. The “autonomous AI agent” trend suggests future AI will be able to execute multi-step sales workflows independently. In fact, Gartner predicted that 35% of Chief Revenue Officers would establish a generative-AI operations team within their go-to-market organization by 2025 Gartner X.
Under this paradigm, many traditional sales roles might shrink or disappear. Transactional sales (where products are simpler and buying decisions are lower risk) could become almost fully self-service, guided by AI. Picture a scenario where a business buyer interacts with a virtual sales agent on a website, asks questions, gets instant tailored answers, negotiates price (the AI has authority within a range), and clicks to purchase – all without ever speaking to a human. It sounds far-fetched, but we’re not too far off technologically. If buyers accept AI agents, companies might decide that human reps aren’t needed for certain product lines or customer segments. Indeed, roles that today are already being affected include junior reps focused on cold outreach or basic qualification, much of that is getting handled by AI tools scanning behavior signals and sending perfectly-timed messages. Inbound lead qualification? AI chatbots can often sort out the serious buyers from the tire-kickers. Sales support roles like list researchers or data cleaners? Those are being replaced by AI that can compile prospect lists and cleanse CRM data in minutes. There’s even talk that AI could soon handle the dreaded cold calls. It isn’t just vendors saying this. SaaStr’s Jason Lemkin notes customers often prefer great AI over mediocre humans for routine sales interactions. Mainstream coverage shows AI voice agents already taking calls and in some cases closing transactions, though top outlets also warn general-purpose agents remain uneven in the wild. The direction of travel is clear (OpenAI’s own line is that agents are ‘joining the workforce’), but the smart play is narrow, compliant, supervised deployments—especially since the FCC now treats AI voice robocalls as illegal without consent.” SaaStr The Wall Street Journal The VergeStratechery by Ben Thompsonfcc.gov
As AI takes on more of these tasks, the composition of sales teams would change. We might see smaller human teams, focused only on the most complex, enterprise-level deals or on crafting strategy, while fleets of AI agents handle the volume. Companies shorten sales cycles, reduce costs, and squeeze more ROI from a leaner salesforce. Sales becomes a high-tech operation with humans managing the machines: a sales manager might oversee a dashboard of AI agents just as much as they coach human reps.
It’s a compelling picture for CFOs looking at the bottom line. However, it comes with a big assumption: that customers will happily buy from robots. And here is where Vision 2 raises a skeptical eyebrow.
Vision 2: The Human-Centered Sales Renaissance - Human-Centric Optimist
No matter how smart the tooling gets, big deals still hinge on human trust. People buy from people, because reading the room, surfacing the unspoken objection, and de-risking a risky choice are deeply human skills. In complex B2B, a great seller isn’t just a messenger; they’re an interpreter of motives and a reducer of fear.
Authenticity beats auto-tone. Steven Bartlett has leaned into AI with 100 CEOs, an AI-voiced spin-off he explicitly labels as machine-generated, while publicly warning that unlabeled, AI-trained content can erode audience trust. The sales translation is simple: scale with AI, but make the you-ness obvious. In a feed flooded with polished, generic outputs, a hand-written note or a 60-second personal Loom can cut through precisely because it couldn’t have been auto-generated. Business InsiderThe Times
As AI takes the admin and synthesis work, the rep’s job shifts upmarket: more consultant than “walking brochure,” more problem-solver than presenter. Let the models crunch, draft, and retrieve; the human reframes, prioritizes, and earns the yes.
Equip for the hybrid. Sellers who can actually wield AI outperform: Gartner reports those proficient with AI are 3.7× more likely to hit quota, yet only a small minority have those skills today. Upskilling is not optional. Destination CRM
Net-net: the future isn’t AI vs humans; it’s AI + humans. Automate the repeatable; make the human moments unmistakably human.
Navigating the AI shift: what to do now
Think “AI with me,” not “AI instead of me.” Use agents to do the mechanical work (research, notes, first-draft outreach), then spend the saved minutes on judgment, framing, and trust. That’s the operating model Microsoft calls human–agent teams. Microsoft
Buy tools that change behavior at the moment of truth. Prioritize systems that help in the meeting (grounded answers, next step recommendations, de-risk options), not just more dashboards. Success looks like faster time-to-answer and fewer “no decision” stalls, not just prettier reports. (Salesforce also shows AI-using teams are more likely to grow revenue: 83% vs 66%.) Salesforce
Measure outcomes, not vibes. Instrument: time-to-answer on live objections, win-rate vs. “no decision,” forecast error, and % of rep time reallocated to customer work. Expand what works; cut what doesn’t.
Make trust non-negotiable. Label bots clearly. Require citations/provenance for claims. Build consent into recording and analytics flows. In the UK/EU, note that the EU AI Act prohibits “emotion recognition” in workplaces (with narrow exceptions), so avoid face-reading to judge staff or buyers. Use speech/text signals instead. The UK ICO has also flagged emotion-recognition risks. artificialintelligenceact.euc 1ico
Pick for fit, not flash. Insist on sub-5s p95 latency on your data, CRM/meeting-native workflows, and auditable evaluations (faithfulness/grounding on your docs). Fancy demos don’t survive real calls.
The 60-Second Sales AI Fit Check
- Does it save seller time this week? (Research, notes, follow-ups.)
- Does it improve call quality? (Clear recommendation, fewer rabbit holes.)
- Is it grounded? (Links + quotes for every factual claim.)
- Is it compliant in my region? (Consent, data minimization; no emotion-recognition on people at work in the EU.)
- Will reps actually use it? (Lives in CRM/Meet/Zoom; minimal screen-hopping.)
This article is for general information only and is not legal, financial, or technical advice.
By Vanja Eriksson, Co-Founder Headsum