Research & strategy

Linked Helper
/ Tandraft

How customer language became a product strategy and a landing page for controlled personalization.

Independent assignment concept B2B SaaS · AI 2026
Tandraft landing-page prototype with personalization preview and human review controls
Tandraft is my concept name for the assignment, not an official Linked Helper rebrand. The prototype uses fictional examples.

The design problem

Linked Helper combines AI messages, custom variables and conditional logic. The challenge was to explain that depth without making the first interaction feel complex.

I worked from the assignment and supplied research materials to develop the positioning, information architecture, content and responsive landing-page prototype. The central question: how can people scale outreach while retaining control of each message?

My role
Research synthesis, product strategy, UX architecture, content design, UI and prototyping.
Inputs
Assignment brief, product documentation, review dataset, search snapshot and competitor sources.
Deliverables
Research presentation, messaging framework and interactive EN/RU landing page.
Status
Independent concept. No client launch or measured business impact is claimed.

What the evidence showed

The detailed thematic review covers 768 G2 reviews: Linked Helper (142), Waalaxy (528) and Expandi (98). The broader competitor comparison uses a different cohort of 1,445 reviews across seven products.

Ease and clarity · 343/768
44.7%
Time saved · 132/768
17.2%
Personalization · 92/768
12.0%
Complexity and learning · 72/768
9.4%

Share of the 768-review cohort. One review can match several themes. This is secondary research using a supplied dataset, not interviews, a representative market survey or a causal study. The two cohorts and their coding are not interchangeable.

Method, sources and limitations

From evidence to design

Make control the value proposition
The reviews describe both useful flexibility and a learning curve.
Design decision: Position the concept around personalization people can control, rather than an inventory of AI features.
Explain the mechanism through decisions
A user needs to know which context is used, what happens when it is missing and who checks the result.
Design decision: Structure the demonstration around data, rules, a draft and human review. Show a fallback example alongside the rule.
Place reassurance before the next commitment
The first campaign and the activation of a trial license are different steps.
Design decision: Make the trial path explicit, explain what happens after the CTA and define a separate first-campaign success event.

What was built

An interactive landing-page concept and a complete, linked research presentation. The prototype lets visitors compare personalization modes and example recipients, inspect fallback logic and see the review step.

My product-design contribution is the traceable connection between evidence, positioning, page structure and interface behavior. The result is a testable design proposal; higher conversion or reply rates have not been demonstrated.

What I would validate next

Before calling the concept successful, I would test the assumptions it depends on:

  • Comprehension: can a new visitor explain the personalization mechanism and the human-review step?
  • Usability: can people choose context, understand a missing-data fallback and locate draft review?
  • Activation: track the path from landing-page visit to trial start and first campaign, with clear event definitions.
  • Business impact: measure against a baseline; use a controlled experiment where traffic permits.

Explore the reasoning

All 12 stages
Next: Advance Education — product design

Let’s make the next
decision clearer.

Tell me about the product, the challenge and where you are in the process.

sonaturum@gmail.com
Or prepare a short project brief

Opens a draft in your email app. You review and send it yourself.