Battery Point Research / AI Playbooks
Sales & Marketing
A customer asks for a quote. What happens next?
With Outlook and Salesforce connected, an agent can identify the account, assemble its history, draft a reply and record the next step. The salesperson approves pricing and handles the conversation.
Illustrative workflow; requires connected systems and agreed permissions.
What the agent does — and what your team does
| Commercial activity | Traditional model | AI-first model | People impact |
|---|---|---|---|
| Account research | SDR or rep researches company, contacts, news and website. | AI continuously researches accounts and prepares briefs.The agent should attach its sources and flag uncertain company or contact matches. | Manual research role largely removed. |
| Contact data | SDR or marketing ops finds contacts and cleans records. | Existing APIs + AI continuously enrich and reconcile data. | Routine data work largely removed. |
| Prospecting | SDR researches, writes outreach, sends sequences and follows up. | AI targets accounts, personalizes outreach, monitors replies and can schedule meetings. | SDR requirement can fall materially. |
| Inbound leads | A person spots, researches, qualifies and routes the lead. | AI enriches, qualifies and routes immediately. | Several human handoffs can disappear. |
| Meeting preparation | Rep checks CRM, email, website, LinkedIn and old notes. | AI assembles account history, activity and suggested questions. | Most preparation labor removed. |
| CRM administration | Reps or sales ops enter notes, activities and tasks. | AI captures activity and updates records automatically. | Significant administrative work disappears. |
| Follow-up | Rep writes email, logs activity and creates reminders. | AI handles routine follow-up and monitors response. | Human can leave the routine workflow. |
| Proposals / RFPs | AE and support staff assemble responses. | AI drafts from approved company information and escalates exceptions. | Routine proposal-support requirement falls. |
| Marketing execution | Marketing staff or agency creates campaign variants. | AI creates and tests variants, then adjusts to performance. | More execution with fewer people. |
| Marketing measurement | People reconcile GA4, ad platforms, email and CRM. | AI connects activity across systems and reconstructs account journeys. | Manual reporting and attribution work falls sharply. |
| Pipeline | Sales ops reports; managers review opportunities periodically. | AI continuously monitors opportunities and surfaces exceptions. | Routine monitoring labor largely removed. |
| Management | Owner receives dashboards and periodic reports. | AI explains changes, their impact and decisions needed. | Less analyst and report-preparation labor. |
| Complex close | Senior salesperson handles negotiation and relationship. | Human seller uses AI-prepared history, analysis and scenarios. | Human remains central today. |
Research: McKinsey: agentic B2B sales workflows and seller productivity
Where does the human come out of the loop?
The key question is where the human still adds real value — and where the handoff can already disappear.
| Activity | Human today? | Could the human be removed now? | Likely next 2–4 quarters |
|---|---|---|---|
| Account research | Yes | Yes | Mostly automated |
| Contact enrichment | Yes | Yes | Automated |
| Lead qualification | Yes | Yes, in many cases | Largely autonomous |
| Routine outbound | Yes | Yes | Agent-run for much of top-of-funnel |
| Meeting scheduling | Yes | Yes | Autonomous |
| CRM updates | Yes | Yes | Background automation |
| Call summaries | Yes | Yes | Background automation |
| Routine follow-up | Yes | Yes | Increasingly autonomous |
| Proposal first draft | Yes | Yes | Automated from approved information and rules |
| Pipeline monitoring | Yes | Yes | Continuous AI monitoring |
| Campaign reporting | Yes | Yes | Continuous |
| Account prioritization | Yes | Yes | AI recommends or acts within rules |
| Pricing / discounts | Usually | Partially | More routine decisions delegated within limits |
| Complex negotiation | Yes | Not reliably yet | Human remains important near-term |
| Strategic account relationship | Yes | Not generally | Human remains important near-term |
This is a snapshot, not a stable boundary. Some human-in-the-loop steps may be transitional.
Human executes → AI assists → AI executes with human approval → AI operates within guardrails → human handles exceptions
In some workflows, the endpoint is not a more productive human. It is the removal of the human step altogether.
Research: Gartner: top-of-funnel roles · Gartner: autonomous sales workflows
Evolution of a Sales & Marketing Team: Traditional vs. AI-Enabled
The old model required adding people before the system could prove itself. AI lets smaller companies build more capability before adding headcount.
| Stage | Traditional team size | Traditional model / weakness | AI-enabled team size | What changes |
|---|---|---|---|---|
| Founder-led | 1–2 | Founder sells; marketing is ad hoc; follow-up and CRM discipline are inconsistent. | 1–2 | AI handles research, meeting prep, first drafts, follow-up, CRM updates and basic reporting. |
| First dedicated team | 3–5 | Separate sales, marketing and operations capability starts to become necessary. | 2–3 | One strong commercial operator can cover much of research, marketing operations, CRM administration and reporting. |
| Repeatable growth | 6–10 | More coverage usually means more SDRs, coordinators, agency support and sales operations. | 3–6 | SDR research, sales coordination, CRM entry and routine reporting can be largely automated or removed as standalone work. |
| Broader market coverage | 10–20+ | Headcount rises with accounts, campaigns, channels and management complexity. | 5–10 | AI monitors more accounts, runs routine workflows continuously and lets sellers cover larger books. |
| Scaled sales & marketing | 20+ | More scale historically required layers of RevOps, SDR management, analytics and administration. | Smaller relative to revenue | Growth comes increasingly from automation and system leverage before adding another layer of people. |
What the recorded work can reveal
Linking each request, action and outcome makes these questions easier to answer:
- Which inquiries wait longest for a useful reply?
- Which missing facts repeatedly delay quotes?
- Which follow-up steps lead to a meeting rather than another unanswered email?