Singapore
AI adoption did not close the gap to the decision-maker

AI adoption did not close the gap to the decision-maker

Sellers have adopted ai automation for business for prospect research and email drafting, and 77% of them still name finding and accessing the right decision-maker as their hardest prospecting task, in a study ValueSelling Associates released on 16 November. The adoption figures and the challenge ranking come from the same respondents, which is what makes the pair worth reading together: the bottleneck sits upstream of the software. The scene has not changed either. A sales manager opens the record for an enquiry that arrived over the weekend and finds a mobile number, a first name and no company, so the thread goes to whoever answers first. Access to the buyer is one half of the problem, and the other is the process that carries an enquiry until someone with authority answers.

Key takeaways

  • AI use among sellers is concentrated in preparation: research, email drafting and call notes, with personalisation close behind.
  • Reaching the right decision-maker is the top prospecting challenge for 77% of sellers, ahead of every tooling complaint in the same list.
  • The training gap is measurable from both sides: 80% of sales leaders say they provide the training, 50% of reps agree.
  • Coaching frequency predicts quota attainment better than any tool, and 76% of sellers say they need more of it than they get.
  • The operational answer is a sequence rather than a purchase: capture each enquiry, identify the account, keep a visible stage, and send a person in when the first contact cannot decide.

What the study set out to measure

The study is a repeat measurement as much as a survey. It goes back to questions first asked in a similar exercise in 2018, so the answers can be read as a before and after: what years of sales technology adoption changed about prospecting, and what stayed where it was. An adoption figure on its own tells you nothing about whether the work got easier. A team can report heavy AI use and a harder quarter in the same conversation.

The questions carried over from 2018

What carried over is mostly the human side of prospecting: whether a rep can get to the person who signs, how they handle a call with a busy executive, and whether outreach gets done consistently enough to matter. Julie Thomas, the company's president and CEO, reads the comparison as continuity. In her statement, the teams pulling ahead this year are winning on capabilities that decided the day in 2018 as well, and she names referrals, confidence, persistence that respects the buyer's time, and the ability to hold a business conversation. She treats AI as support for those capabilities rather than a substitute for them.

Where AI is already doing the work

The adoption figures are specific enough to act on. Among the sellers surveyed, 55% use AI to draft prospecting emails, 55% use it to research prospects, 46% use it to personalise outreach, and 43% use it to build call-prep notes. Those four uses sit in one layer of selling: the reading, the writing and the background work that happens before a conversation starts. Not one of them touches the conversation itself, and none of them decides who receives it.

Research, drafting and personalisation

What the numbers describe is a rep who arrives at a call better briefed than the rep in the 2018 study was, with a first draft in the outbox and a summary of the account on screen. That is a genuine change, and most of it is ai automation for business doing the reading and writing before a person takes over. It also has a ceiling that shows up quickly. Preparation helps only when the conversation happens with someone who can say yes, so a well-written email to a user who cannot approve a budget is a careful piece of work pointed at a dead end. The same ranking suggests that is where much of the effort lands, since the tooling complaint sits at the bottom of the list and the access complaint at the top.

Reaching the right decision-maker in B2B sales

The challenge list puts access first and software last. Finding and accessing the right decision-maker is a challenge for 77% of sellers. Giving up too easily on cold calls is named by 35%, outreach that is neither organised nor consistent by 27%, and being unskilled at using sales tech for prospecting by 24%. The distance between the first figure and the last is the story. On their own account, sellers are not held back by technology they cannot operate. They are held back by not reaching the person who decides, and by the habits that would carry them there when the first attempt fails.

A salesperson holding a phone pauses by a corridor of closed office doors, a contact list open on the desk behind.

Why a better email does not fix a wrong recipient

AI sharpens the message and leaves the recipient untouched. Research can tell a rep the company, the role, the recent announcement and the likely concern. It cannot tell them whether the person reading has authority to approve. That judgement comes out of a conversation, a referral, or a record that already holds the answer. The working rule is to check authority before spending a sequence on a name: ask a question that reveals who signs, and treat a reply that avoids the question as information rather than as momentum. A useful companion read is the note on what changes when emails start answering questions, because a quicker reply from a junior contact can look like progress while the authority question is still open.

What the smaller challenges add

The other answers describe what happens after first contact. A rep who gives up on cold calls too easily never reaches the decision-maker on the second or third attempt, and 35% admit to that pattern. A rep who cannot keep outreach organised loses the thread between attempts, which is the 27%. The 24% who say they are unskilled with sales tech are telling a smaller story: most sellers do not believe their problem is the software.

The training gap behind the tooling gap

The study measures training from both ends, and the two answers do not meet. 80% of sales leaders say they provide sales training on the skills needed to prospect and sell effectively. 50% of reps agree. One side counts sessions delivered. The other counts skills that survived contact with a live deal, and the distance between the two numbers is the space where reinforcement was supposed to sit. A quarterly group session is a poor carrier for a skill like holding a confident call with a busy executive, which is the example Thomas uses.

What managers are stretched across

Coaching is the mechanism that would close that distance, and the numbers show why it rarely runs. 73% of sellers receive no consistent coaching at all. 76% say they need more of it. 55% of managers do not know how to coach. My own observation from client work is that most managers were promoted for their own sales results and never taught the skill, so a review can drift toward activity totals even when the time exists. Time is the other constraint: the coaching product page describes managers stretched across 8 to 12 reps while carrying their own number and their own deals.

Coaching frequency and the wrapper around it

The most quoted figure in this territory comes from the same product page. Sales organisations that use AI in their coaching activities report 3.3x greater year-over-year growth in quota attainment, a number the page attributes to ValueSelling and Aberdeen Strategy & Research. The comparison is between organisations that already coach and organisations that do not. It does not say that buying a coaching tool produces the growth. It says that teams which put AI inside an existing coaching routine grew faster than teams without that routine, which is a claim about a wrapper, not about a product.

What the figure does and does not say

The mechanism inside a good coaching loop is unglamorous. A manager or a coach looks at the deal in front of the rep, points at the skill underneath it, and the rep applies the correction in the next call. Software can put the review on a calendar and remember what was said last time. It cannot make anyone change a habit without a person holding a standard. That is also why practice-only tools lose their users: the rehearsal is easy to skip.

A manager leans over a laptop beside a rep, pointing at one deal while the rep writes a note.

What enough looks like operationally

If the answer is not more tooling, the next question is what to build instead. The test I would apply to any follow-up setup is narrow: every enquiry, from every channel, ends up in one record with a named owner, a next step and a visible stage. A team that can answer that for all of its enquiries knows where its decision-maker problem sits. In the Singapore teams I see, the gap shows first in chat and phone enquiries, which live outside the CRM. That is also where web application development decides whether the fix holds: a workbench and a customer portal help only when they read and write to the same record the sales team works from. Visible fields change the first call too, because nobody spends the opening minutes working out who they are talking to.

Capture, identify, stage, escalate

The four steps are worth naming in order, because each one fails in its own way.

  1. Capture. Enquiries from forms, chat, calls and marketplace messages land in one list, not in an inbox, a shared spreadsheet and a rep's phone.
  2. Identify. The record holds a company, a contact, a source and, when offered, a budget and a timeline, so a rep can judge whether the person is senior enough to matter.
  3. Stage. The account carries a stage that moves on events and stamps each change with a time, which turns silence into something visible.
  4. Escalate. When the first contact is not the buyer, a person reads what was said and decides who to approach next.

What Topkee's systems do about it

Topkee's Custom AI Agent Development service is scoped around those four steps. Its stated principles are that data has a source, execution has permissions, results are verifiable and the client can take the delivery over, and outward sending or budget changes still need human confirmation. Access work starts at the front door: the YME Chat Agent answers visitors in their own language, reads the intent behind a question, captures the lead with a human handover point and syncs it to Leadbox. The workbench a rep uses to read that record is web application development, built on the same backend as the agent, so the portal, the chat and the sales list share one set of fields. Identity resolution through TTOclid keeps one customer file across devices and platforms, and Leadbox holds the fields that reveal authority, including company, contact, phone, email, source, message and budget. Sales stages move on events, routing sends an account to the right rep, and records carry calls, chats, emails and the next step, so an account stays visible while the first contact is only a user. Verification is agreed before development, on samples, scoring rules and pass thresholds, and the delivery is scoped without a promise of a conversion rate or a revenue-growth percentage. What it fixes is the record: who the person is, who else at the account can decide, and what happens next.

A budget test for next year

The useful way to read the study is as a sorting exercise for a budget, whether the line is another seat of ai automation for business or an afternoon of a manager's time. Each purchase below fails in a different place.

OptionWhat it fixesWhat it cannot fixWhat to ask before signing
More AI prospecting seatsResearch, drafting and personalisation speedWho receives the message and whether they decideThe share of sequences that reached a named decision-maker
Manager coaching timeSkill gaps on calls and on consistent outreachReach on its ownThe number of live deals reviewed, and the habit that changed
A follow-up system on one recordOwnership, stage visibility, the 27% who cannot stay consistentThe quality of the conversation itselfThe share of enquiries with a named owner and a dated next step

The first row is where most teams have already spent, which is why the study's adoption figures look healthy. The second and third rows are where the challenge ranking says the money is missing. A team that puts next year's increase into the first row gets faster at reaching the wrong person, and can measure how fast.

One caution on the table. Every option depends on whether the data underneath it is traceable, and that question is worth asking before the purchase. Enquiries from paid search are only as good as the record that follows each click, which is the point the note on Google Ads brand AI conversions makes when it traces a brand's paid traffic into tracked conversions.

What to watch next quarter

Three things are worth watching next quarter, and none of them is an AI release. The first is the share of enquiries that carry a named owner and a dated next step, which tells you whether the follow-up system works. The second is whether reps can name who signs before they send a sequence, a detail that shows up in call notes as a person rather than a job title. The third is how a manager spends review time: reading activity totals off a dashboard, or discussing the deal that stalled last month. The study's training numbers suggest the third is where the gap opens.

A manager and a rep review a laptop together in a meeting room at quarter end, a notebook open beside them.

FAQ

Why do sellers still miss the decision-maker when they have AI tools?

The tools work on the preparation layer. They read an account and draft a message, and neither task answers who has authority to approve. That answer comes from a conversation, a referral or a record that already knows the account, which is why the study's sellers rank access above tooling.

Is AI prospecting still worth the spend?

For research, drafting and personalisation, yes, on the evidence of how widely it is used and what it does well. It is worth less as a route to the buyer. A team buying it to solve an access problem should expect a faster version of the same problem unless the follow-up process changes too.

How much coaching is enough?

The study gives no frequency, and its finding is directional: coaching frequency predicts quota attainment more reliably than anything else measured, and 76% of sellers say they want more of it. A workable start is a short review on live deals at a rhythm a manager can keep.

Should I buy more tooling or hire a coach?

Look at where enquiries die. If they die because nobody follows up, fix the follow-up process first, since hiring does not solve a routing problem. If they die in conversations with the right person, coaching is the better buy, because the study puts the skill gap at the centre of quota performance.

What is the first thing to fix in the follow-up process?

Ownership. Every enquiry needs one named owner and a next step with a date on it, because that single field turns a quiet record into a visible one. Stage and routing can follow; without an owner, both are decoration.

Sources

  1. "ValueSelling: Sellers Struggle to Reach the Right Decision-Makers Despite AI Adoption", Demand Gen Report, 2026-11-17. https://www.demandgenreport.com/industry-news/news-brief/valueselling-sellers-struggle-to-reach-the-right-decision-makers-despite-ai-adoption/55348/
  2. "ValueCoach AI by ValueSelling | AI Sales Coaching Platform", ValueSelling Associates, undated product page. https://www.valueselling.com/solutions/revenue-tech/valuecoach-ai
Share to:
Date: 2026-11-17
Terry Wong

Article Author

Terry Wong

Website Production Manager