In this episode of the Thoughts on Selling podcast, Torquil Thomson, joining me from Barcelona, explore something most revenue leaders know but rarely address head-on: your go-to-market engine is only as good as the data, systems, and human judgment beneath it.

Torquil brings 20+ years of experience across SaaS, sales ops, and country leadership, and today helps companies fix the messy operational foundations that stall growth — especially in the age of AI.

We start with his background, his move from Scotland to Spain, and how he kept gravitating toward building systems that help people perform, no matter the role he was in. From there, we dig into what The Conversion Architects do — and why so many companies think they have a “pipeline problem” when they really have a data quality and process problem.

Torquil shares astonishing numbers:

• Many companies have 40–50% duplicate records.

• Another 5–10% of contacts have no name.

• In one case, a customer had 2 million accounts… and only 1 million contacts.

As Torquil puts it, “You can’t build AI on chaos.”
AI isn’t magic — it’s a learning model. If your CRM is full of mud, your revenue engine runs on mud.

🔍 The Hard Truth About CRM Data

Torquil shares astonishing numbers:

• Many companies have 40–50% duplicate records.

• Another 5–10% of contacts have no name.

• In one case, a customer had 2 million accounts… and only 1 million contacts.

As Torquil puts it, “You can’t build AI on chaos.”

AI isn’t magic — it’s a learning model. If your CRM is full of mud, your revenue engine runs on mud.

🤖 AI, GTM Systems & the Illusion of the “Magic Wand”

We break down why so many organizations bolt AI tools onto broken systems and expect miracles. Torquil explains how:

• AI fails when the underlying data is incoherent.

• SDR/BDR tech stacks have become increasingly fragmented.

• Go-to-market engineers often build “Frankenstein systems” companies can’t maintain.

The Conversion Architects focuses on cleaning, deduping, enriching, and restructuring CRM data using a bow-tie model — because revenue systems must work underneath the automation, not just on top of it.

🔄 Humans > Machines (at Least for Now)

One theme that comes up again and again:

Sales is — and will remain — human.Automation should eliminate admin work, not connection, nuance, or judgment.

Torquil talks about using AI to help reps:

• Manage more opportunities without burning out

• Keep customer handoffs clean

• Identify when key roles change within an account

• Flag renewal risks before it’s too late

But not to replace the human element that actually moves deals.

🎯 Where Companies Gain the Most Leverage

Torquil sees opportunity across the full bowtie, especially here:

• Fixing handoffs between Sales → CS

• Aligning systems that don’t speak to each other

• Renewals and expansions (where companies often lose track of decision-makers)

• Automating the long-tail of high-propensity customers that reps never have time for

He shares a success story involving product-triggered outreach that drove huge lift — from understanding behavior, rather than blasting cold lists.

🧠 The Human Insights Sales Leaders Keep Missing

We talk about:

• Why sales engineers and first-line managers remain the highest-leverage roles

• Why reps struggle to talk about change management

• How AI allows buyers to become dramatically more informed

• Why customers don’t want more information — they want confidence

This leads to a discussion of The JOLT Effect (Matt Dixon & Ted McKenna), and how true sales impact comes from helping customers feel safer moving forward, not drowning them in detail.

⭐ Key Takeaways

• You can’t automate your way out of bad data.

• AI amplifies whatever foundation you have — good or bad.

• Sales is still human, especially when change and risk are involved.

• Fixing the middle of the bowtie (handoffs, renewals, role changes) unlocks huge value.

• Sales engineers and CS handoffs remain massively underutilized assets.

• Confidence beats information — especially in complex sales.