Enterprise Sales Deck for AI Startups

An enterprise sales deck is the document that does most of your selling when you are not in the room. A champion at the customer forwards it to their boss, their IT lead and procurement. Each of them opens it alone, on a laptop between meetings, and decides in a few minutes whether your startup is worth more of their time.

If you just raised, have three to six people and no designer, your sales deck is probably your pitch deck with the market slide deleted. But investors and enterprise buyers ask different questions, and a deck built for one rarely convinces the other. This guide is how I would build a sales deck for an AI startup selling to enterprise.

I was founding designer at Almetra AI from the first prototype through a $19M Series A. Almetra sells AI software to manufacturing companies, and I spent a lot of time with the people on the other side of the table: operators, line leads and the managers who sign off. What they need to see before they say yes is what this post is about.

The short answer

A good enterprise sales deck answers five questions, in this order:

1. Do you understand my problem? In my words, with my context.
2. What changes for my team? The workflow before and after, shown with real screens.
3. Can I trust it? How the AI reaches its output, and how people stay in control.
4. Is it safe to adopt? Security, integration, rollout effort, and who else uses it.
5. What do I do next? A concrete pilot with scope, success criteria and a timeline.

Everything else is appendix.

Sales deck vs pitch deck

I covered the investor side in pitch deck design after seed for AI startups and the slide order in pitch deck structure for AI startups. The two decks should tell the same story, but they do different jobs:

	Pitch deck	Enterprise sales deck
Reader	Investor	Champion, their boss, IT, procurement
Core question	Why does this become a big company?	Why should my team act now?
Opens with	Your claim and the market	The buyer’s problem
Proof	Growth, traction, team	Workflow change, results at similar companies, safety
Ask	The round	A pilot or contract

The mistake I see most is mixing them. TAM slides, fundraising history and “the future of AI” slides inside a sales deck tell the buyer you are talking to someone else.

Start with the buyer’s problem, not your company

Most startup sales decks open with “About us”, then the founders, then the technology. By slide five the buyer still does not know why they are reading.

Open with their problem instead. Use the language you heard in discovery calls. When I interviewed manufacturing teams for Almetra AI, nobody said “we lack visibility into production data”. They said things like “I find out about the problem at the end of the shift, when it is too late”. The second sentence lands. The first one sounds like every other vendor.

A practical way to get there:

• Collect the exact phrases buyers use in calls. Your notes or call recordings are the source.
• Pick the one problem that most of your best customers share.
• Write the problem slide headline as that sentence, close to how they said it.
• Add the cost of the problem in their terms: time, scrap, delays, risk. Only use numbers you can defend.

Write every headline as a claim

This rule matters even more in a sales deck than in a pitch deck, because the deck gets read without you. A slide titled “Platform overview” says nothing. A slide titled “Line leads see the cause of a stoppage while the line is still running” tells the reader exactly what to believe, and the slide body proves it.

Test it: read only the headlines, top to bottom. If that reads like a short, convincing argument, the deck works. If it reads like a table of contents, rewrite the headlines before touching any visuals.

Show the real workflow, before and after

Enterprise buyers buy a change in how their team works. So show that change.

• Before: how the job is done today. Spreadsheets, radio calls, a weekly report, whatever it really is. One slide, simple, no mockery of their current tools.
• After: the same job with your product. Two or three real screens, in the order a user would see them.

Use real screens with realistic data, not empty states or lorem ipsum. Crop to the part that matters, and annotate the one thing the reader should notice. Screenshots from three different product versions make the product look unstable, so update them together.

Be honest about what is live. If a screen is designed but not shipped, label it. Buyers forgive a roadmap. They do not forgive discovering in the pilot that the hero screen from the deck does not exist. I wrote about the same rule for live demos in how to design a sales demo for enterprise SaaS: separate shipped, designed and directional.

Make the AI legible

AI products have an extra burden in a sales deck. The buyer has to understand what the model does without a data science background, and the person who signs has to feel the risk is controlled.

What works for me is a simple pattern on the product slides:

• Claim: what the system tells the user, in plain language.
• Evidence: why it says that. The data, the signal, the comparison.
• Action: what the user can do next, including overriding or correcting it.

That pattern does more for trust than any slide about model architecture. It also matches what the product should do anyway, which I go deeper into in enterprise-ready UI for AI startups and SaaS dashboard design for enterprise buyers.

Add one slide on control: who sees what, how humans review outputs, what happens when the model is unsure. Enterprise buyers do not expect AI to be perfect. They expect you to know where it is not.

Design for the forward, not only the meeting

Your champion is your real presenter. They will forward the deck to people you never meet. So build two versions from one master file:

• The presenting version: sparse, big headlines, real screens. You talk over it.
• The send-ahead version: the same slides with short captions under screens, a one-slide summary at the front, and contact details at the end.

The summary slide is the most underrated slide in enterprise sales. Three lines: the problem, what changes, the proposed next step. Many readers will see only that slide. Make it good enough that your champion can paste it into an email to their boss. For the single page that stays behind after the meeting, see sales one-pager for enterprise SaaS.

Keep the file in a tool your team can edit, like Pitch, PowerPoint, Keynote or Google Slides. A deck that only a designer can change will be out of date within a month, and outdated screens are worse than plain ones.

Answer the risk questions before they are asked

Every enterprise buyer has a list of silent objections: “you are too small”, “this will take my team months to set up”, “IT will block it”, “what happens if you disappear”. You cannot remove those worries with confidence on a call. You can address them with a calm, specific slide or two:

• Rollout: what week one, month one and month three look like, and how much time their team needs to give.
• Integration: what systems you connect to, and what you need from their IT team.
• Security and data: the short version, with a pointer to the full documentation.
• Who else uses it: only customers and logos you have permission to show. If you do not have permission yet, describe the type of company and the result instead.

Put the detail in the appendix, keep the main slides short. The goal is that the IT lead who opens the deck finds their answer in under a minute.

End on a concrete pilot

The last slide should not be “Thank you” or “Questions?”. It should be the next step, designed as carefully as the product slides:

• Scope: which site, line, team or use case.
• Success criteria: what the buyer and you agree counts as success, written in their metrics.
• Timeline: start date, check-ins, decision date.
• People: who needs to be involved on their side and on yours.

A clear pilot slide makes it easy for your champion to say yes internally, because they can show their boss exactly what they are agreeing to. If your product side of the pilot is still weak, that is a design problem worth solving before the deck. The admin views, onboarding and first-value moment decide whether a pilot turns into a contract.

A structure that works

There is no perfect template, but this order fits most enterprise AI sales decks:

1. Summary. Problem, change, next step in three lines.
2. The problem in the buyer’s words. One per segment.
3. What it costs them today. In their terms, defensible numbers only.
4. Before: how the job is done now.
5. After: the workflow with your product. Two or three real screens.
6. How the AI reaches its output. Claim, evidence, action.
7. How people stay in control. Review, override, permissions.
8. Results at similar companies. Only what you can share.
9. Rollout and integration. What it takes from their team.
10. Security and data. Short, with a pointer to the full docs.
11. The pilot. Scope, success criteria, timeline, people.
12. Appendix. Security detail, integrations, team, FAQ.

Keep it tight. If a slide does not help the buyer decide, it belongs in the appendix.

Common mistakes I see

• The pitch deck in disguise. Market size and fundraising slides in front of a buyer.
• Feature lists instead of workflows. Twelve icons with one word each tell the buyer nothing about their day.
• Stock AI imagery. Glowing brains and neural network wallpapers signal “early” to enterprise teams.

One story across deck, website and demo

Your buyer will check your website after the meeting and probably watch the demo again. If those three tell different stories, trust drops quietly and you never hear why.

The Design OS method I use with Grauberg clients treats the sales deck, pitch deck, website and product as one system: the same problem framing, the same language, the same screens, adapted to each audience. When the positioning changes, everything changes together. For the website side, see startup website design for AI and B2B teams.

When to bring in a designer

You can get far yourself with a strict template, claim headlines and real product screens. Bring in a designer when:

• The story is clear, but the deck still does not look like a company an enterprise trusts.
• The product screens are the weak part, and the deck exposes it.
• The deck, website and demo have drifted apart and nobody owns the system.

For a single deck, a scoped project is usually enough. If the product, site and sales material all need the same care, a founding designer on retainer keeps them in sync as your proof grows. I wrote about that choice in founding designer vs agency.

Bottom line

An enterprise sales deck is not a smaller pitch deck. It is a document your champion uses to sell you internally. Start with their problem in their words, show the real workflow before and after, make the AI legible and controllable, answer the risk questions calmly, and end on a pilot they can say yes to. Get that right and the design becomes simple, because every slide finally has one job.