Most companies ask which roles AI can replace. The stronger question is which processes AI should carry, and whether the recovered time improves the customer experience the role was there to create.
AI should usually automate the repeatable non-selling work inside a commercial role before a company redesigns the role itself. The useful question is which processes AI can carry without weakening buyer experience, judgment, escalation, or accountability.
FCP maps sales, customer success, and commercial operations work by process type: repeatable, information-heavy, judgment-led, and accountability-led. That map shows where AI can safely increase capacity and where human ownership still matters.
The AI conversation in commercial teams has settled on the wrong unit of analysis.
Most companies are asking which roles AI can replace. Whether the SDR function can run without human staff. Whether customer success headcount can be reduced. Whether the next product launch needs the same team depth as the last one.
These are role-level questions. They are also the wrong place to start.
The unit that matters is not the role. It is the process. And the outcome that should drive the design is not cost reduction. It is customer experience.
Salesforce's 2026 State of Sales report found that sales reps spend 40% of their time on selling activity and 60% on non-selling work. The AI opportunity sits squarely in that 60%. Capturing it requires understanding what those processes are, which of them AI can carry, and what the human does with the time that recovers.
The mistake is treating a role as the unit of automation. The right unit is the process. The right test is whether automation improves the customer experience the process was there to create.
A role is a label. Inside that label is a collection of processes: some repeatable and pattern-driven, some requiring synthesis of multiple information sources, some requiring judgment, and some requiring human accountability.
When a company automates at the role level, it removes the person without mapping what the person was actually doing. The visible activity disappears. What does not disappear is everything embedded alongside it, including the customer experience the role was producing.
McDonald's ran a two-year pilot of AI-powered drive-through ordering across roughly 100 restaurants before ending the programme in mid-2024. On paper, order-taking is a repeatable process. The AI should have handled it.
What the analysis missed was what the process was there to deliver. A drive-through interaction is a customer experience touchpoint. The customer expects speed, accuracy, and a frictionless transaction. The trial produced repeated misheard orders, incorrect items, and friction at the moment the brand was most exposed to judgment. The lesson is not that voice AI has no future in drive-through. It is that the design question was wrong from the start.
They asked whether the order-taking process could be automated. The better question was whether automation would improve the customer experience that process was there to deliver.
Klarna's 2024 AI assistant announcement framed the tool as doing the equivalent work of 700 full-time agents. Later reporting described a more mixed result: the company still needed human agents for complex customer issues where judgment, escalation, and accountability mattered. The judgment embedded in the customer service role had not disappeared: knowing when a response was inaccurate, reading a frustrated customer correctly, making the call to escalate before a complaint became a loss.
In both cases, the role was automated. The processes inside it were not properly understood.
Before any commercial role is restructured around AI, two questions need answers.
First: what are the processes inside this role, and which of them should AI carry?
Second: what customer experience does the human in this role produce, and does the restructure improve it?
The second question is the one most companies skip. It is also the one that determines whether the transition creates commercial value or erodes it.
Four categories are useful.
Repeatable processes follow a clear pattern, do not require contextual judgment, and produce consistent outputs. They are candidates for full automation.
Information-heavy processes require synthesis of signals from multiple sources: market data, conversation history, CRM behaviour, and customer context. They are candidates for AI support, where the tool improves speed and coverage without replacing the decision.
Judgment processes depend on experience, pattern recognition, and contextual interpretation that is difficult to formalise. They require more careful assessment before any transition.
Accountability processes require someone to own the outcome, absorb the commercial risk, or make a decision with downstream consequences. These belong with a person until there is clear evidence the risk can be reliably managed otherwise.
The table below applies the framework to common non-selling processes in SDR and customer success roles.
| Process | Category | AI recommendation |
|---|---|---|
| Prospect research and account enrichment | Repeatable | Automate |
| CRM logging after calls and emails | Repeatable | Automate |
| Outreach sequence creation and management | Repeatable | Automate |
| Meeting scheduling and coordination | Repeatable | Automate |
| Post-call summaries and notes | Repeatable | Automate |
| Pre-call briefing compilation | Information-heavy | AI support |
| Account prioritisation based on engagement signals | Information-heavy | AI support |
| Pipeline reporting and forecast preparation | Information-heavy | AI support |
| QBR and review meeting preparation | Information-heavy | AI support |
| Usage pattern monitoring and alerts | Information-heavy | AI support |
| Objection detection and early naming | Judgment | Human-led |
| Account anomaly recognition | Judgment | Human-led |
| Buyer relationship development | Judgment | Human-led |
| Client escalation decisions | Accountability | Human-owned |
| Commercial negotiation and deal judgment | Accountability | Human-owned |
Most commercial roles contain processes across all four categories. The ratio varies by company, team, and growth stage. Drawing the map makes that ratio visible, often for the first time.
For the SDR, repeatable and information-heavy processes represent a significant share of the working week. When those move to AI, the SDR's time with buyers changes. More conversations per account means more context. More context means a more accurate read of where the buyer is. More time in conversation means the SDR can detect the objection before it becomes a pattern, notice the account not behaving as the segment model predicted, and surface the signal that would eventually appear in a board slide, but only because someone was close enough to catch it early.
That is not only an efficiency improvement. It is a customer experience improvement. It is also market intelligence that accumulates through the relationship and does not exist anywhere else.
For the customer success manager, the same logic applies at the retention and expansion level. AI carries the QBR prep, usage alerts, renewal administration, and routine reporting. The CS manager carries the client. More face time. More understanding of what is actually happening in the customer's business. More of the presence that keeps a client bought in when things get difficult and converts a retained account into an expanded one.
Cognism's 2026 research found that 93% of CROs and sales leaders are already embedding AI into prospect research and account prioritisation. The adoption question is settled. The question that remains is whether the time being recovered is being reinvested in customer contact or absorbed by other demands.
Use this as a starting point before any transition.
| Step | What to inspect |
|---|---|
| 1 | List every recurring task in the role. Not the job description. The actual daily and weekly work. |
| 2 | Categorise each task as repeatable, information-heavy, judgment-led, or accountability-led. |
| 3 | For repeatable and information-heavy tasks, identify which AI tools can carry them with the quality and reliability required. |
| 4 | For judgment and accountability tasks, confirm these remain with the human and that the restructure does not accidentally remove them. |
| 5 | Calculate the time recovered when the right tasks move to AI. |
| 6 | Define which buyer or customer interactions increase, and by how much. |
| 7 | Ask whether the restructure improves the customer experience or only reduces cost. |
| 8 | If the answer is only cost reduction, revisit the design. |
Should AI replace SDRs? The better starting point is not whether the SDR role can disappear. It is which parts of SDR work are repeatable, information-heavy, judgment-led, or accountability-led, and whether automation improves the buyer experience.
What is the difference between automating a role and automating a process? Automating a role starts with the job title and asks whether a person can be removed. Automating a process starts with the work itself and asks which tasks can move to AI without removing the judgment or customer experience the role was producing.
Which sales tasks are usually suitable for AI? Prospect research, CRM logging, account enrichment, meeting coordination, first-draft follow-up, pre-call briefing, and routine reporting are usually stronger candidates than qualification judgment, negotiation, escalation, or relationship development.
How can AI improve customer experience in sales? AI improves customer experience when it frees people from non-selling work and the recovered time is deliberately reinvested into buyer conversations, faster follow-up, better context, and more useful escalation.
What should stay human? Work that requires judgment, accountability, relationship context, commercial negotiation, or sensitive escalation should remain human-led. AI can support those decisions with better information, but it should not quietly inherit responsibility for them.
What should a company measure after automating non-selling work? Measure whether buyer contact increases, response quality improves, follow-up becomes faster, customer issues are escalated earlier, and sales or customer success teams have more useful context before conversations. If the only metric is headcount reduction, the design is incomplete.
A company that removes the visible activity without mapping what is embedded alongside it does not become more efficient. It becomes fragile.
The signals that were being detected get missed. The objections that were being named early go unnamed until they appear as lost deals. The customer relationships that were being built quietly disappear with the role that was building them.
Emergence Capital's 2025 survey of more than 560 venture-backed technology companies found that 36% had restructured their SDR or BDR teams in the past year, the highest rate of change of any sales function. The restructures that produce durable commercial outcomes will be those built from a process map and designed around a customer experience outcome. The ones that are not will require rebuilding.
The AI opportunity in commercial teams is real. The non-selling work, the 60% that currently sits between the salesperson and the buyer, is a legitimate target for automation and AI support.
But the transition works when it is designed around processes, not aimed at roles. The process map is the prior work. The customer experience question is the filter. Without both, the transition removes the right activity and the wrong judgment in the same move.
The companies that get this right will end up with commercial teams doing more of what they were hired to do: spending more time with customers, building the understanding and relationships that produce revenue that is harder to replicate.
If your business is deciding where AI belongs in sales, customer success, or commercial operations, FCP can help identify which work is ready for automation and which customer moments need human judgment.
Start a conversation → View diagnosticsFull Court Press is a Singapore revenue growth advisory for companies across Asia Pacific. FCP works across go-to-market strategy, enterprise sales systems, AI search visibility, agentic growth systems, commercial diagnostics, and the operating rhythm behind repeatable revenue.
Related pages: Revenue Growth Advisory Services, Growth Intelligence Framework, Commercial Diagnostics, AI Search Visibility, and Agentic Growth Systems.
Salesforce: 15 Sales Trends Shaping 2026, citing the seventh edition of the Salesforce State of Sales report.
Cognism: The Cold Calling Competitiveness Gap and How To Close It, including 2026 research on AI use in prospect research and account prioritisation.
CNBC: McDonald's to end AI drive-thru test with IBM, published June 17, 2024.
SaaStr: The Great SDR Downsizing, summarising Emergence Capital survey findings on SDR and BDR team changes in 2025.
Klarna: AI assistant handles two-thirds of customer service chats in its first month, and AP: AI shakes up the call center industry, but some tasks are still better left to humans.
Plain answers on AI in sales, SDR automation, customer experience, and where human judgment still belongs.
The stronger starting point is not whether AI should replace SDRs. It is which processes inside the role are repeatable, information-heavy, judgment-led, or accountability-led, and whether automation improves the customer experience.
Repeatable and information-heavy work is usually the best candidate: CRM logging, meeting coordination, first-draft follow-up, prospect research, account enrichment, and pre-call briefing. Judgment and accountability work should remain human-led.
Companies should map the processes inside the role, identify which customer moments each process affects, then test whether automation improves speed, accuracy, trust, escalation quality, and relationship depth.
Automating a role starts with the job title and asks whether a person can be removed. Automating a process starts with the work itself and asks which tasks can move to AI without removing the judgment or customer experience the role was producing.
AI improves customer experience when it removes administrative and research work that keeps salespeople away from buyers. The recovered time should be reinvested into better conversations, faster follow-up, clearer context, and stronger escalation.
Work that requires judgment, relationship context, escalation, negotiation, or accountability should stay human-led. AI can support the decision with better information, but the commercial owner should remain responsible for the outcome.