1. Knowledge Base
  2. Features and Tools

AI Profiles: How Resquared adapts to you

AI Profiles are a personal, advanced system for optimizing your Resquared campaigns over time. This guide includes both a high level overview for the curious, and a deeper dive for technical teams.

TL;DR (for the busy)

  • That bar you see training up measures how well Resquared understands you: your tone, offers, audience, and what actually gets replies.

  • The fastest way to improve it is simple: send campaigns, try small variations, and accept/adjust the AI’s suggestions.

  • As it fills, your account becomes “tuned” to your market. Templates start pre-fitting your style, targeting/guidance sharpens, and results get more consistent.

  • We’ve back‑tested this on a large historical dataset across thousands of users. The system spots what moves open, reply, and lead rates—then adapts those patterns to your specific profile.

  • Bottom line: treat the first campaigns as the learning phase. The compounding personalization you gain becomes a lasting advantage.


Why AI Profiles exists

Resquared’s purpose is straightforward: get you leads. Most customers who win with outbound do two things right away:

  1. They ship early (send campaigns), and

  2. They iterate lightly (change small things and keep what works).

AI Profiles converts that winning behavior into a product experience. It aligns expectations around learning-by-doing, personalizes your recommendations, and—most importantly—helps you compound gains over time so your account becomes hard to beat (or leave).


What the Training Bar actually measures

The Training Bar is a personalization progress indicator. As you work, Resquared gathers signal from three places:

  1. Your campaigns – structure, tone, clarity of value, call‑to‑action, length, and more.

  2. Your targets – business types and audiences you pursue and how they respond.

  3. Your profile – the way your company sells (B2B vs B2C, deal size, sales motion, etc.) and how you interact with AI suggestions.

We don’t need you to label anything manually. The system infers patterns from outcomes and behavior and builds a profile that gets sharper with each send.

Key idea: The bar rises with useful signal. Activities that teach the system—like sending a full campaign, trying a variant, or tagging a positive outcome—move it fastest.


What you get as the bar fills

0–40%: Smart onboarding

  • Helpful defaults and baseline best practices.

  • Clear, low-effort suggestions to get you shipping quickly.

  • Reliant on global best pracitces (i.e. the types of campaign tactics that work for most people)

40–80%: Adaptive guidance

  • Recommendations start reflecting your market and the tones that work best.

  • You may notice larger variations in performance at times, due to the system testing wider ranges of tactics to see if less common tactics can lead to unique insights.

80–100%: Personalization flywheel

  • Templates and hints feel pre-tuned to you.

  • Suggestions get more specific; you’ll do fewer edits to get a strong campaign out the door.

  • The system anticipates your best performing “moves” and prioritizes what usually wins in your niche.


How it works (high‑level)

1) Data layer (your activity → structured signal)
Campaign performance and simple profile inputs become structured features. Think: tone, opening style, value clarity, type of ask, degree of personalization, and characteristics of the businesses you target.

2) Profiling layer (you vs. your segment)
We compare your patterns to segment‑level priors—learned from historical performance across similar users and audiences. This gives a starting point even if you’re new.

3) Recommendation engine (best next change)
For each campaign, the system weighs profile fit (what’s worked for you) and segment fit (what works for similar users) to suggest the highest‑leverage tweaks. It also attaches a confidence signal so you know how strongly the system believes a change will help.

4) Feedback loop (compounding improvements)
Your results feed back in. Wins are reinforced; misses are deprioritized. Over time, the model becomes more certain about your right combinations.

Conceptual view:
Lift ≈ Baseline × Profile Fit × Message Fit × Audience Fit.
As your bar grows, the “fit” terms become more accurate—and lift becomes more predictable.


What we’ve learned from the data (global patterns)

Across a large corpus of real campaigns (templates sent many times by hundreds of accounts), a few themes consistently matter. Directionally:

  • Opens respond most to clear value propositions and a well‑chosen type of ask, with audience alignment and the appeal you use also playing large roles.

  • Replies & leads depend heavily on who you’re aiming at (target category fit) and how you frame the benefit (appeal type), followed by sender positioning (how you present yourself) and tone.

  • On the user side, the strongest predictors tend to be primary sales channels (in‑person vs online vs phone), engagement with features (who experiments and tracks), revenue model/deal economics, and client type.

These aren’t one‑size‑fits‑all rules—the system localizes them to your context. But they inform the defaults you see on day one and the adjustments we recommend as your bar grows.

Illustrative outcome: In controlled tests, moving from a generic setup to a profile‑aligned setup commonly produced meaningful lifts (e.g., opens stepping up from the mid‑30s to around ~40% and replies per 200 emails ticking up accordingly). Your mileage will vary, but the direction is clear: fit beats generic.


Why rely on AI

  • Grounded in real outcomes. We learn from what recipients actually do—opens, replies, positives—not just theory.

  • Segment‑aware, not cookie‑cutter. The engine blends your own history with segment priors so even new accounts start smart, then quickly become yours.

  • Confidence‑guided. We surface how sure the system is about a suggestion, so you focus on high‑signal changes first.

  • Built for iteration, not guesswork. Tiny edits (subject framing, ask type, appeal) are suggested one at a time so you can move fast and see the impact.


How to get to 100% faster (and start feeling the “snap”)

  • Ship a complete first campaign. The system needs a full pass to get your initial fit.

  • Try one small variant per send. e.g., keep the tone, change the opening; or keep the opening, change the ask. The AI will do this automatically, so if you let it create the campaigns it will usually train faster.

  • Accept suggestions you agree with. Edits are welcome; your approvals still teach the model.

  • Tag positives. When somebody books, buys, or asks for a quote, make sure it’s marked as a lead in Resquared—you’re telling the system what “good” looks like. We also ID positive replies automatically.

  • Stay consistent for a few cycles. Consistency helps the engine lock in what’s repeatable for you. Send campaigns frequently, particularly in your first month. Once your training bar is nearly full, continue to send campaigns regularly to keep it trained up and adapting to market changes.

Pro tip: Quick, consistent action beats long deliberation. The bar measures useful learning, not reading time.


What changes you’ll actually see (examples)

  • Subject & opening guidance aligned to your audience (e.g., “question‑style” vs “direct offer”).

  • Value framing nudges (ROI, convenience, social proof, community, etc.).

  • Ask type tuning (request a reply vs schedule time vs send info) matched to your typical sales motion.

  • Tone & positioning that fits your brand (friendly local peer vs professional advisor, concise vs consultative).

  • Campaign shaping (where a message sits in the cadence, when to follow up) to maintain momentum without fatigue.

These are representative examples. The exact signals evolve as your profile trains.


For the curious: methodological notes (light technical)

  • We analyze outcomes both email‑weighted (what works across lots of sends) and user‑balanced (what’s robust across different accounts).

  • Signals like value clarity, type of ask, target type, appeal, and sender positioning repeatedly emerge as high‑leverage, especially when adapted to segment context.

  • We combine historical patterns with on‑the‑fly learning from your own results, and we attach confidence to recommendations so exploration feels purposeful rather than random.


Frequently asked questions

Does this replace my judgment?
No. It accelerates it. You still approve, edit, or skip any suggestion.

Will it write in my voice?
Yes—your edits teach the system. As your bar grows, drafts arrive closer to “your way,” needing fewer changes. Though keep in mind, “your way” that works best may be different than the approach you were likely to take before starting with Resquared. 

What if I’m new to outbound?
Perfect. Send your first campaign; the defaults are informed by what tends to work for similar users and audiences. You’ll get early wins and clearer next steps.

What if I’m experienced?
You’ll move quickly into the high‑confidence zone. The system will prioritize nuanced tweaks that stack on your existing strengths.

Can I see why it’s suggesting something?
All AI features in the product are clearly labelled. When you use them, it is always incorporating your AI profile and learning tactics into the final results. 


The moment it clicks

A well‑trained profile feels like this: you open Resquared, start a campaign, and what appears is fine tuned to you, your business, and your target client—immediately. The suggested target framing makes sense. The subject line “lands.” You spend your time refining, not reinventing. That’s the Training Bar doing its job.

Send your next campaign. The sooner you ship, the sooner your account becomes unmistakably yours.