# AEO for B2B SaaS: get named when buyers ask AI for software

> In short: Software shortlists in AI answers come from roundups, review sites and vendors' own pages, and when an assistant looks a product up by name it reads the vendor's docs and pricing. So make your pricing, product and help pages state plainly who you're for and what you cost, publish honest comparison and alternatives pages, keep review-site profiles current, and track the shortlist and alternative questions buyers ask.

Guide from Keats, published 2026-11-13, updated 2026-11-13. Canonical: https://getkeats.com/guides/aeo-for-b2b-saas

A software buyer's first shortlist is increasingly written by an assistant. "What's a good project management tool for a 20-person agency that needs client portals?" gets a list of four products and a sentence on each, and the buyer books demos from it. If you're not on it, you're not in the deal, and you won't see it happen.

## What software answers draw on

Software questions are among the most heavily searched by assistants, because products, features and prices change. Look at a sample of eight ChatGPT answers to questions from buyers of AEO software ([the CSV](https://getkeats.com/data/chatgpt-sample-2026-11-12.csv)). The pattern was clear:

- **Shopping the category**, it cited roundups and review-style articles, and named the products they listed.
- **Looking up a product by name**, it cited the vendor's own site and, often, its help centre.
- **On a niche question** no roundup covered, it named whichever vendors had a page about it, cited only from their own sites.

That gives a B2B SaaS company three surfaces: its own site and docs, the roundups and review sites in its category, and the niche questions where it can be the only good page.

## Your docs are marketing now

When an assistant checks whether a product does something, it reads the help centre. A support article titled "Setting up client portals" that opens with what the feature does, who it's for and which plans include it is, in effect, a sales page for the question "which project management tool has client portals?".

- **Make help articles answer first**: what the feature does, in one or two sentences, before the steps.
- **Say which plans include it**, plainly.
- **Keep them public and crawlable.** Docs behind a login can't be read.
- **Retire old articles** about features you've removed, or they'll be repeated.

## Pricing pages

"How much does X cost?" and "cheaper alternatives to Y" are among the commonest software questions. A pricing page that says "contact sales" for everything leaves the assistant to guess or to quote a third-party page from two years ago. Even if you sell enterprise contracts, state a starting price or the basis of pricing (per seat, per usage) in text, and say what each plan includes.

## Comparison and alternatives pages

Buyers ask "X vs Y" and "alternatives to X". Rivals' comparison pages get cited for exactly these questions. Write your own, honestly:

- **Name the rival and cite your facts about them**, with dates.
- **Say where the rival is better.** It's what makes the page credible to a buyer and quotable to an assistant.
- **Be specific about who should choose you**: "for agencies that need client portals", not "for teams that want more".

## Review sites and roundups

Software review sites appear in most category answers. Keep your profiles complete: category, pricing, integrations, screenshots, and a steady flow of reviews from real customers. For roundups, find the ones assistants cite on your buying questions and work your way on, with a fact sheet and a trial for whoever writes it; the [roundups and threads guide](https://getkeats.com/guides/third-party-placements) has the outreach steps.

## The niche questions are yours to take

Every product has questions few others answer: an integration, a compliance need, a workflow for a specific industry. Our sample showed what happens: with no roundup to read, the assistant named whoever had a page. Find those questions in your sales calls and support tickets, and write one answer-first page for each. The page structure is in [writing for AI search](https://getkeats.com/guides/answer-first-content).

## Which questions to track

| Tier | Software examples |
|---|---|
| 1, buying | "best CRM for a 10-person sales team", "alternatives to [rival] that are cheaper", "[category] tool with [integration]" |
| 2, preparing | "how to choose a CRM", "what should I look for in a project management tool" |
| 3, facts | "does [your product] integrate with Slack?", "how much does [your product] cost?" |

Weight the buying tier heavily; see [question tiers](https://getkeats.com/glossary/question-tiers). Include the "alternatives to [rival]" questions for each main rival: they're where you meet buyers already in a sales cycle.

## Check the plumbing

Software sites often block bots at the CDN, render pricing and docs in JavaScript, or host docs on a separate subdomain with its own robots.txt. OpenAI says sites opted out of OAI-SearchBot "will not be shown in ChatGPT search answers"; check the docs subdomain too.

## When you're losing to one rival

If one competitor keeps being named instead of you, diagnose it question by question: which searches the assistant ran, which pages it read, what those pages say about them that yours don't. Our guide to [why ChatGPT recommends your competitor](https://getkeats.com/guides/why-chatgpt-recommends-your-competitor) walks through it.


## Your agent does this

Ask your agent: "Which of our buying questions name rivals from their help docs or pricing pages, and what do those pages say that ours don't?" It reads `list_citations` and `explain_prompt`, and drafts updates to the matching pages of your docs and site.

## Sources

- [Keats: eight ChatGPT answers to buyers' questions about AEO tools (CSV)](https://getkeats.com/data/chatgpt-sample-2026-11-12.csv), read 2026-11-13
- [OpenAI: Overview of OpenAI crawlers (OAI-SearchBot)](https://developers.openai.com/api/docs/bots), read 2026-11-13
