Superior Commercial Decisions Through Strategic AI Usage

AI only matters if it improves decisions. Across Life Sciences, the gap between experimentation and execution remains significant. AI and Commercial Excellence are firmly on the leadership agenda. Yet many organisations still struggle to translate ambition into measurable impact, particularly when it comes to CRM maturity, faster insights, and execution at scale.

For this reason, I am pleased to exchange perspectives with Matteo Bacchin Bertling, Founder of It’s AI Right, committed to the intersection of Commercial Excellence, AI, and practical implementation in the Life Sciences sector.

Karsten: Matteo, thank you for taking the time to share your insights today. AI is now on almost every leadership agenda in Life Sciences. Where do you see it genuinely accelerating commercial decisions, and where is it still adding more complexity than value?

Matteo: Thank you, Karsten. I am glad to contribute to this important discussion. AI genuinely accelerates decisions when it is embedded into existing commercial workflows and improves decision latency (how fast we go from signal → action). The strongest examples I see are: faster insight generation from messy data (CRM + engagement + sales + market signals), better targeting and prioritisation (who to focus on, why, and what to do next), and execution support (e.g., preparing field teams with sharper pre-call insights, consistent messaging, and clear next steps). These are “high frequency” decisions that happen every day, so even small improvements compound.

A very special use case and close to my heart is AI-driven market research by using “synthetic users”. You create AI personas (HCPs, Patients) and interview them to get insights on your desired topic. It’s faster, cheaper and provides comparable insights to real with accumulating research supporting this approach.

Where it adds complexity is when AI becomes a parallel universe: pilots that sit outside the CRM/operating rhythm, models chasing theoretical accuracy while adoption stays near zero, or heavy automation built on shaky data foundations. In those cases, AI doesn’t simplify work—it increases coordination cost, governance burden, and user scepticism. If AI doesn’t reduce friction for the field and the commercial managers, it’s usually not value yet.

Illustration: From AI Experimentation to Embedded Commercial Decisions

Karsten: Many organisations run AI pilots but struggle to turn them into sustained impact. From your experience, what typically breaks between proof of concept and day-to-day execution?

Matteo: Three things typically break:

1. Ownership and operating model: the pilot has a project team, but production needs a product owner, clear accountability, and a cadence (release cycles, feedback loops, governance). If “who runs this” is unclear, it dies (quietly and slowly).

2. Workflow integration: pilots often prove a model can predict something, but they don’t answer: Where does it show up in the user’s day? If the output isn’t delivered in the tools people already use (often CRM/Teams/email), it becomes “one more dashboard.”

3. Change management + trust: users need to understand what the AI is for, what it’s not for, and how to challenge it. Without training, incentives, and manager reinforcement, adoption collapses, even if the model is good.

Karsten: From a Commercial Excellence perspective, how should leaders prioritise AI use cases so that effort is focused on the few areas that truly move sales effectiveness and customer engagement?

Matteo: I prioritise around decisions, not technologies. Start with the 5–10 decisions that drive performance and consume real time (e.g., targeting, call planning, channel selection, next-best-action, resource allocation, performance steering). Then filter AI use cases through three lenses:

➤ Business leverage: Will it materially shift sales effectiveness, customer experience, or manager effectiveness? (Not “nice to automate,” but measurable lift.)

➤ Adoption likelihood: Will the field and managers actually use it weekly? If not, it’s a lab experiment.

➤ Data + process readiness: Do we have the minimum viable data and the operating cadence to sustain it?

This usually leads to a short list: 

(1) CRM and call effectiveness enablers, 

(2) targeting and planning, 

(3) performance steering and exception-based management—before anything exotic.

My advice is also: start small. Take the initiative(s) which are the easiest to implement, even though they have a lower impact to start. Building early wins helps building confidence to build something bigger.

Karsten: What role should data quality, CRM maturity, and operating model readiness play in deciding whether an organisation is actually AI ready?

Matteo: They’re not “nice-to-haves”. They are very important, but they also shouldn’t paralyse the company to get started. Companies often think they need “the perfect data”, but it will never be the case.

➤ Data quality: AI will amplify what’s there. If customer master, interactions, and activity tagging are inconsistent, the model output becomes noise with authority.

➤ CRM maturity: CRM is the behavioural backbone of commercial execution. If CRM is mostly compliance logging, AI on top won’t create insight-driven selling; it will generate suggestions nobody trusts.

➤ Operating model readiness: You need clear decision rights, governance (including medical/legal where relevant), and a mechanism to act on insights (who does what differently on Monday). Without that, AI produces recommendations that don’t translate into action.

In practice, “AI ready” often means: good-enough data + reliable cadence + accountable owners, not perfect data.

Karsten: You often emphasise implementation over experimentation. What distinguishes teams that successfully embed analytics and AI into everyday commercial decisions from those that do not?

Matteo: Successful teams treat AI like a commercial product, not a data science deliverable.

They start small, ship something usable, and iterate weekly with users. They put AI outputs directly into the workflow (CRM tasks, manager huddles, call planning routines). They measure adoption and behaviour change—not just model metrics. And they invest in enablement: managers learn how to coach with AI signals, and reps learn how to use suggestions critically.

Unsuccessful teams optimise for the pilot demo: impressive dashboards, complex models, and unclear ownership. The gap is execution discipline, governance, and user-centric design.

Karsten: How do you see the role of Commercial Excellence evolving as AI increasingly supports CRM optimisation, planning, and performance steering?

Matteo: Commercial Excellence will shift from being primarily a reporting and process function to a decision-enablement and orchestration function.

As AI reduces manual analytics and “slide production,” Commercial Excellence can focus on: defining the right decision routines, standardising best practices across markets, building scalable playbooks, and owning the commercial “product stack” (CRM + insights + enablement + governance). It also elevates the importance of ComEx as the bridge between business, data/tech, and the field—ensuring AI is not only built, but used correctly and consistently.

In short: less time explaining what happened, more time shaping what happens next.

Karsten: Looking ahead, which practical AI or advanced analytics developments are most likely to reshape commercial operating models in pharma over the next 12–24 months?

Matteo: Three developments stand out as particularly impactful for reshaping commercial operating models in the near term: 

1. Copilots/agents inside CRM and daily tools (CRM, email, Teams): summarising interactions, proposing next steps, preparing pre-call briefs, and reducing admin burden while staying compliant.

2. AI-powered segmentation and targeting: platforms can help Life Sciences companies refine launch planning and ensure that no relevant HCP segments are overlooked.

3. Self-service analytics: With tools like Claude in Excel, ad hoc dashboards can now be built in minutes by clearly describing the required output.

There is, however, a critical caveat.  It’s not enough to just give access to a tool: companies need to put AI training and coaching at the centre, otherwise they are just expensive pilots.

Karsten: Thank you, Matteo, for sharing such grounded and implementation-focused insights. What makes your perspective particularly valuable is the clear emphasis on decisions, workflows, and adoption rather than on technology alone.

The discussion shifts the AI narrative from ambition to execution and clarifies what truly drives impact in Commercial Excellence: embedding AI into daily routines, ensuring clear ownership, focusing on high-frequency commercial decisions, and building adoption through trust and enablement. The message is simple but powerful. AI creates value when it reduces decision latency and friction for the field, not when it adds another layer of complexity.

In short, the future belongs to organisations that treat AI as a commercial product embedded in their operating model, not as an isolated pilot.

If you look at your current AI initiatives, which ones are truly embedded in daily commercial decisions, and which are still operating as pilots?

#CommercialExcellence #ArtificialIntelligence #LifeSciences #SalesEffectiveness