Attach your own presentation template to Airspeed Chat, say who the upcoming meeting is with, and the agent returns a deck built for that specific call. It cuts the slides that do not apply, writes the rest from what was actually said on your calls, and hands it back ready to download.
Why a generic deck no longer survives the meeting
Expectations moved. It is not acceptable any more for an AE to walk into a discovery call or a demo with an out-of-the-box presentation and hope it lands. Buyers can tell within two slides whether the deck was built for them or for anybody, and the second answer costs credibility you then spend the rest of the call earning back.
The problem was never that reps disagree with this. It is that doing it properly is expensive. Customizing a deck for one meeting means rereading the calls, remembering which stakeholder raised which requirement, pulling the relevant proof, and cutting the eleven slides that do not apply. That is an afternoon, and an AE with six live deals does not have six afternoons a week.
So the deck ships generic, or it ships late. Both are worse than the alternative.
What the flow actually looks like
Three steps, and none of them involve opening a blank slide.
- Write the prompt. Something close to plain speech: you have a template for a presentation, the agent should use it as guidance, and there is a call coming up with the CFO. Ask it to generate the deck.
- Attach the template. Drop your existing deck into the conversation as an attachment, the same way you would hand an agent any other file.
- Click Ask. The agent works through the context and comes back with the deck, plus a short written description of what it did and why.
Then download it and open it. The useful thing to do at that point is put the template and the generated version side by side, which is exactly what the video does, because the differences are the whole product.
What changes between the template and the deck
Four things, and they are the four a rep would do by hand given the time.
Slides get cut. The template carries everything your team might ever need. The generated deck carries what this meeting needs. Removing the irrelevant nine tenths is most of the work of customizing a deck, and it is the part people skip first.
The agenda comes from the calls. What the deck says it will cover is drawn from what was actually raised in conversation with this buyer, in their words rather than your category’s words.
Account-specific detail gets written in. Not a merge field with a company name in it, the actual substance that came up.
The feature list gets distilled. You offer a lot. This buyer cares about a few things. The agent narrows to those rather than parading the rest past a CFO who did not ask.
Where does the content come from?
Three sources, and the third is the one that matters most.
The agent reads the context across your deals, the email back and forth between seller and buyer, and, crucially, similar deals your team has already won and lost. That last one is institutional memory nobody has time to consult. Knowing that the last three deals with this profile turned on the same procurement objection is the kind of thing that changes a deck, and it is sitting in your own history where no rep is going to go digging for it at 5pm.
This is the argument for the deck being built where the conversations already live. A general-purpose model can format slides beautifully, but it only knows what you paste into it. Assembling the source material is the expensive part, and that is the part being removed here.
Who decides how it sounds?
You do, at an admin level. The prompt behind the deck can be controlled centrally, which means the voice, tone and structure your team has agreed on get applied every time rather than re-decided by whoever is building a deck on a Thursday afternoon.
That is a quieter benefit than the time saving, and probably a larger one. Deck drift is real: six months after enablement ships a template, there are as many versions of the pitch as there are reps, each one a reasonable local edit that nobody reviewed. Centralizing the instruction means a new rep’s third deck reads like your best rep’s, and the experience a prospect gets is consistent across whoever happens to own the account.
Stop triggering it by hand
Even asking is a task, and tasks that depend on someone remembering them decay. So the last step in the video is to save the conversation as an agent.
Saving keeps all the context you built up in the chat and lets you apply it in set circumstances. The obvious one is after a discovery call: the call ends, the agent generates the deck, and it is waiting in Airspeed for the rep to review and adjust. Nobody requested it. The prompt stays editable, so you tune it to exactly how your team sells.
Agents can be created and edited in plain English, and deck generation shipped alongside charts and code execution on September 3, 2026. See agents for how the rest of them work.
What this does not do
It does not make your template better. If the underlying deck is weak, the customized version is a well-targeted weak deck. It does not know your strategy either, because it was not in the room when you decided how to play the account.
What it removes is the reconstruction: the forty minutes of scrolling back through calls, emails and old deals before a single relevant slide exists. You still edit it. You just start from a draft that is already right about the facts.
Want to watch it build one against your own pipeline? Book a demo, bring your deck template and a live deal, and we will generate the deck in the room so you can tell us what it got wrong.
Frequently asked questions
How do I get Airspeed to build a demo deck?
Attach your presentation template to a chat, write a prompt telling the agent to use it as guidance and who the upcoming meeting is with, then click Ask. The agent works through your deals, emails and past won and lost deals, and returns a customized deck you can download.
Does it use my template or its own layout?
Yours. You attach the template and the agent treats it as guidance, which is what keeps the output recognizable as the deck your team already uses rather than a generic one. It will cut slides from the template that are not relevant to the meeting.
What does the agent change between the template and the deck?
It removes slides that are not needed, writes the agenda around what was actually said on your calls, pulls in detail specific to that account, and narrows a full feature list down to the things that matter to this buyer.
Where does the content come from?
The context across your deals, the email back and forth with the buyer, and similar deals your team has won and lost in the past. That last source is the one a rep building a deck by hand almost never has time to go back through.
Can we keep reps on-message?
Yes. The prompt behind the deck can be controlled at an admin level, so the voice, tone and structure your team has agreed on are applied every time rather than left to whoever is building the deck that afternoon.
Do I have to ask for the deck every time?
No. Save the conversation as an agent and it keeps the context you built up. From there it can run in set circumstances, for example after every discovery call, so the deck is waiting in Airspeed for the rep to review rather than waiting to be requested.
