AI for Marketing & Sales Beginners: 7 Days to Results

Think of modern “chat AI” as a prediction engine for language. It has learned patterns from large amounts of text, and it guesses the next best words based on your request. That’s why it can write a subject line, suggest a value proposition, or turn bullet points into a landing page section. It is not “thinking” the way a person does, and it does not automatically know your business, your customers, or your latest pricing unless you provide that information.

This is the first milestone: treat AI as a helper, not a magic button. When you ask it for marketing or sales work, it will often sound confident even if it’s wrong. In marketing, that shows up as inflated claims (“guaranteed results”), fake statistics, or generic positioning. In sales, it shows up as assumptions about a prospect’s needs, made-up company details, or awkward personalization. Your job is to supply the key facts and constraints, then evaluate the output like you would evaluate a draft from a junior teammate.

Engineering judgment matters here. A good rule: AI is strong at drafting and variation; you are responsible for correctness and strategy. Use AI to accelerate the parts that are slow (first drafts, ideation, restructuring), and keep humans responsible for the parts that carry risk (claims, pricing, compliance language, competitor comparisons, and anything that must be accurate).

• AI can: generate options, rewrite for clarity, adapt tone, summarize notes, propose test ideas, and help you structure campaigns.
• AI cannot (reliably): verify facts, access your private data unless you provide it, or guarantee conversion outcomes.

If you keep this boundary clear, AI becomes a consistent productivity tool instead of a source of random, risky text.

Your second milestone is choosing your first use case. Beginners progress faster when they pick one “lane” for the first week—content, leads, or outreach—then expand. AI can help in all three, but the best starting point depends on what would move your business forward this week.

Marketing tasks AI speeds up: drafting landing page sections, generating social post variations, creating ad angle ideas, rewriting copy to match a brand voice, producing FAQs from product notes, and turning a webinar transcript into multiple content pieces. These are high-leverage because you can review quickly and ship improvements daily.

Sales tasks AI speeds up: writing outreach drafts, preparing call agendas, summarizing call notes into next steps, generating objection-handling bullets, and producing follow-up emails that reference the buyer’s stated goals. The value comes from consistency and speed—fewer “blank page” moments and better follow-through.

Lead work AI speeds up: turning a target market description into simple customer profiles, creating qualification questions, drafting a basic prospect list format (you still need to source the data), and suggesting segmentation fields for email campaigns.

Common beginner mistake: trying to do everything at once. Instead, choose one:

• Content-first: if you need assets (landing page, ads, posts) to support demand.
• Leads-first: if you don’t have enough qualified names to contact.
• Outreach-first: if you have leads but follow-up is inconsistent or slow.

You’re not locking yourself in—this is just your Day 1–2 focus so you can create momentum and measurable output.

AI output quality is mostly determined by input quality. A “prompt” is simply your instruction, but the best prompts include three ingredients: context (what you sell and to whom), constraints (tone, length, rules), and examples (what “good” looks like). This is the foundation for Milestone 5: creating your first good prompt and saving it.

Here’s a practical prompt structure you can reuse across marketing and sales tasks:

• Role: “Act as a B2B SaaS copywriter…”
• Goal: “Draft landing page hero copy that drives demo requests…”
• Audience: “For operations managers at 50–200 person logistics companies…”
• Offer facts: “We provide X, priced at Y, with these proof points…”
• Brand voice: “Plain, confident, no hype, short sentences…”
• Constraints: “No unverified claims, avoid these words, 3 options…”
• Output format: “Return a table with headline/subhead/CTA…”

When you don’t provide facts, AI fills gaps with plausible-sounding guesses. That’s the source of many marketing problems: claims that cannot be supported, features you don’t actually have, or “industry stats” without sources. Your fix is simple: add a short “fact box” to your prompt. Even 6–10 bullets dramatically improves accuracy.

Examples are the fastest way to get on-brand. Paste one paragraph of existing copy you like (from your site or emails) and say, “Match this style.” If you don’t have copy yet, describe the tone using opposites: “friendly but not casual; expert but not academic; persuasive but not pushy.”

Finally, ask for options instead of “the best.” Marketing and sales are iterative. A good prompt produces 3–10 viable variants you can test, combine, and refine.

Milestone 3 is setting up a simple workspace so your AI work doesn’t disappear into chat history. You need four basic tools: a chat interface (your AI assistant), a document for “source of truth,” a spreadsheet for tracking, and a notes file for prompts and snippets.

1) Chat (your drafting room). Use chat for ideation, drafting, rewriting, and structured outputs. Keep one conversation per project (e.g., “Landing page for Product A”) so context stays clean. If a thread gets messy, start a new one and paste your fact box again—clean context beats long context.

2) Docs (your source of truth). Create one living document called “Brand + Offer Basics.” Store your approved positioning, customer profiles, proof points, pricing rules, and compliance notes. This is what you paste into prompts to keep outputs consistent. Beginners skip this and then wonder why every AI draft sounds different.

3) Sheets (your results tracker). Set up a simple table with columns like: Date, Asset type (post/ad/email), Prompt used, Time spent, Output shipped (Y/N), Metric (CTR/replies/conversions), Notes. This makes progress visible and keeps you focused on outcomes, not word count.

4) Notes (your prompt library). Save your best prompts and your best outputs. Treat this like reusable code. Name prompts clearly: “LP-Hero-3Options,” “ColdEmail-FirstTouch,” “AdAngles-ProblemAware.”

This toolkit is intentionally basic. You don’t need complex automation to get results in 7 days; you need repeatability: the same inputs, the same checks, and a record of what worked.

Before you publish or send anything generated by AI, run three safety checks. This is Milestone 4, and it prevents most beginner failures.

Check 1: Truth (is it accurate and supportable?). Highlight every specific claim: numbers, time savings, “best,” “leading,” guarantees, competitor comparisons, customer logos, certifications, and results. Verify each one against a real source (internal data, case study, policy, or approved sales deck). If you can’t verify it, rewrite it into a safe version (e.g., “helps reduce manual work” instead of “cuts work by 37%”). Also watch for invented “industry statistics.” If you need a stat, require a citation and verify it yourself.

Check 2: Brand voice (does it sound like you?). AI defaults to generic marketing language. Create a short voice checklist and apply it every time:

• Is it plain language or buzzwords?
• Does it match our confidence level (no hype)?
• Is it specific about who it’s for and what it does?
• Does it use our preferred terms (product names, feature labels)?

If it fails, don’t just say “make it better.” Give targeted instructions: “Shorten sentences to under 12 words; remove adjectives; replace ‘revolutionary’ with concrete benefits; keep 1 CTA only.”

Check 3: Compliance (is it allowed?). Depending on your industry, compliance may mean legal review, platform policies, privacy rules, or brand safety. At minimum: avoid collecting sensitive personal data, avoid discriminatory targeting language, avoid unsupported medical/financial claims, and respect email laws (unsubscribe, truthful subject lines, identification). If you have internal rules, paste them into the prompt as constraints so the first draft is safer.

These checks take minutes and turn AI from “risky shortcut” into “controlled acceleration.”

This course is designed for results in one week, but “results” must be measurable. Milestone 6 is committing to a short plan and a simple scoreboard. The point isn’t to generate more content; it’s to move a real marketing or sales metric.

Use this beginner-friendly 7-day rhythm:

• Day 1: Choose your lane (content, leads, or outreach). Build your “Brand + Offer Basics” doc with 10–20 bullets of facts.
• Day 2: Create one customer profile and one value proposition draft from your fact box. Save your first reusable prompt.
• Day 3: Draft and finalize one core asset (landing page section, ad set, or email #1–2). Apply the three safety checks.
• Day 4: Create variations (headlines, CTAs, subject lines). Ship them into a real channel (post, ad platform draft, email tool).
• Day 5: Build the next assets (finish a 5-email sequence or a week of posts). Keep prompts consistent.
• Day 6: Create a basic prospect list format and write personalized outreach messages using safe personalization (role, company type, known pains—not private data).
• Day 7: Review your tracker, pick one winner, and document what changed the metric.

How to measure “real results” as a beginner: choose one primary metric and one secondary metric. Examples: replies (primary) and meetings booked (secondary) for outreach; CTR (primary) and conversions (secondary) for ads; sign-ups (primary) and bounce rate (secondary) for landing pages. Also track time saved, but treat it as a bonus metric—not the goal.

End the week by saving your best prompt, best-performing output, and the exact metric lift. That becomes your starting kit for the next campaign—and proof that AI is a reliable helper when you run it with process, facts, and checks.

Fast customer understanding comes from four anchors: customer (who), problem (why now), promise (your outcome), and proof (why believe you). Most beginner marketing skips straight to “promise” (big claims) without specifying the customer or proving anything, which creates generic copy and weak conversion.

Start with a simple template you can reuse in any channel:

• Customer: “For [specific group] who [situation/context]…”
• Problem: “They struggle with [pain] because [cause/constraint].”
• Promise: “We help them [achieve outcome] by [mechanism].”
• Proof: “Here’s why it works: [evidence: results, demo, credentials, reviews, process, guarantees].”

This structure is the foundation for Milestone 1 (benefits) and Milestone 4 (value proposition). AI can draft multiple versions quickly, but you must set boundaries: what you can honestly claim, how fast results typically appear, and which outcomes are realistic for most customers.

Reusable prompt: “You are a marketing strategist. Using the inputs below, write 5 versions of a customer-problem-promise-proof statement. Keep claims conservative and verifiable. If proof is missing, list what proof to collect. Inputs: Offer description, target industries, price range, delivery timeline, past results (if any), differentiators, constraints.”

Common mistakes: using a job title as a “customer” without context (e.g., “marketers”); stating features as outcomes (“AI-powered dashboard”); and using proof that doesn’t reduce risk (awards that don’t matter, vague “trusted by” lists). Your practical outcome is a short paragraph you can paste at the top of a landing page and test immediately.

AI outputs only get as specific as your inputs. The fastest path is to collect a small, high-signal “source pack” in 30–45 minutes—then stop. You are not writing a thesis; you are building a working model of your customer.

Collect these items (even if imperfect):

• Your offer facts: what you sell, pricing structure, onboarding steps, time-to-value, what success looks like.
• Customer language: 5–10 lines from emails, DMs, call notes, support tickets, or reviews (yours or competitors’).
• Competitive context: 2–3 competitor homepages and one comparison/alternatives page.
• Objections you hear: “too expensive,” “no time,” “already tried X,” “need approval,” etc.
• Proof assets: a case study, testimonial snippets, metrics, screenshots, certifications, or a short demo outline.

If you don’t have customer language yet, don’t panic. Use proxy sources: Amazon/G2/Capterra reviews for adjacent products, Reddit threads, LinkedIn comments, YouTube reviews, or sales Q&A pages. The key is to capture real phrasing, not your internal jargon.

Reusable prompt: “Act as a research assistant. From the pasted notes, extract (a) recurring pains, (b) desired outcomes, (c) constraints (time, budget, skills), (d) exact phrases customers use, and (e) implied decision criteria. Output as bullets with short quotes. Do not invent facts; if data is missing, flag it.”

Engineering judgment: prefer fewer, cleaner inputs over massive paste-dumps. Too much noise leads to bland averages. Your practical outcome is a one-page source pack you can reuse across prompts in the rest of the chapter.

Customer profiles are useful only if they drive decisions: what to say, what to show, and what to offer. Forget 20-field persona templates. Build 2–3 profiles that represent distinct buying situations, not just demographics.

A usable profile includes: (1) context, (2) job-to-be-done, (3) success metric, (4) top 3 pains, (5) top 3 objections, (6) buying triggers, (7) decision process, and (8) best channels/tone. AI can draft these quickly from your source pack.

Reusable prompt: “Using the source pack below, create 3 distinct customer profiles for this offer. Each profile must differ by buying situation (e.g., new vs scaling, budget holder vs evaluator). For each, include: context, goals, pains, triggers, objections, decision criteria, what they need to see to trust us, and a 2-sentence ‘how to talk to them’ note. Use real phrases from the source pack where possible. If uncertain, mark as hypothesis.”

Then validate. AI-generated profiles are hypotheses until proven. Use a simple validation loop:

• Spot-check with reality: do these profiles match your last 10 prospects?
• Run a “contrast test”: if two profiles would respond to the same headline and offer, you don’t have distinct profiles yet.
• Ask 3 people: a customer, a sales rep, a support person. “Which profile is closest? What’s wrong/missing?”

Common mistakes: profiles that are too broad (“small business owners”), too flattering (“growth-minded visionary”), or too internally focused (“needs our platform”). The practical outcome for Milestone 2 is three profiles you can label and use: e.g., “Time-poor operator,” “Skeptical finance gatekeeper,” “Ambitious marketer under pressure.”

Conversions improve when you handle resistance before it becomes a “no.” Objections are not a nuisance; they are the map of perceived risk. Pair them with triggers (why buy now) and decision factors (how they choose).

Start by separating objections into categories:

• Value: “Is it worth it?”
• Fit: “Will it work for my situation?”
• Trust: “Do I believe you?”
• Effort: “Do we have time/skills to implement?”
• Risk: “What if it fails? What’s the downside?”

Then list buying triggers: a new quarter, a pipeline gap, churn spike, ad costs rising, a product launch, a new boss, or a compliance deadline. Finally, capture decision factors: budget owner vs influencer, required integrations, security concerns, contract terms, and what proof they need (case studies, ROI calculator, demo, trial, references).

Reusable prompt: “From these profiles and source notes, produce a table with columns: Objection, Category (value/fit/trust/effort/risk), What they’re really afraid of, Best response (1–2 sentences), Proof asset to use, and Where to address it (landing page section, ad, email, sales call). Keep responses honest and specific; avoid hype.”

Practical outcome: you now have copy direction for FAQs, email replies, and sales scripts. This is also where Milestone 3 starts to emerge: each objection points to a message you must include in your problem-to-solution map.

A message map connects what the customer feels (pains) to what they want (gains), to what you do (features), to what changes (outcomes). This prevents the classic beginner error: listing features without translating them into results.

Build the map per profile. One offer can have different “lead” pains and outcomes depending on who’s buying. Use a simple grid:

• Pain: what’s frustrating, costly, risky, or time-consuming
• Gain: what they want instead (speed, clarity, predictability, confidence)
• Feature/Mechanism: what you provide (process, tool, service element)
• Outcome: measurable or observable result (time saved, fewer errors, higher conversion)

Reusable prompt: “Create a message map for each customer profile. For each profile, list 5 pains, the corresponding gains, the feature/mechanism of our offer that addresses it, and the outcome phrased as a believable result. Then generate: (a) one core value proposition sentence, (b) 5 tagline options, and (c) 3 headline/subheadline pairs for a landing page. Use the source pack language; avoid absolute claims.”

This is where Milestone 4 becomes straightforward: your value proposition is simply the best-performing pairing of customer + problem + promise, supported by proof and framed in the customer’s words.

Common mistakes: mixing gains and outcomes (“feel confident” vs “ship campaign in 2 days”); using features that don’t map to a pain; and creating taglines that sound clever but communicate nothing. Your practical outcome is a map you can use to generate ads, landing pages, and email sequences in the next chapters without starting from scratch.

Consistency is a sales advantage. A simple voice guide prevents AI from producing a different personality every time you run a prompt. Your goal is not “brand poetry.” It’s a one-page starter kit that keeps tone, vocabulary, and claims aligned with your real positioning (Milestone 5).

Define voice using practical sliders: direct vs playful, premium vs friendly, technical vs plainspoken, bold vs cautious. Then codify it as rules AI can follow.

• Voice principles (3–5): e.g., “Clear over clever,” “Specific over hype,” “Teach, don’t boast.”
• Words to use: 10–20 preferred terms customers recognize (from your source pack)
• Words to avoid: jargon, exaggerated promises, vague fillers (“revolutionary,” “game-changing”)
• Claim boundaries: what you can/can’t say (timeframes, guarantees, ROI claims)
• Formatting preferences: sentence length, bullets, emoji policy, CTA style

Reusable prompt: “Create a one-page brand voice guide for this business. Inputs: audience, positioning, source pack phrases, compliance constraints. Output: voice principles, tone sliders, words to use/avoid, do/don’t examples (rewrite 3 sample lines), CTA patterns, and a short ‘brand bio’ paragraph. Keep it practical for use in future prompts.”

Engineering judgment: keep the guide short enough that you will actually paste it into prompts. The practical outcome is a consistent voice that makes your landing pages, ads, and emails feel like one coherent brand—while still being tailored to each customer profile.

Most beginners ask AI for “a post” or “an ad” and then wonder why results are random. The fix is to decide the job of the content before you generate it. For practical marketing and sales, most content fits into three goals: awareness (get noticed and understood), trust (reduce uncertainty), and action (create the next step). AI performs best when you tell it which goal you’re targeting and how you’ll measure success.

Awareness content is about clarity, not persuasion. You’re teaching the market what category you’re in, who you help, and what problem you solve. Trust content is about proof: how it works, what makes it safe, what results are typical, and what to expect. Action content is about a single next step: book a call, start a trial, download a guide, or request a quote.

A simple rule for beginners: one piece of content should have one primary goal. When you mix goals, AI tends to write generic copy (“We are passionate about…”). Instead, specify the stage and the audience temperature. For example: “cold audience who doesn’t know us” (awareness), “warm leads who are comparing options” (trust), or “high-intent visitors on a pricing page” (action).

• Common mistake: asking for “high-converting copy” without defining the offer, audience, and proof available.
• Better instruction: “Write 3 options aimed at trust: explain process, address top 2 objections, and include one proof point we can substantiate.”

Milestones in this chapter build on these goals: your week of posts should intentionally mix awareness, trust, and action; your landing page sections should move a reader from understanding to confidence to next step; and your ads should align the hook and benefit to one goal at a time.

Your productivity breakthrough is to stop briefing every asset from scratch. Create one “source of truth” prompt—your one brief—that includes the details AI needs to stay on-brand and accurate. Then you can ask for a week of posts, landing page sections, and ad angles using the same inputs.

Use this reusable template (fill in brackets):

ONE BRIEF TEMPLATE
You are a marketing copy assistant. Write in this brand voice: [3-6 adjectives, e.g., practical, friendly, direct, not hypey].
Business: [what you sell, to whom, where].
Primary customer: [role, industry, context].
Core problem: [pain + why it matters].
Desired outcome: [result customer wants].
Your mechanism/process: [how it works in 3-5 steps].
Differentiators: [3 bullets].
Proof available: [testimonials, case studies, years, numbers you can verify]. If none, say “no quantified proof.”
Offer: [what they get].
CTA: [single action].
Constraints: Do not invent facts. Avoid guarantees. No competitor claims. Keep compliant with [any industry rules].
Message map (must reflect): [1) headline promise, 2) key benefits, 3) top objections + responses].

Now you can hit Milestone 1 quickly: “Using the one brief, generate 7 social post drafts (mix awareness/trust/action). Provide a posting schedule, a hook, body copy, and CTA for each.” If the output feels repetitive, don’t throw it out—tighten the brief. Add a “content pillars” line (e.g., education, behind-the-scenes, customer story, offer), and ask for one post per pillar.

• Common mistake: leaving “proof” blank, then the AI invents statistics.
• Fix: explicitly say what proof exists and instruct “use placeholders like [INSERT TESTIMONIAL] if needed.”

Once your one brief exists, every future prompt becomes a small request: “Using the brief, write X in Y format for Z audience.” That is how you create content and ads in one afternoon without losing coherence.

A landing page is not a brochure. It’s a guided decision page with a single job: move a specific visitor to a specific next step. AI can draft landing page sections quickly, but only if you tell it the page purpose and the visitor intent (cold vs. warm). For Milestone 2, you’ll draft sections that match your message map so your page doesn’t feel like stitched-together paragraphs.

Use this practical structure as your default:

• Headline: clear promise + audience (“Get X without Y”).
• Subhead: expand with context, who it’s for, and time/effort truthfully.
• Proof: testimonial, client logos, before/after (only if real), process preview, or credibility statements you can verify.
• Offer: what they get, what’s included, what’s not included, and who it’s best for.
• CTA: one action, one button label, one expectation statement (“Takes 2 minutes”).

Prompt example (using your one brief):
“Draft landing page sections: (1) hero headline + subhead + primary CTA, (2) 3 benefit blocks tied to the message map, (3) ‘How it works’ in 4 steps based on our process, (4) objection handling (pricing/time/risk), (5) proof section using only provided proof or placeholders, (6) final CTA with a low-friction reassurance. Keep language concrete and avoid guarantees.”

Engineering judgment: landing pages fail when they are vague. Replace “streamline” and “optimize” with specific outcomes (“reduce manual follow-ups,” “get qualified leads in your calendar”). Also watch for hidden claims: AI may imply outcomes (“double your sales”) that you can’t back up.

Finally, consistency matters: if your ads promise “book a demo,” your landing page CTA should also be “book a demo,” not “download a guide.” Mismatched CTAs are a common beginner mistake and create drop-off even with good copy.

Ads are compressed strategy. You have seconds to earn attention, so structure matters. A simple, reliable ad formula is: Hook (pattern interrupt), Benefit (what improves), Audience (who it’s for), and CTA (what to do next). AI is excellent at producing many options quickly, which is why Milestone 3 is to create 10 ad angles and pick 3 to test.

Start by defining “angles” correctly: an angle is not just a different headline—it’s a different reason to care. Common angles include: save time, reduce risk, avoid a mistake, social proof, process transparency, cost control, speed, simplicity, personalization, or compliance/safety (if relevant).

Prompt example:
“Using the one brief, generate 10 distinct ad angles. For each: (a) angle name, (b) 2 hook options, (c) primary benefit statement, (d) who it’s for, (e) CTA line. Constraints: no guarantees, no invented data, no ‘best/number one’ claims. Keep under 30 words per ad concept.”

Then choose 3 to test using practical criteria:

• Truth: you can deliver the promise reliably.
• Clarity: a stranger understands it in one pass.
• Relevance: matches a real pain for your target profile.

Common beginner mistake: testing too many variables at once. If you change the angle, creative, audience, and landing page all at the same time, you won’t learn what worked. Keep your test clean: pick 3 angles, keep the same offer and landing page, and vary only the hook/creative within each angle.

Repurposing is where AI saves the most time. You already have a message map and a one brief, so you can produce coherent variations without rewriting from zero. This is Milestone 4: on-brand rewrites for different platforms. The key is to repurpose the idea, not just reformat text. Each platform has a different reader expectation and constraint (length, tone, structure, hashtags, links, visual pairing).

Pick one “seed” asset—often a landing page section, a customer story, or an objection-handling paragraph—and create a repurposing batch:

• LinkedIn post: insight + takeaway + soft CTA.
• Instagram caption: punchy hook + short lines + CTA to comment/DM.
• X/Twitter: one sharp point + optional thread of 3–5 posts.
• Email snippet: personal opening + one benefit + single CTA link.
• Short video script: hook in first 2 seconds + 3 beats + CTA.

Prompt example:
“Repurpose this seed text into: (1) LinkedIn post (120–180 words), (2) Instagram caption (max 120 words), (3) X thread (5 tweets), (4) email to warm lead (100–130 words). Keep the same core claim, but adapt pacing and format. Maintain brand voice. Do not add new facts; if proof is needed, insert [PROOF PLACEHOLDER].”

Engineering judgment: create a “house style” so repurposed content still sounds like you. If your brand avoids hype, explicitly ban phrases like “game-changer,” “crushing it,” “10x,” or “guaranteed.” Also watch for repetitive openings—AI loves starting every post with a question. Tell it: “No question hooks in this batch,” and you’ll get more variety.

When done well, one idea can produce the full week of posts from Milestone 1 and also feed ads and email—without drifting off-message.

AI makes drafting fast; editing makes it effective and safe. To ship content quickly, you need a repeatable editing approach and a lightweight approval checklist (Milestone 5). Your job in editing is not to “make it sound nicer.” It’s to ensure the copy is clear, specific, and compliant with your business reality and any platform/industry rules.

Use a three-pass edit:

• Pass 1: Clarity. Can a stranger explain what you do after reading it once? Remove jargon. Replace “solutions” with what you actually deliver. Keep one CTA.
• Pass 2: Specificity. Add concrete details: who it’s for, what’s included, what happens next, time/effort expectations. Remove generic filler (“we’re passionate”).
• Pass 3: Compliance and truth. Check claims, proof, pricing, guarantees, and comparisons. If you can’t substantiate it, rewrite it or remove it.

Now convert that into a simple approval checklist you can reuse:

• Claims are accurate and verifiable (no invented stats, no implied guarantees).
• Offer and CTA match the landing page and funnel step.
• Audience and pain are explicitly stated (no “for everyone”).
• Proof is real or clearly marked as placeholder.
• Tone matches brand voice (approved vocabulary; banned phrases avoided).
• Compliance checks passed (industry rules, testimonials permissions, platform policies).

Common mistake: approving copy based on “it sounds good.” Instead, approve based on whether it’s true, clear, and aligned. If you want AI help editing, ask it to critique against your checklist: “Review this draft. Flag any vague phrases, unsupported claims, or compliance risks. Suggest safer rewrites.” That keeps AI in the assistant role while you retain final responsibility.

By the end of this chapter, you should have a repeatable system: one brief → week of posts → landing page sections → ad angles → platform rewrites → checklist-based approval. That’s how beginners get real output in an afternoon, without turning their brand into generic AI copy.

Email marketing works when your list is built on permission and your content has a clear purpose. Permission means the subscriber expects to hear from you (opt-in, checkout, webinar registration). Purpose means you know what this email is trying to accomplish: welcome, educate, convert, retain, or re-engage. Without those fundamentals, AI-generated copy will amplify the wrong thing—more volume, more noise, more unsubscribes.

Start by defining your list segments in plain language. Example: “New leads from the ebook,” “Trial users,” “Customers in month 1,” “Past customers,” “Cold subscribers (no opens in 90 days).” You do not need dozens of segments to begin; you need one audience you can describe clearly. This sets you up for Milestone 2 (a 5-email nurture sequence for one audience) and Milestone 5 (a re-engagement email for cold subscribers).

Next, write down your sending expectations: frequency, content types, and what subscribers will get. This becomes the backbone of Milestone 1: a welcome email that sets expectations clearly. A strong welcome email reduces spam complaints because it reassures the reader: “You signed up for X; here is what will happen next.”

• Engineering judgment: Keep promises small and specific. “Weekly practical tips” beats “valuable insights.”
• AI prompt starter: “Draft a welcome email for [audience] who opted in via [source]. Set expectations for [frequency], topics [A/B/C], and the next step [CTA]. Tone: [brand voice]. Constraints: no hype, no false urgency, 120–180 words.”

Common mistake: treating every email as a sales email. If the list feels like a funnel instead of a relationship, your metrics will drop quickly and deliverability will suffer long-term.

“Sounds human” is often just good structure. AI is excellent at generating words; your job is to enforce the shape of a helpful message. Use a repeatable frame: Subject (promise), Opening (context and empathy), Value (one idea), Proof (credibility), CTA (one next step).

Start with the opening. The opening should prove you understand the reader’s situation. Avoid generic lines like “Hope you’re well.” Instead: “If you downloaded the checklist, you’re probably trying to [job-to-be-done] without [common pain].” That immediately feels written for them.

Then deliver one piece of value, not five. A frequent AI mistake is to overstuff. One email = one main point. Your proof can be light: a quick stat, a mini case example, a quote, or your own experience. Finally, use a single CTA. Multiple CTAs read like a template and reduce clicks.

Milestone 3 fits here: create subject lines and preview text for A/B tests. Subject and preview work as a pair. The preview should not repeat the subject; it should add information or create curiosity ethically.

• A/B idea: Test “Benefit” vs “Curiosity.” Subject A: “3 ways to cut demo no-shows” vs Subject B: “The no-show fix nobody wants to do.” Keep the body identical.
• Prompt: “Generate 10 subject lines and 10 preview texts for this email body. Requirements: 35–55 characters for subject, 70–100 for preview. Avoid spam words. Include 3 ‘how-to’ subjects, 3 ‘mistake’ subjects, 2 ‘question’ subjects, 2 ‘short’ subjects.”

Engineering judgment: only test what you can learn from. If you change subject, CTA, and offer at the same time, you won’t know what caused the result.

A nurture sequence is not “five sales emails.” It is a short curriculum that moves a reader from interest to confidence. The simplest 5-email structure is: (1) Welcome + orientation (Milestone 1), (2) Quick win, (3) Common mistake + fix, (4) Proof + story, (5) Offer + clear next step. If your audience is already warm (trial users), you might shorten education and emphasize activation. If your audience is cold (ebook leads), you educate and reassure more.

When you build Milestone 2 (a 5-email nurture sequence for one audience), begin with two lists: what they want and what they fear. They want an outcome (more leads, faster sales cycles). They fear wasting time, buying the wrong tool, or looking foolish. Emails 2–4 should reduce that fear with clarity and small wins.

AI helps you produce variations, but you must supply the strategy inputs: audience, pain points, value proposition, and offer. A practical workflow is: write one “sequence brief,” then ask AI to draft each email using the same brief, then you edit for truth and specificity.

• Sequence brief template: Audience = [who]; Desired outcome = [result]; Main obstacles = [3]; Your approach = [how you solve]; Proof assets = [case study/testimonial/stat]; Offer = [call/book/buy]; Brand voice = [adjectives].
• Prompt: “Using this sequence brief, draft Email #2 (Quick win). Include: a 1-sentence promise, a 3-step tip, a short reassurance line, and a soft CTA to [resource]. Keep under 200 words.”

Common mistake: making every email “high stakes.” Human sequences alternate intensity: helpful email, then a slightly more direct one, then helpful again. That rhythm keeps attention and reduces unsubscribes.

Personalization is not adding a first name. Real personalization is relevance: using context the subscriber expects you to have. “You attended the webinar on Tuesday” is expected. “I noticed you visited our pricing page at 2:14pm” feels invasive unless you’ve clearly disclosed tracking and it’s used sparingly. Your goal in Milestone 4 is to add personalization safely using a small data set—enough to be relevant, not enough to be spooky.

Use “safe fields” you collected directly: first name, company, role, industry, use-case selected on a form, product tier, lifecycle stage (lead/trial/customer), and the specific asset they requested. Then personalize the opening and the example, not the entire email. Over-personalized emails often look machine-generated because they cram tokens into every line.

Practical method: create 3–5 message variants by segment, not 500 one-to-one emails. For example, one version for agencies, one for SaaS, one for local services. AI can write each variant quickly if you give it guardrails and your approved data fields.

• Small data set example: {first_name, company, industry, goal, lead_source}
• Prompt: “Rewrite this email into 3 variants for industries: [A], [B], [C]. Use only these fields: [list]. Do not infer revenue, location, or behavior. Keep the CTA identical.”

Engineering judgment: if you didn’t explicitly collect a field, don’t imply you know it. Also, don’t personalize sensitive categories (health, finances, personal circumstances) unless you have explicit consent and a strong reason.

Emails fail for two reasons: they don’t get delivered, or they get ignored. Deliverability problems often come from patterns that look like spam: excessive exclamation points, ALL CAPS, “free money” phrasing, aggressive urgency, misleading subjects, and link-heavy layouts. AI will sometimes drift into these because it has seen them in training data. Your job is to remove them.

The second failure is vagueness. “Unlock your potential” is not an offer. “Get tips to grow” is not a reason to click. Humans trust specificity: “Steal our 12-line follow-up that gets replies” or “See 3 subject line formulas we use for onboarding.” Vague emails also make AI edits harder because there’s no truth to check.

Use an editing checklist before sending:

• One promise: Can you summarize the email in 10 words?
• One reader: Is this written to a specific segment?
• One CTA: Is there a single next step?
• Truth check: Are claims accurate and provable?
• Spam scan: Too many links? Too much urgency? Too many images?

This section also supports Milestone 5 (re-engagement for cold subscribers). Re-engagement emails are especially prone to spammy language (“We miss you!!!”). Instead, be direct and respectful: remind them what they signed up for, offer a clear choice, and make unsubscribing easy. Counterintuitively, a clean list improves deliverability and performance.

Email success is measured by movement, not vanity. Opens can be useful, but they are imperfect due to privacy features and image blocking. Still, opens can indicate subject line fit when compared within the same audience and timeframe (useful for Milestone 3 A/B tests). Clicks indicate interest in the content and strength of your CTA. Replies are gold for sales-led motions because they reveal objections and buying signals. Sales and pipeline are the final measure—especially for nurture sequences.

Match the metric to the email’s purpose. Your welcome email (Milestone 1) should be judged by: low spam complaints, solid click-through to the “start here” resource, and early replies (“Here’s my situation…”). Your 5-email nurture sequence (Milestone 2) should be judged by: clicks on educational emails, conversions on the offer email, and overall unsubscribe rate across the series.

Practical tracking setup: tag links with UTM parameters, track conversions on the landing page, and label sequence emails consistently (e.g., NURTURE-LEADS-01). After one week, review the funnel: delivered → opened → clicked → converted. If clicks are low, the value/CTA is weak. If opens are low, the subject/preview or sender reputation needs work. If clicks are high but conversions are low, the landing page or offer is misaligned.

• Prompt: “Given these metrics for 5 emails (opens, clicks, replies, conversions), diagnose the most likely bottleneck and propose 3 copy changes and 1 sequencing change. Keep recommendations realistic and testable.”

Engineering judgment: optimize one step at a time. Email is a system. When you make small, disciplined changes—subject line tests, clearer CTAs, safer personalization—you get compounding gains without losing the human feel.

Section 5.1: What a “lead” is and where leads come from

A lead is a person or company that has a plausible reason to buy and a plausible path to buying. Beginners often label anyone with a job title as a lead. In practice, a lead becomes valuable when you can answer three questions: (1) Why might they care? (2) Why now? (3) How would they buy?

Leads come from two broad sources: inbound and outbound. Inbound leads find you (content, ads, referrals, events, SEO, partnerships). Outbound leads are prospects you identify and contact directly (email, LinkedIn, calls). Neither is “better”; they serve different timelines. Inbound compounds but takes time. Outbound is immediate but requires targeting discipline and consistent follow-up.

• Inbound examples: a webinar signup, a demo request, a download, a referral intro.
• Outbound examples: a list of 50 companies with a specific tool stack, hiring signal, or growth stage.

AI is useful here to turn your business description into plain-language lead definitions. Prompt it to produce a clear “lead rule” you can apply quickly. Example: “A good lead is a US-based HVAC company with 5–50 technicians, running Google Local Services Ads, and mentioning missed calls or scheduling issues.” That definition becomes your Milestone 1 foundation: a prospect list described in plain terms that any teammate can understand.

Common mistake: treating interest as the only qualifier (“They liked our post”). Interest matters, but without fit (budget, need, capability, timing), you’ll chase friendly conversations that never convert.

Section 5.2: Prospecting basics: industries, roles, and signals

Prospecting is the craft of choosing where to focus and who to contact. Start with two axes: industry (who tends to have the problem) and role (who feels the pain and can act). Beginners often pick roles by seniority alone (“I’ll message the CEO”). Instead, message the role closest to the problem and the buying path—sometimes that’s operations, revenue, marketing, sales enablement, or IT.

Use signals to add a “why now” that is not creepy and not speculative. Good signals are public, verifiable, and directly related to your offer. Examples: hiring for a role your product supports, launching a new location, recent funding, a new product line, a tool change, or a public complaint in a review about a solvable issue (e.g., slow response times).

• Industry filter: Where is the pain frequent and expensive?
• Role filter: Who owns the metric you improve?
• Signal filter: What changed that makes action likely now?

Milestone 1 output should be a small list you can actually work: 25–75 prospects is plenty for a beginner. Ask AI to help you define the list criteria and suggest data sources (directories, LinkedIn search, review sites, job boards). But keep your judgment in control: do not scrape private data, do not buy questionable lists, and do not include people without a legitimate interest basis.

Common mistakes: (1) building a huge list to avoid the discomfort of outreach, (2) mixing very different segments in one campaign, and (3) using vague ICP language (“fast-growing companies”) that produces random results.

Section 5.3: Outreach basics: relevance, proof, and one clear ask

Your first outreach message is not a brochure. It’s a short note that shows relevance, offers a credible reason to listen, and makes one easy next step. A reliable structure is:

• Relevance: “I noticed X” (industry/role/signal) + “This usually leads to Y problem.”
• Proof: one line of credibility (a result, a recognizable customer type, a small artifact like a checklist).
• One clear ask: a simple question with low effort (reply yes/no, 10-minute call, permission to send a resource).

Milestone 2 is to draft a cold email plus one follow-up that gets replies. Keep the email under ~120 words. Use plain language. Avoid buzzwords, attachment-heavy emails, and exaggerated claims. AI can generate variations, but you must constrain it: define your voice, banned phrases, and the specific outcome you want (a reply, not a sale).

Example “one clear ask” patterns that work well for beginners:

• “Worth a 10-minute call to see if this is relevant, or should I talk to someone else?”
• “Want me to send a 1-page checklist we use to diagnose this?”
• “Are you the right person for improving X, or is that owned elsewhere?”

Common mistakes: asking for 30–60 minutes too early, packing in multiple offers, or hiding the ask in a paragraph. If a prospect must think to understand what you want, you won’t get replies.

Section 5.4: Personalization from public info (safe and respectful)

Personalization should feel like basic professionalism, not surveillance. The safe approach is to personalize using public, business-relevant information: their role, company focus, recent press release, job post, website messaging, or a specific product/service page. Avoid personal details (family, health, political views) and avoid guessing sensitive attributes.

Milestone 3 is to create short LinkedIn-style messages and call scripts. The same rule applies: keep it short, specific, and respectful. LinkedIn messages work best when they read like a normal human wrote them—one idea at a time. A call script should not be a monologue; it’s a sequence of short questions that earns permission to continue.

• Safe personalization: “Saw you’re hiring SDRs—often a sign the team is scaling outbound.”
• Not safe: “I saw you recently moved / had a life event…”

How AI helps: you can paste a company’s homepage text (or a job post excerpt) and ask AI to propose 3–5 “safe personalization hooks” and convert them into a 250-character LinkedIn note and a 20-second opener for a call. Your judgment decides which hook is real and relevant. Never fabricate (“Congrats on your funding”) unless you verified it.

Common mistakes: over-personalizing with trivia, using flattery as a substitute for relevance, and copying lines that feel “AI-ish” (“I hope this message finds you well…”). Your goal is to sound direct, helpful, and grounded in what you actually know.

Section 5.5: Handling objections with short, helpful replies

Objections are often not rejection; they’re requests for clarity or risk reduction. The beginner trap is arguing. Instead, reply in a way that (1) acknowledges, (2) reduces effort, and (3) offers a next step. Keep replies short enough to read on a phone.

Milestone 4 is to build a simple qualification checklist (fit vs. not fit). This prevents you from pushing leads that should never have been contacted. When an objection arrives, check it against your fit criteria. If they’re not a fit, exit politely and preserve goodwill. If they are a fit but timing is wrong, move them into a follow-up stage with a date.

• “Not interested.” Reply: ask one clarifying question or offer a resource. “Totally fair—should I close the loop, or is this something you revisit next quarter?”
• “We already have a vendor.” Reply: position as complementary. “Makes sense—are you open to a quick comparison checklist to sanity-check coverage?”
• “No budget.” Reply: confirm and reduce scope. “Understood—if I could show a low-cost way to test this in 2 weeks, would that be worth a look?”
• “Send details.” Reply: send one tight asset and ask one question. “Happy to—what matters more for you: speed, cost, or reliability?”

Common mistakes: sending long rebuttals, adding pressure (“Just 15 minutes!”), or continuing after a clear “no.” Your reputation compounds; respectful exits often create referrals later.

Section 5.6: Simple pipeline tracking: stages, next step, owner

Outreach only works when follow-up is consistent. Most beginner systems fail because they are too complex. You need three fields per prospect: stage, next step, and owner (even if that owner is you). This is Milestone 5: set up a basic follow-up system you will maintain.

Use a spreadsheet or a simple CRM. The tool matters less than the habit. Define 5–7 stages max, with clear meanings. Example stages:

• Prospect Identified (meets ICP rules)
• Contacted (message sent)
• Engaged (replied or accepted)
• Qualified (passes checklist)
• Meeting Scheduled
• Closed Won / Closed Lost
• Nurture (future date)

“Next step” must be a verb with a date: “Send follow-up #1 on Wed,” “Ask for referral to ops owner,” “Share checklist + ask one question.” If there is no next step, you do not have a pipeline—you have a history.

AI can help by generating follow-up tasks and reminders from your message logs, but you should keep the system simple enough that you can run it without AI. Common mistakes: creating too many stages, failing to record outcomes, and letting “Nurture” become a graveyard. If you can maintain 10–15 touches per week with accurate tracking, you will outperform most beginners who send 200 messages once and then stop.

Section 6.1: Template library: briefs, prompts, and checklists

Your first system asset is a small “template library.” Templates prevent you from starting from scratch, and they reduce risk because you’re repeating a known-good process. The goal is not clever prompts; it’s consistent inputs that produce on-brand outputs you can approve quickly.

Create three kinds of templates: briefs (what the work is), prompts (how to generate the draft), and checklists (how to review). Keep them in one place: a doc, Notion page, or a “Templates” tab in your spreadsheet.

• Brief template (1 page): audience, offer, proof, tone, CTA, constraints (length, channel, compliance notes), and “do not say” claims.
• Prompt template (copy/paste): role + task + context + style rules + output format. Include brand voice bullets and any required sections (headline, subhead, 3 benefits, CTA).
• Checklist template: accuracy (facts verified), brand (tone matches), compliance (no prohibited claims), clarity (one idea per paragraph), and CTA (specific next step).

A practical way to build this library: take your best-performing post, email, or landing section from earlier chapters and reverse-engineer it. What inputs would you need to reliably produce something similar again? Capture those inputs as fields in the brief, then convert them into a prompt.

Common mistake: templates that are too broad (“Write a marketing email”). Broad prompts create generic output and force you to edit heavily. Better: narrow and structured (“Write Email #2 of a 5-email sequence for webinar leads who downloaded X; objective is to overcome objection Y; include 1 story + 3 bullets; end with CTA to book a call”). Your templates should make good work easier than bad work.

Section 6.2: The weekly cadence: Monday plan to Friday review

Next, put your AI work on a weekly cadence. A cadence is a schedule you can keep even when life gets busy. If you run marketing “when you feel like it,” AI will mostly generate drafts that never get published, measured, or followed up. A simple Monday-to-Friday loop fixes that.

Monday (Plan): pick one campaign focus (one offer, one audience). Use AI to brainstorm themes, but you choose the priority. Output of Monday is a short plan: 1 landing page tweak (optional), 2–3 social posts, 1 email, and 10–20 outreach targets. Lock your metrics for the week (see Section 6.3).

Tuesday (Produce): use your templates to draft content in batches. Generate variations, then select one. Your job is editor-in-chief: remove weak claims, add specific proof, and align with your value proposition. Keep everything in a single “This Week” folder so you don’t lose drafts.

Wednesday (Publish): schedule posts, publish the email, and update landing copy if planned. Use AI for final polish (grammar, clarity, shortening), but do not outsource factual accuracy. If you need citations or product details, pull them from your real sources and paste them in.

Thursday (Follow up): sales motion. Use AI to personalize outreach messages at scale, but keep guardrails: one personalization detail, one clear ask, one CTA. Record who you contacted and what you sent. If you promised a resource, send it same day.

Friday (Review): open your mini dashboard. Look for one insight: what performed best and why. Decide one small test for next week (headline change, new CTA, different audience angle). This is where AI becomes a learning engine instead of a content machine.

Common mistake: trying to do all tasks daily. Batching reduces context switching and makes it easier to measure cause and effect. Another mistake: producing more content instead of publishing and following up. If time is tight, prioritize: publish one asset and do follow-up over drafting five unused ideas.

Section 6.3: Basic metrics that matter for beginners

You don’t need advanced analytics to run a measurable AI system. Beginners do need a few “signal metrics” that connect effort to outcomes. Choose metrics you can collect weekly without friction, and avoid vanity metrics that look good but don’t help decisions.

Use a simple funnel view: Attention → Action → Conversation → Conversion. Track one or two metrics per stage.

• Attention: impressions or reach (by channel). Useful for context, not success by itself.
• Action: link clicks, landing page visits, or content saves. This tells you whether the message earned interest.
• Conversation: replies to outreach, booked calls, demo requests, or inbound inquiries.
• Conversion: purchases, proposals accepted, or trial-to-paid upgrades.

For email, keep it simple: sends, opens (directional), clicks, replies, unsubscribes. For outreach, track: prospects contacted, reply rate, meetings booked. For landing pages, track: visits and conversion rate (form submits / visits).

Engineering judgment matters here. A metric is only useful if it drives a decision. Example: if your post got high reach but low clicks, your next test might be a clearer CTA or a stronger hook. If clicks are high but conversions are low, test the landing page headline or add proof (testimonials, guarantees, FAQs). If outreach reply rate is low, test the first sentence and the offer relevance, not a new AI model.

Common mistake: changing five things at once and then “measuring.” Your metrics should support learning, so keep changes small and deliberate. Also avoid measuring too soon: give each asset enough time to collect a baseline (e.g., 48–72 hours for social, a week for email, two weeks for landing page if traffic is low).

Section 6.4: Iteration: how to learn from results without overwhelm

Iteration is the skill that turns weekly activity into compounding improvement. The trick is to run “small tests” that are cheap, safe, and easy to interpret. You are not trying to perfect everything; you’re improving one bottleneck at a time.

Use a simple loop: Hypothesis → Change → Measure → Decide. Write the hypothesis in one sentence. Example: “If we lead with the customer’s pain instead of our features, outreach replies will increase.” Then change one element: the opening line. Keep the rest constant. Measure the reply rate. Decide: keep, revert, or iterate again.

AI helps you create variations fast, but you should choose the variable. Ask AI for 10 hooks that match your brand voice, then select 2 to test. Avoid testing 10 at once unless you have enough volume; otherwise results will be noisy.

• Content test ideas: headline angle (pain vs. outcome), CTA wording, proof placement (early vs. late), length (short vs. medium), format (bullets vs. paragraph).
• Email test ideas: subject line style, first paragraph, single CTA vs. two options, adding a relevant case study.
• Outreach test ideas: personalization type (role-based vs. company event), the ask (15-min call vs. resource), follow-up timing.

Common mistake: interpreting every fluctuation as meaning. Small numbers swing wildly. When volume is low, focus on qualitative signals too: what objections appear in replies, what questions prospects ask, what people quote back to you. Use AI to summarize responses, categorize objections, and propose improved positioning—then you decide what’s true.

Practical outcome: after four weekly cycles, you’ll have one proven message angle, one higher-performing CTA, and one outreach opener that reliably earns replies. That is a real system: fewer guesses, faster execution.

Section 6.5: AI risk basics: privacy, copyright, and sensitive data

A repeatable AI system must be safe. Most problems come from three areas: privacy, copyright, and accuracy. The rule is simple: if you wouldn’t paste it into a public document, don’t paste it into an AI prompt—unless your organization has an approved, private setup and policy.

Privacy and sensitive data: Never include customer PII (full names, emails, phone numbers), payment details, contracts, internal financials, or health-related data. For personalization, use placeholders (“{FirstName}”, “{Company}”) and store the real data in your CRM or spreadsheet. If you need AI to tailor messaging, summarize the prospect in non-sensitive terms (industry, role, public company info).

Copyright and brand risk: Don’t ask AI to mimic a living author or a competitor’s voice. Don’t paste copyrighted content you don’t own and ask for “rewrites” as a shortcut. Instead, provide your own brand voice guidelines and your own source material (your product pages, your policies, your case studies). Keep a “Claims you can make” list based on real proof.

Accuracy and approvals: AI will invent details confidently. Set an approvals checklist: (1) verify factual claims (prices, guarantees, results), (2) verify compliance (regulated industries, testimonials rules), (3) verify tone (no promises you can’t keep), (4) verify links and CTAs. If you work with a team, define who approves what: e.g., marketing approves voice, sales approves offer, legal/compliance approves claims.

Common mistake: treating AI output as “draft = done.” Your system should assume every output is a draft that requires review. Safety is not a one-time step; it’s built into the workflow.

Section 6.6: Your next 30 days: scaling content, leads, and outreach

To scale without chaos, commit to a 30-day operating plan that uses your templates, cadence, dashboard, and safety rules. The goal is steady throughput: publish consistently, follow up consistently, and learn weekly.

Week 1 (Stabilize): finalize your template library: one content brief, one landing page prompt, one social prompt, one 5-email sequence prompt, and one outreach prompt. Build your spreadsheet dashboard and record baseline metrics. Publish at least one email and two posts. Do outreach to 10 prospects using a single message angle.

Week 2 (Improve one bottleneck): choose the biggest gap (low clicks, low replies, low conversions). Run one small test. Example: two different CTAs on the same post format, or two different email subject styles. Keep everything else constant and log results.

Week 3 (Increase volume safely): add one more content batch per week (e.g., +2 posts) or increase outreach volume (e.g., from 10 to 20 prospects). Do not increase volume until your approval checklist is frictionless. If review time is slowing you down, tighten templates and reduce variation.

Week 4 (Systemize what worked): turn the winning elements into defaults: update the template with the best hook pattern, the best CTA, and the best structure. Document your “Do/Don’t” list (claims, tone, compliance notes). Plan the next month around the message angle that proved itself.

Practical outcomes after 30 days: you’ll have a repeatable weekly workflow, a small library of prompts that produce on-brand drafts, a basic dashboard that shows real movement, and a sales follow-up habit that turns content into conversations. That’s the point of AI in marketing and sales: not replacing you, but giving you leverage—safely and measurably.