Can Tidio’s Lyro AI Agent Handle the First Conversation? A Hands-On Test for Service Businesses

A website visitor can be interested in hiring you and still leave without making contact.

Sometimes the reason is simple. They cannot tell which service fits, they have a question after business hours, or they do not want to fill out a form before understanding what your company does. For a service business, that first unanswered question can quietly become a lost lead.

We wanted to see whether an AI agent could help with this early part of the conversation. So we set up Tidio’s Lyro AI Agent using information from the WD Strategies website and a small set of manually approved answers.

Disclosure: WD Strategies may receive compensation from Tidio if this article is approved under its content campaign. This content contains affiliate links. We may earn a commission if you sign up through our link, at no extra cost to you. Tidio did not review or control our hands-on observations.

This was a focused product demonstration, not a long-term deployment. We did not install the Tidio widget on the live WD Strategies website, connect email or social channels, or expose Lyro to real visitors. The goal was narrower: test how quickly a service business could give Lyro useful information, then see how it answered a few common questions in Tidio’s Playground.

The result was promising, with a few important limits.

Why the first website conversation matters

Service businesses often ask visitors to make several decisions before speaking with anyone. A potential client may need to choose between website development, digital marketing, branding, consulting, or another service they only partly understand.

The website can explain each option, but visitors do not always read it in order. They arrive on a project page, skim a service description, then wonder whether the company can help with their particular situation.

A traditional contact form captures people who are ready to explain the project. Live chat works when someone is available to respond. An AI agent covers the space between those 2 moments by answering common questions immediately and helping the visitor find a sensible next step.

That job sounds modest. It is also where many chatbots become annoying. A scripted bot may force the visitor through a menu that does not match the question. A useful AI agent needs to understand natural wording, stay within approved business information, and know when the conversation needs a person.

What Tidio and Lyro do

Tidio combines live chat, ticketing, automation, and AI-supported customer service. Lyro is its conversational AI agent. Businesses can add knowledge through website pages, manual question-and-answer pairs, CSV or PDF files, Zendesk content, and other supported sources.

Lyro uses that material to answer customer questions. Tidio also provides Guidance settings for communication style and instructions, a Playground for testing answers before launch, and handoff controls for cases that need a human agent.

Tidio reports that Lyro resolves 67% of customer conversations on average. That is Tidio’s platform-wide figure, not a result from our demonstration. Actual performance will depend on the quality of the source material, the questions visitors ask, and the rules used for escalation.

For a service business, the appeal is easy to understand. Questions about services, process, contact details, industries served, or the next step are usually repeatable. A more specific request involving budget, strategy, timing, or technical requirements still belongs with the team.

How we set up the test

The initial setup started with public pages from the WD Strategies website. Lyro scanned those pages and converted the available information into knowledge it could use in its answers.

We then added 5 manual Q&As covering:

•      The services WD Strategies provides

•      Website improvement work

•      The types of businesses the company works with

•      How to contact the team about a project

•      What a visitor should do when they are unsure which service they need

The manual entries gave us direct control over the answer. They also helped keep replies short enough for a visitor who wants direction, not a full service-page rewrite.

In our hands-on test, Lyro combined manually approved Q&As with information collected from selected WD Strategies website pages.

This mix of website content and manual answers is practical. The website gives Lyro breadth, while approved Q&As can tighten the wording around high-value questions. The Data sources screen also shows which entries came from the website and which were added manually, making it easier to review or update them.

After saving the knowledge, we used the Playground to test questions without publishing the agent on the website. Tidio recommends this step before launch so teams can find missing, irrelevant, or inaccurate answers.

What happened when we tested Lyro

1. It gave a clear overview of the services

We asked, “What services does WD Strategies provide?”

Lyro returned a structured list that included website design and development, software development, digital marketing, systems integration, multimedia production, branding, and business consulting. It then invited the visitor to describe their goals if they were unsure which service fit.

Using approved business information, Lyro summarized the services available to a website visitor.

The response was useful because it did more than repeat a page title. It gave the visitor a quick view of the available services and ended with a reasonable path forward.

It was also longer than a one-line answer. That is where testing matters. A team may decide that the extra context is helpful, or shorten the approved Q&A if most visitors use the chat window on a phone.

2. It handled uncertainty without inventing a recommendation

The stronger test was: “What if I’m not sure which service I need?”

Lyro told the visitor to describe their goals or current challenges so the team could help identify the right service or next step. It did not assign a service based on missing information.

When a visitor was unsure where to begin, Lyro guided them toward the team for more specific help.

That restraint matters for consulting and agency work. A website redesign request may turn out to be a positioning problem. A request for marketing may depend on analytics, content, or technical work. Our guide to choosing a digital marketing agency makes a related point: the process starts with getting clear on what the business needs and asking the right questions. The first conversation should reduce confusion without pretending that discovery has already happened.

Our screenshot also shows that the response was based on the connected data sources. This gives the team a useful review trail when testing how Lyro arrived at an answer.

3. The quality of the answer depended on the knowledge we supplied

Lyro worked well with questions already covered by the site or manual Q&As. It was less meaningful to test vague or highly specific project questions because the setup did not include private discovery notes, proposals, pricing rules, or client information.

That is the right boundary for this kind of test. Public business information can support common first-contact questions. Recommendations involving scope, cost, feasibility, or commitments need approved rules and often require a person.

Where the human team still belongs

Tidio’s handoff settings can transfer a conversation to an agent, keep it with Lyro, or create a ticket when Lyro lacks the answer or a visitor asks for a person. We reviewed the available handoff controls, but we did not test a live transfer because no chat widget or support channel was connected.

For a service business, human involvement is still essential when the conversation reaches:

•      Project scope and technical requirements

•      Budget, estimates, or contractual questions

•      Strategy that depends on business context

•      Urgent or sensitive concerns

•      A visitor who directly asks to speak with someone

This follows a broader pattern in generative AI and business operations, where AI handles repeatable work and people provide the judgment needed for more complex decisions.

Lyro can prepare that conversation by answering routine questions and pointing the visitor toward the team. It should not make commitments the business has not approved.

Current Tidio and Lyro pricing

Tidio’s public pricing page was checked on August 12, 2026. The figures below reflect the annual-billing prices displayed at the time:

•      Free: $0 per month, with 50 billable conversations and the first 50 Lyro AI conversations available as a one-time allowance.

•      Starter: $24.17 per month, with 100 billable conversations, live chat, ticketing, and basic analytics.

•      Growth: Starts at $49.17 per month, from 250 billable conversations, with added team and analytics features.

•      Plus: Starts at $300 per month plus monthly usage, with custom limits and added support and management features.

•      Premium: Custom pricing, with managed AI services and a guaranteed 50% Lyro resolution rate among the listed features.

Lyro is also listed as a stand-alone product or add-on starting at $32.50 per month for 50 AI conversations. Tidio’s AI Agent page separately promotes pricing at $0.50 per conversation. Because the entry package and per-conversation figure do not produce the same effective rate at every volume, check the live pricing calculator and confirm the quota, billing term, and included features before making a cost comparison.

Pricing can change. Review Tidio’s current plans through our partner link before subscribing.

Who should consider Tidio

Based on this test, Tidio is worth considering for a service business that receives recurring website questions and already has clear service pages, FAQs, or other source material. It is especially relevant when slow first replies are creating a gap between website interest and a real sales conversation.

It may be a weaker fit for a business that depends mainly on phone support, has very little documented information, or needs the AI to make complex recommendations without human review. The agent can only work with the knowledge and instructions it receives.

There is also ongoing work after setup. Services change. Contact details move. A new offer may need its own Q&A. Someone should review conversations, correct weak answers, and keep the knowledge current.

A sensible way to test it

Start with 10 to 20 questions that real prospects ask. Add the relevant website pages, then write short manual answers for questions where wording or accuracy matters most.

Use the Playground before publishing. Try normal questions, vague questions, misspellings, and requests that should reach a person. Check the answer itself, the tone, and the next step it gives the visitor.

Our demonstration showed that Lyro could learn from public service-business content, answer a basic service question, and guide an unsure visitor toward the team. It did not prove the 67% average resolution claim, measure lead conversion, or test a live handoff. Those results require real traffic and a longer evaluation.

If your website receives questions that could be answered from information you already have, try Tidio and Lyro AI through our partner link. Give it a small, realistic test set first.

The useful question is whether it helps more visitors reach the right next step without creating extra cleanup for your team.