Sales Plays
Sales Plays
A sales play is a document in your knowledge base that gives your agents specific guidance for a particular buyer type or scenario. Without plays, agents give generic answers. With them, agents know exactly how to talk to a solo founder vs an enterprise buyer, what pricing to recommend, how to handle objections, and what next step to offer.
Why Plays Matter
Your best sales rep knows things that are not written down anywhere: which plan to recommend for a 5-person team, how to respond when someone says "we tried AI before and it didn't work", when to push for a demo vs offer a free trial.
Sales plays capture that knowledge in a format your agents can use. Instead of giving every visitor the same generic pitch, your agents tailor the conversation based on who they are talking to.
The difference is significant. An agent without plays says "we have three plans starting at $49/month." An agent with plays says "for a team your size, most customers start with our Business plan at $149/month - it includes the reporting features you just asked about, and you can add users as you grow."
Creating Plays
- Go to AI > Knowledge
- Create a new knowledge base (or open an existing one)
- Click Add Document
- Write your play in markdown format
- Save and attach the knowledge base to your agent workflows
You can have multiple plays in a single knowledge base. One play per buyer segment or scenario works well.
What to Include in a Play
A good play answers five questions for the agent:
- Who is this person? - The buyer persona and how to recognize them
- What do they care about? - Their pain points and priorities
- What should I recommend? - The right product, plan, or next step
- What if they push back? - Objection handling for common concerns
- What should I avoid? - Things that will hurt the conversation
Target Audience
Describe who this play is for in concrete terms. Not "small businesses" but "solo founders or teams under 5 people, usually bootstrapped, price-sensitive, looking for quick setup."
The more specific you are, the better the agent can match visitors to the right play.
Pain Points
List the problems this buyer type typically has. Be specific about the frustration, not just the category.
Instead of "needs better CRM", write "spending 2+ hours a day updating CRM records manually, data is always out of date, reps hate it."
How Your Product Solves It
Map each pain point to a specific capability. Agents use these mappings to connect the visitor's problem to your solution in a natural way.
Pricing Guidance
Tell the agent which plan to recommend for this segment and why. Include the reasoning so the agent can explain it naturally in conversation.
Objection Handling
List the most common objections from this buyer type and how to respond. Be specific. Write the actual response you would want your best rep to give.
Next Steps
Define what the agent should push toward. A free trial? A demo booking? A proposal? Different buyer types have different natural next steps.
Rules
Explicitly state what the agent should not say or do with this buyer type. "Do not mention our enterprise plan to this segment - it is out of their budget and makes us look expensive." "Do not offer a discount unless they specifically ask about pricing flexibility."
Play Template
Use this structure as a starting point for each play. Copy it and fill in the details for your buyer segments.
# Play: [Segment Name]
## Who is this for
[Describe the buyer persona - role, company size, industry, budget,
what signals identify them in conversation]
## Their pain points
- [Specific pain point 1 - describe the frustration, not just the category]
- [Specific pain point 2]
- [Specific pain point 3]
## How we solve it
- [Pain point 1] -> [Your specific feature/capability and why it helps]
- [Pain point 2] -> [Your specific feature/capability and why it helps]
- [Pain point 3] -> [Your specific feature/capability and why it helps]
## Pricing guidance
- Recommend: [Plan name] at [price]
- Why: [Reasoning the agent can share naturally]
- Upgrade trigger: [When to suggest a higher plan]
## Objection handling
- "It's too expensive"
-> [Your response - focus on value, not discount]
- "We tried AI before and it didn't work"
-> [Your response - what is different this time]
- "We need to talk to our team first"
-> [Your response - offer to help build the case]
- "Can we get a discount?"
-> [Your response - what you can and cannot offer]
## Next steps
- Primary: [What to push toward - demo, trial, signup]
- Fallback: [If they are not ready - resource, follow-up email]
## Rules
- [What not to say or do with this buyer type]
- [Topics to avoid]
- [Promises not to make]
Example Play
Here is a concrete example to show what a finished play looks like:
# Play: Solo Founders and Small Teams
## Who is this for
Founders or small team leads (1-5 people). Usually bootstrapped
or seed-stage. They do everything themselves - sales, support,
product. Price-sensitive but willing to pay for tools that save
real time. Often come from a "spreadsheet and inbox" workflow.
## Their pain points
- Spending hours each week on manual data entry across tools
- Losing track of leads because follow-ups slip through the cracks
- Cannot afford to hire a dedicated sales or support person yet
- Tried cobbling together Zapier automations but they keep breaking
## How we solve it
- Manual data entry -> Automatic CRM updates from every conversation
- Lost follow-ups -> AI agents that respond instantly, 24/7
- No budget for headcount -> Agents cost a fraction of a hire
- Fragile automations -> Purpose-built workflows that just work
## Pricing guidance
- Recommend: Starter plan at $49/month
- Why: "It covers everything a small team needs - you get 2 agents,
your CRM connected, and chat on your website. Most founders
start here and upgrade when they hire their first sales rep."
- Upgrade trigger: If they mention hiring soon or needing more
than 2 agents, mention the Growth plan
## Objection handling
- "It's too expensive"
-> "Totally fair concern. Most founders on our Starter plan
tell us it replaces 5-10 hours of manual work per week. At
$49/month, that works out to less than $2/hour for the time
you get back."
- "We tried AI before and it didn't work"
-> "That's really common - most AI tools give generic responses
because they don't know your business. The difference here is
you give the agent your actual pricing, your FAQ, your sales
process. It responds based on YOUR knowledge, not general
training data."
- "I just need a chatbot, not all this"
-> "You can absolutely start with just the chat widget and one
agent. That takes about 10 minutes to set up. The rest is
there when you need it."
## Next steps
- Primary: Free trial signup (no credit card required)
- Fallback: Send them the getting started guide
## Rules
- Do not mention enterprise features (SSO, custom SLAs)
- Do not push for a demo - this segment prefers self-serve
- Do not offer discounts - the Starter plan is already positioned
as affordable
- Do not compare to specific competitors by name
How Plays Work with Corrections
Sales plays give your agents a starting point, but the real power comes from combining plays with human-in-the-loop corrections over time.
Here is how the cycle works:
- You write a play with your best knowledge of each buyer segment
- The agent uses the play to handle conversations
- You review responses with human-in-the-loop enabled
- You correct responses that are not quite right - maybe the tone is off, or the agent is recommending the wrong plan, or missing an objection
- The agent learns from those corrections and applies them to future conversations
Over time, your agents get better at each scenario. The plays provide the foundation, and the corrections refine the execution. After a few weeks of active correction, your agents start sounding like your best sales rep - because they have learned from your feedback on real conversations.
This is why it is worth enabling human-in-the-loop review early on, even if it feels slow. Every correction makes the agent better for every future conversation with that buyer type.
Tips for Getting Started
- Start with 2-3 plays. Cover your most common buyer types first. You can always add more later, but three good plays are better than ten incomplete ones.
- Use real conversations as source material. Look at your last 20 sales conversations. What questions came up? What objections? What worked? Use those patterns as the basis for your plays.
- Be specific about pricing. Vague pricing guidance leads to vague responses. Tell the agent exactly which plan to recommend and exactly what to say about cost.
- Update plays when things change. New pricing? New feature? New competitor objection? Update the play. Agents use whatever is in the knowledge base right now, so outdated plays mean outdated responses.
- Review and correct actively for the first two weeks. The combination of good plays and active corrections during the first two weeks will set the quality baseline for everything after that.
