One Marketer, 8 Hours a Week, 13 Weeks: A B2B AI Search Plan Built Around Competitors

One marketer. Eight hours a week. Thirteen weeks. That is 104 hours, and I would spend the quarter putting competitors’ names on useful pages, not publishing another thought-leadership blog.

B2B buyers ask ChatGPT for shortlists, comparisons, pricing, and reasons to drop one vendor in favour of another. 94% of business buyers now use AI in the buying process. If your site answers those questions plainly, it has a shot at being cited and at earning the visit that follows. If it offers a thoughtful essay about the future of your industry instead, the buyer still needs an answer. They will get it elsewhere.

This is the plan I would give one marketer with no spare team behind them. The hours are fixed. So are the cuts.

The 104-hour budget: decide what does not get done

The quarter has five jobs, in this order:

1. Weeks 1–2: Find the buyer questions. Cost: 16 hours. Buys: a page list.
2. Weeks 3–8: Build comparison, alternatives, and pricing pages. Cost: 48 hours. Buys: answers a buyer can use and an AI answer can quote.
3. Weeks 9–10: Remove technical barriers. Cost: 16 hours. Buys: pages that search systems can read.
4. Weeks 11–12: Strengthen third-party profiles. Cost: 16 hours. Buys: evidence beyond your own site.
5. Week 13: Measure enquiries and citations. Cost: 8 hours. Buys: a decision about what to improve next.

That is eight hours in every week, not a 48-hour page sprint somehow squeezed into four Fridays. Nothing in this plan depends on a heroic weekend.

What gets cut? The thought-leadership calendar, broad guides with no buyer behind them, a blog rebrand, and the urge to make every page a 2,000-word essay. I used to tell clients to publish weekly to stay visible. I was wrong. If the choice is one comparison that answers a live purchasing question or 12 essays nobody uses to make a shortlist, I know where the hours go.

1. Weeks 1–2: Find the 15 shortlist questions

Cost: 16 hours. Buys: the page list. Do sales homework before you write a headline. Pull 10 lost deals, 10 won deals, and the last 20 demo-call transcripts if you have them. Mark every time a prospect names a competitor, asks what something costs, or wants to know who else they should consider. Then write the questions in the language buyers use, not the language on your positioning slide.

The list will usually contain variations on a few shapes:

• Best vendors for a specific use case.
• [Competitor A] vs [Competitor B].
• Alternatives to the market leader.
• The market leader’s pricing.
• Whether your product integrates with a named system.

Do not turn those examples into pages automatically. Rank the actual questions from your deals by how often they appear and how close they are to a decision. A comparison raised in lost deals is more useful here than a broad topic that happens to have a large search-volume estimate. If sales keeps hearing the same objection, write it down even if it makes your product look less tidy.

72% of B2B professionals use AI to scope a category, 62% to compare vendors, and 48% to narrow a shortlist. After an AI mention, 71% visit the vendor website and 63% Google its name. The answer and the follow-up visit both matter. Your inventory should cover the question that gets you named and the question a buyer asks when they arrive to check you.

I learned the value of that second question looking at groas’s Search Console. Our competitor review and alternatives pages draw demand while generic explainers lag. The pattern also appears in prompt-shaped searches: a competitor’s name plus review, pricing, or alternative. One of our target queries sits at position 12.3. I would rather improve the page answering that query than publish another general explanation and hope it wanders into a shortlist.

Finish the 16 hours with 15 questions, an order of priority, and a note beside each: what price, competitor, integration, or tradeoff must the page address? If you cannot name the decision the page helps someone make, it does not make the list.

2. Weeks 3–8: Publish pages a buyer can compare

Cost: 48 hours. Buys: the pages most likely to help with a shortlist. These are six ordinary eight-hour weeks. Ship three page types, in priority order, and stop when the hours run out. A complete, honest page is worth more than a folder of thin drafts.

Put your strongest rivals on “competitor vs us” pages

Start with the rivals named most often in those deals and calls. Publish one page per rival, up to four pages total. Name a competitor’s genuine strength before explaining where you win. Otherwise the comparison reads like an ad written by someone who assumes the buyer cannot open another tab.

Use facts the buyer can check: price band, implementation time, integrations, support hours, and the conditions attached to each. If a detail is not public or you cannot stand behind it, leave it out rather than filling the cell with confident fog. State who should choose the rival, then state who should choose you. That is not surrendering the comparison. It is giving the buyer a reason to trust the part where you win.

The mechanism is straightforward: a page that answers a specific comparison gives search systems and buyers something more useful to work with than a page saying your team is passionate about innovation. Comparison pages convert at 4–10% versus 0.1–0.5% for problem-aware blog content. I used to soften comparisons to avoid annoying competitors. That politeness did not help the buyer decide.

Cut the adjectives before you cut the tradeoffs. If the page cannot explain a reason to pick either product, it is not finished.

Give “alternatives to” pages a real shortlist

Next, build an Alternatives to [Market Leader] page with 5 to 10 real options, including you. Include the leader and two vendors you respect. Give each option a one-line “best for” and a reason someone might rule it out. A buyer should be able to leave with a shorter list, even if your product does not make theirs.

Put the comparison in a real HTML table. Use pricing, minimum contract, and one tradeoff column where those details are available. AI Overviews can pull from tables, and one cited analysis reports 81% extraction for tables versus 23% for prose. Put a direct, 40- to 60-word answer under a question heading near the top, too: 44.2% of citations in that analysis came from the first 30% of page text.

The table is not a trick for a model. It forces you to make the comparison legible. If your alternatives page could not help a buyer eliminate you, rewrite it. A list of names wrapped around a sales pitch wastes both parties’ time.

Make pricing readable without a sales call

Reserve 8 of the 48 page-building hours for pricing. Publish one page with either three tiers or a “starting at” anchor. Say what is included, what costs extra, and who should not buy. A CFO should be able to understand the shape of the purchase in 30 seconds.

A procurement question such as “what does X cost for 50 seats?” needs a number the vendor has actually published. A blank Contact Us button cannot supply one. Pages that hide behind “contact for pricing” carry 38% higher bounce; the same source reports that transparent pricing converts demos 17% better and closes 14% faster. It also reports that a starting anchor converts 22% better than a blank Contact Us page.

This does not require pretending every buyer has an identical contract. It requires publishing the price information you can stand behind instead of making the buyer book a meeting to learn the first useful number. If pricing is part of the decision, give it a page.

3. Weeks 9–10: Remove the technical citation blockers

Cost: 16 hours. Buys: readable, discoverable pages. Do this after you have pages worth checking. Perfecting a crawl setup for a comparison that still does not exist is a peculiar use of a small quarter.

Work through the blockers in order:

1. Check robots.txt. Allow OAI-SearchBot and PerplexityBot to access the pages while keeping GPTBot blocked for training if that is your policy.
2. Confirm the pages appear in the Bing index and that their answers are present in server-rendered HTML.
3. Check the page itself: a question heading with a direct answer, an HTML table where a comparison needs one, and a numbered list where the order genuinely matters.
4. Log what you changed so the next person publishing a buyer page does not undo it.

FAQ sections are cited at 2.3 times the rate in one analysis. That is a reason to answer real questions clearly, not to bolt a dozen empty FAQs onto every page. The same source describes different discovery clocks: days for Perplexity, 1 to 3 weeks for ChatGPT, and 4 to 8 weeks for AI Overviews. Do not declare a page dead because one engine has not picked it up the morning after publication.

The test is whether a system can read the answer you wrote. If it cannot, fix that before debating another headline.

4. Weeks 11–12: Put evidence on pages you do not own

Cost: 16 hours. Buys: third-party proof. Your own pages can explain a comparison. They cannot provide independent confirmation of every claim you make about yourself. That is why the next two weeks go to review profiles and an industry directory, not a speculative press pitch.

Brands are 6.5 times more likely to be cited through third-party sources than through their own domains, and 89% of unbranded prompts are fulfilled by third-party sources. I would not respond by trying to manufacture agreement across the web. I would make the places where buyers already check vendors accurate and complete.

Claim or update G2, Capterra, and one relevant industry directory. Add pricing bands where you can support them. Add genuine customer quotes with headcounts and timelines where you have permission and the details. Ask five happy customers for fresh reviews; do not write those reviews for them. Brands with active review profiles appear in 75.3% of AI answers versus 1% without, and the same source puts complete G2 or Capterra profiles at roughly three times the citation odds.

This is where other people’s pages can beat your blog for an unbranded shortlist. Treat a directory entry as a buyer-facing page, not a form somebody filled in three years ago. Correct the facts first; ask for reviews second. A fresh testimonial beside an outdated price is not proof of much.

5. Week 13: Count enquiries before you count positions

Cost: 8 hours. Buys: a decision about the next page to improve. Rankings belong in the appendix. A page at position 6 that produces three demos a month is doing more for sales than a position 1 blog that produces none.

Use a spreadsheet, not a new visibility dashboard. Record four things:

• Demos or contact forms from the new pages.
• Cited mentions in ChatGPT and Perplexity for the 15 questions.
• Visits referred by an AI answer where you can identify them.
• Pipeline value, if your CRM can connect it to those enquiries.

Test the same questions in clean sessions and log the date, prompt, engine, whether you were cited, and which page was linked. Screenshot the answers. Test engines separately: only about 11% of domains are cited by both ChatGPT and Perplexity. An appearance in one is not evidence that you appear in the other.

For pages published early enough to have had six weeks, check the first screen if there is no enquiry path, cited mention, or visit. Does it answer the question? Can the buyer see a price or a tradeoff without digging? Rewrite that screen before commissioning something new. For later pages, record the baseline and give them time; the clocks are different.

I have seen teams pay for dashboards that report 40 keywords going up while sales hears nothing. That is what the deck calls synergy. I call two people doing the same job. Report enquiries per page to your CEO. If two comparison pages and the pricing page each drive one qualified enquiry a month, that is a better quarter than a stack of blogs nobody brings to a buying meeting.

What stays cut when the 104 hours get tight

The temptation at this point is to sneak the blog calendar back in, or to publish a thin alternatives page because the outline is already sitting in a document. Do neither. If a page takes longer than planned, cut the lowest-priority comparison or the second alternatives page. Do not cut the pricing facts, the honest tradeoffs, or the enquiry path from every page just to hit a page count.

A thin alternatives list with three vendors and no prices does not do the job of a real shortlist. A sharp comparison with numbers can. The accounting-firm version of this 90-day idea makes the same call: fewer pages tied to enquiries beat more pages tied to traffic. The constraint should reduce the number of answers you publish, not the usefulness of each answer.

When the constraint lifts, add names—not a blog calendar

Do the hourly math once. Your only marketer is spending 104 hours on comparisons, pricing, technical checks, reviews, and measurement. If that person also owns campaigns, events, and sales support, even an eight-hour weekly plan can slip. That is when handing the work over may cost less than doing it late.

groas runs SEO and AI Search for a flat $1,499 a month, with new pages, technical fixes, and authority work; it goes live in under 24 hours after onboarding. No percentage of spend. No junior learning on your budget. You set the enquiry goal and guardrails, a named human owns the outcome, and actions are logged with their reasoning.

When the constraint lifts next quarter, do not add blogs by default. Add the next three competitors, the next integration page, and one partner site that lists you unprompted. Keep answering the questions buyers use to cut a shortlist. That is how the first 104 hours keep working after you stop writing.