commercial leadership
WHAT DOES AN AI-AUGMENTED SALES DEPARTMENT LOOK LIKE?
A seller's week is mostly research, preparation and administration. Not selling. The change that matters is that preparation, memory and follow-through stop being personal traits and become an operating layer every seller has. The judgment stays with the person. The rest becomes institutional.
Capacity goes to everything
except selling.
A mid-market commercial function is typically ten to sixty people: a sales director or CCO, country or regional managers, account executives, key account managers, sales development, sales operations, bid support and sometimes solution engineers. Marketing, customer success and pricing may report into the same leader or sit beside sales.
It drives the top line through acquisition, expansion, retention and price realisation. It also shapes contribution margin, working capital and delivery quality, because a poorly qualified deal can create revenue while destroying margin or consuming scarce delivery capacity.
Ten processes| 01 | Territory and account planning | 06 | Pipeline review and forecasting |
| 02 | Prospecting and outbound | 07 | Key account management and renewals |
| 03 | Meeting preparation and follow-up | 08 | Pricing decisions |
| 04 | Proposal and quote production | 09 | Handover to delivery |
| 05 | Tender responses | 10 | CRM hygiene |
Where the information lives is the problem. The CRM, mail, calendars, call recordings, chat, ERP order history, CPQ tools, service tickets, document stores, proposal folders and private spreadsheets. Relationship history and product judgment sit disproportionately in senior sellers' heads. CRM records carry stages and amounts without the evidence behind them.
Sellers spend roughly forty percent of their time selling. The rest goes to research, preparation and administration.
Salesforce's 2026 State of Sales survey puts selling time at 40 percent. LinkedIn's 2025 research describes roughly six hours a week of manual CRM work. Different samples, the same constraint: seller capacity is absorbed by everything that surrounds the sale.
Performance becomes
institutional.
Two things change at the core of the function.
Seller performance becomes partly institutional rather than biographical. Every seller can receive the preparation, retrieval and follow-up discipline that used to belong to the best people. The company extracts patterns from won bids, account histories, call transcripts and product expertise, and makes them available during live work. The result is a composite seller: a person exercising commercial judgment, with an operating layer that remembers, prepares and follows through consistently. The performance distribution narrows. Relationship judgment and negotiation skill still separate the best from the rest.
Commercial management changes from collecting reports to managing exceptions. Account intelligence refreshes continuously. Pipeline review starts with unexplained changes, missing evidence and deals that need intervention. Managers stop spending the meeting reconstructing what happened and spend it deciding what to do.
The same design improves speed and quality and reduces unit cost at once. Preparation covers more accounts, proposals arrive sooner, the record gets fuller, and senior time moves toward customers.
Where it is weakestThe combination is weakest where success depends on politics, trust or novel deal design. Those stay with human judgment, and should.
A day, a week,
a month.
A function that invests heavily still employs sellers and account managers in 2028. Its shape changes. Fewer roles concentrate on data entry, list production or report assembly. More capacity sits in strategic account management, solution design, commercial operations and knowledge stewardship. One commercial operations owner maintains data definitions, agent permissions and evaluation sets. A twenty-person team may remain a twenty-person team while handling more accounts and more bids.
The dayEvery seller receives a short account brief: recent correspondence, open commitments, delivery issues, changes at priority accounts, deals losing momentum. The seller chooses which recommendations deserve action.
Routine research has been assembled from approved company material and, where connected, the CRM, mail and public sources. The seller reviews the buying group, prior objections, installed products, open service matters and a suggested conversation plan. A consented transcription captures decisions and commitments.
A follow-up draft, a CRM update and a next-step plan appear for approval. The seller approves the record, and the record is right.
The weekThe pipeline agent compares CRM fields with meetings, mail, historical conversion and delivery constraints. The review opens with ten exceptions rather than every opportunity. Managers test the evidence, coach the seller and make resource decisions. Forecast changes carry reasons and source links.
The monthThe commercial leader reviews account movement, price realisation, forecast error and agent reversals. Their week centres on account allocation, deal strategy, pricing guardrails, coaching, and coordination with delivery and finance. CRM administration becomes a control activity rather than a management ritual.
Continuous attention to every material account, structured preparation for every meeting, disciplined follow-up across the whole portfolio. Those used to be reserved for the largest opportunities. They become the standard.
What runs, and what
stays with the person.
| Process | What the agent does | What stays with the person |
|---|---|---|
| Territory and account planning | Assembles account potential, relationship coverage, product fit and whitespace | Ownership, coverage intensity, the strategic bets |
| Prospecting and outbound | Researches approved segments, drafts account-specific approaches, routes promising replies | The market hypothesis, campaign approval, the real conversations |
| Meeting preparation and follow-up | Prepares the brief, captures consented notes, extracts commitments, drafts follow-up | Conversation strategy, approval of the record |
| Proposal and quote production | Creates a sourced draft from approved language, customer context and product rules | The commercial story; product, finance and legal approve their sections |
| Tender responses | Parses requirements, builds the compliance matrix, retrieves approved answers, flags missing evidence | Win themes, the submission |
| Pipeline review and forecasting | Detects stale deals, contradictory signals and unusual slippage | Customer politics, intervention |
| Key account management and renewals | Maintains the relationship map, commitments, renewal plan and whitespace hypotheses | Whom to involve, how to handle tension |
| Pricing decisions | Recommends a range from comparable transactions, customer context and guardrails | Exceptions explained by sellers; material deviations approved by managers |
| Handover to delivery | Assembles scope, commitments, exclusions and unresolved risks | Joint approval by sales and delivery |
| CRM hygiene | Proposes records from meetings and mail, checks required fields, explains conflicts | Approval; later, routine verified activities log automatically |
Four stages, and the
CRM position today.
A realistic path has four stages. They map onto the rungs in The Access Ladder, and the order is the same.
Meeting packs, proposal drafts and tender responses begin with curated product material, approved bids, price policies and playbooks. No live-system access. Outputs are drafts, and the seller supplies account context by hand.
The CRM, calendar and selected mailbox data come first. These remove repeated retrieval and support current account briefs. Meeting transcripts follow once notification, retention and purpose are settled.
The agent proposes notes, tasks, contacts and deal-field changes. The responsible seller sees the source and approves the write. Proposal and quote workflows then connect product, ERP, CPQ and contract data.
Activity logging, internal reminders and verified contact enrichment run within defined limits. Customer messages, opportunity overrides, contractual commitments and prices outside approved bands keep named human approval.
Point connections become inadequate when forecasting, pricing and account health depend on several years of CRM, ERP, service and contract history. A warehouse then gives every agent the same definition of revenue, activity and churn.
In the headless model the CRM keeps records, permissions and audit history. Sellers work mainly through meeting, mail and briefing surfaces and open the CRM for exceptions and investigation. Every action log identifies the agent, the authenticated user, the sources, the proposed change, the approval, the before-and-after values and the system's response.
Six things we believe,
from building this.
We have built this inside commercial organisations, not read about it. Six things hold every time, and they are the ones a leadership team most often gets the wrong way round.
Every rollout measures drafting time and adoption rates, and both are real. A commercial function can get faster at everything it already does and not close one more deal. The return sits in the quality of the account choice, the preparation and the pricing call, and nobody has built a dashboard for that yet.
The cost side is where the business case writes itself, which is why everyone starts there and why it plateaus. The market nobody had time to analyse properly, the pricing model nobody dared reopen, the hundred customer conversations nobody had the capacity to prepare: that is where a commercial function's impact per hour actually moves.
Prospecting is only as good as the market choice before it. Meeting preparation can only show history if the last report was written. A proposal cannot contain what has not been agreed. Tooling dropped into the middle of a chain with holes underperforms, and it is almost never the tool's fault.
Sales management runs on a pipeline nobody maintains, because maintaining it competes with selling. When notes, actions and fields are proposed from the work and merely approved, the pipeline reflects reality for the first time, and the forecast becomes a reading rather than a ritual.
The pitch, the read of the room, the negotiation and the relationship stay with the person. Everything around them, the research, the assembly, the write-up, the reminder, is the operating layer. Get that boundary wrong in either direction and the sellers either stop trusting it or stop thinking.
A management team can only set ambition for what it has used itself. Hand the tools to the floor and steer from outside, and the ambition caps at a faster version of last year. Put the leaders through it first, on their own accounts, and they decide from experience.
None of the six is a technology position. They are positions about how a company sells, and every tool laid on top inherits them.
Roles, rhythm, and
where it fails.
Sellers move from information gathering toward account judgment, customer dialogue and negotiation. They learn to supply missing context, inspect sources, challenge weak recommendations and recognise which outputs need approval. The average seller receives more of the best seller's preparation. Judgment still develops through experience.
Sales development shifts from sequence volume toward segmentation, message design and reply interpretation. Cheap prospecting raises competitive noise, so track positive replies, complaints and opt-outs beside meetings booked. Sales operations becomes the steward of commercial definitions, permissions, evaluation sets and agent-written data. Conversation intelligence stays a coaching input, never a hidden surveillance system.
The rhythmA document-grounded workflow changes one seller's week
A stable team rhythm
Connected read-and-write workflows
Role, territory and incentive redesign
Where it failsThin CRM data. Conflicting definitions between countries. Generic drafts nobody trusts. Automatic writes that create false certainty. Review work left on top of the old process instead of replacing it. Incentives tied to usage rather than outcomes. And the commonest of all: a successful pilot that stops, because it looked like success for about a quarter.
Involve the people whose work changes before it changes, and involve their representatives where the company has them. That is not a legal note. It is the difference between a rollout and a way of working.
Twelve questions for the
commercial leader.
Each one names something the function could do, and asks what it would change. None has a right answer.