// INSIGHTS

AI + Human Commercial Execution

Most commercial teams have now bought AI tools. Far fewer can point to a number that changed as a result. The gap is rarely the technology — it is the absence of a decision about where the machine helps and where a person stays responsible.

Leif Sundström··6 min read

The issue

The typical adoption pattern is familiar. A tool is bought, a few people try it, output quality drops somewhere visible, trust erodes, and within a quarter usage has quietly stopped. The tool is still being paid for.

This is usually blamed on the model, or on change management. More often it is that nobody decided in advance which specific steps AI would handle, which a person would always check, and who owned the setup once the enthusiasm faded.

Why it matters

Commercial teams have a genuine time problem. Research before a meeting, account intelligence, first drafts, keeping the CRM current — this work is real, it is repetitive, and it consumes hours that would be better spent in front of customers.

That is precisely where AI earns its place. The mistake is applying it to the part of the job that depends on judgement, relationships and reading a room, and then concluding it does not work when the output sounds generic.

The tools handle speed. The human still decides what ships. Reverse those and you get more output that nobody wants to receive.

The Sales Advantage view

AI belongs inside the existing workflow, at named steps, with an explicit line about what a person always reviews. That line is not a limitation — it is what makes the rest of it trustworthy enough to use.

  • Research and preparation: strong fit. Gathering, summarising and structuring information is exactly what these tools do well.
  • First drafts and personalisation at scale: good fit, provided the voice is tuned on material you have already written and sent.
  • Data hygiene and CRM upkeep: strong fit, and the least glamorous place with the highest return — notes, next steps and activity logged without typing.
  • Deciding what to say to a specific customer, and what a relationship needs: poor fit. That stays human.

There is also a question most teams meet too late. When you sell into enterprise, procurement will eventually ask which systems your AI touches and what data it processes. Deciding that at setup is considerably cheaper than discovering it during a security review on a live deal.

What to do

Map a week of your commercial team's actual work and mark the steps that are repetitive, information-heavy and low-judgement. That short list is your AI stack — usually far shorter than the number of tools already bought.

Then name one owner and write a one-page runbook covering what runs, what it costs, and what to check. Setups that depend on the person who built them stop working within a quarter of that person moving on, and that is the most common way this work is wasted.

// QUESTIONS

Common questions

Where does AI genuinely help a commercial team?
Research, account intelligence, meeting preparation, first drafts, personalisation at scale, and keeping CRM records current. These are information-heavy, repetitive and low-judgement — the conditions where the tools are reliable.
How do you stop AI-generated outreach sounding generic?
Tune templates on material your team has already written and sent, so the voice is yours rather than the model's default, and keep a person reviewing anything that reaches a customer. AI is used for speed and structure, not for deciding what to say.
Why do AI pilots fail after a few weeks?
Because no one decided in advance which steps the tools own, which a person always checks, and who maintains the setup. Without those three decisions, quality drifts, trust goes, and usage stops — while the subscription continues.

Want a second opinion on yours?

Book a working session