SUCCESS STORY
See how Contraktor scaled support in a multi-product operation, reduced the backlog accumulated outside business hours, and freed the team to focus on what really required human intervention.
Metric | Result |
|---|---|
Tickets resolved with AI | 75% |
First response time | 15s |
Chat backlog in the morning | 0 |
About Contraktor
Contraktor operates in the contract management market and has been expanding its operation with different products within this ecosystem.
As a result, the company needs to serve both direct customers and end users who interact with the platform to sign contracts and complete steps in the process.
In practice, this increases the volume of questions, the variety of support tickets, and the pressure for quick, accurate responses. Scaling support without losing efficiency has become an increasingly clear necessity.
The main bottlenecks of the operation at that time
With the operation evolving, Contraktor began to face a common challenge in SaaS companies with a broader portfolio: sustaining speed and quality in support without turning support into an increasingly heavy structure.
Title | Description |
|---|---|
High volume of repetitive tickets | A large part of the operation was consumed by recurring questions, which overloaded support and took time away from cases that required more context and analysis. |
Multiple products, more complexity | With different product lines in operation, the team had to deal with distinct journeys, workflows, and knowledge bases within the same support interaction. |
Pressure for speed without losing quality | It was necessary to respond quickly in a sensitive context, without compromising clarity and security in interactions related to contracts and subscriptions. |
Scale without inflating the structure | As the operation grew, it became clearer that expanding the team at the same pace would not be the most efficient or sustainable path. |
Contraktor also tested some tools to try to scale support. First, Intercom, which offered robustness, but the billing model in dollars and per retained ticket was too heavy for a high-volume operation.
Then, the company tried Crisp, which was more affordable, but still did not deliver the depth and retention needed in AI support. In practice, the challenge remained unresolved.
“With the evolution of the operation, we started to feel an ever-growing complexity in the support area.”
Rafael Salomão
Head of Customer Success at Constraktor
ClaudIA as the new scaling layer for support
Contraktor’s response was to structure the initial support experience with ClaudIA, using AI to absorb N1, handle recurring questions, and make ticket intake smoother.
Main changes:
ClaudIA began taking the first contact with the customer, reducing the burden of manual support right at the start of the operation.
While the AI handled recurring questions and initial triage, the human team was able to focus more on cases that required analysis, context, and greater depth.
The operation no longer depended as much on dividing the team’s attention across different platforms, resulting in more consistent support.
The Help Center and the existing workflows began feeding the AI, enabling faster responses that were more aligned with the operation’s context.
The operation became more balanced, faster, and a more efficient way to scale without increasing the team structure at the same pace
“While we’re giving attention to a more serious case, ClaudIA is able to handle the support right from the start.”
Alana S. Souza
Senior Support Analyst at Constraktor
Fast implementation, continuous evolution
The implementation of ClaudIA happened in a simpler way than the team had imagined at the beginning. Instead of requiring a parallel structure or major technical mobilization, the project relied on a foundation that Contraktor already had: its support content, workflows, and the practical experience of its operations.
Step | Title | Description |
|---|---|---|
01 | Initial base built from real operations | The Help Center, existing workflows, and the team's knowledge served as the starting point for structuring ClaudIA's responses and behavior. |
02 | Fast entry with validated rollout | In about a week, the AI was already live at Constraktor Sign. Based on what was learned in this stage, Constraktor expanded ClaudIA to the other support lines. |
03 | Calibration in the first few weeks | After go-live, the team monitored conversations, reviewed responses, and refined ClaudIA's performance to improve relevance, quality, and retention. |
04 | Evolution with close support and auditing | The implementation did not end at setup. The maturity gain came from continuous follow-up, audits, and the team's dedication over the first few months. |
More than just an implementation project, ClaudIA's rollout was treated as an operational evolution process. The setup was quick, but the results came from the combination of a well-structured base, constant calibration, and the direct involvement of people who knew the support routine.
With close support from Cloud Humans and the active participation of the Contraktor team, the operation matured without interrupting day-to-day service.
“The implementation was also very smooth. I can say I had a lot of support from the entire Cloud Humans team from the beginning.”
Alana S. Souza
Senior Support Analyst at Constraktor
The results of ClaudIA in the operation
With ClaudIA handling the initial support, Contraktor began to see clear gains in speed, resolution, and operational efficiency. The results showed up both in the most visible indicators, such as first response time and AI retention, and in the team’s day-to-day routine, which became less overloaded and had more room to operate strategically.
Metric | Result | Impact |
|---|---|---|
Tickets resolved by AI | 75% | More support capacity without increasing the structure at the same rate |
Backlog at the start of the day | 30 chats → 0 | Less queued backlog and a lighter start to operations |
First response time | 40 min → 15s | Almost immediate response right as support begins |
Beyond the results shown in the table, the change also became clear when compared to the operation’s earlier attempts. Before Cloud Humans, Contraktor had already used Intercom and Crisp, but had still not found a truly efficient combination of cost, AI quality, and retention capacity in support. In the case of Crisp, for example, the operation did not reach 15% of support via bot.
The gain also appeared in the way the operation became less dependent on the technology team in some routines. With the evolution of the implementation and the use of backoffice integrations, Contraktor gained more autonomy to resolve some demands that previously required direct technical support.
“We were able to use her time for things that were truly more strategic.”
Rafael Salomão
Head of Customer Success at Constraktor
What Contraktor learned in the process
Topic | Description |
|---|---|
Good implementation depends on the tool and a committed team | The result didn't come from technology alone. The combination of a good solution and an engaged team was essential to make the implementation truly work. |
Automation doesn't replace the team, it frees the team | With ClaudIA absorbing part of the initial support, the human team gained more space to focus on what required analysis, context, and strategic action. |
Becoming more efficient in N1 changes the entire operation | When initial support becomes faster and more efficient, the impact goes beyond the queue: the operation becomes smoother, the team works better, and the customer feels the difference. |
Continuous tuning helps the AI evolve with the operation | Going live quickly was important, but the best results came from constant monitoring, reviewing conversations, and adjustments made in the first few weeks. |
“A good tool and a committed team make for a good implementation. You can't expect miracles from the vendor alone.”
Rafael Salomão
Head of Customer Success at Constraktor
Cloud Humans answers
What was Contraktor's main challenge before ClaudIA?
Contraktor needed to maintain speed and quality in support across an operation with multiple products, repetitive tickets, and an increasingly complex knowledge base, without inflating the structure at the same pace.
How did ClaudIA begin to operate in the workflow?
ClaudIA took over initial support, absorbing N1, handling recurring questions, and routing to the human team the cases that required more analysis, context, or depth.
Was the implementation complex?
No. The implementation went live quickly and relied on a foundation Contraktor already had, such as support content and operational workflows. The maturity gains came from continuous tuning in the following weeks.
How soon did the operation start seeing results?
The AI was already live for testing in about a week. After that, the adjustments and audits in the first few weeks helped accelerate the operation's growth curve.
Did ClaudIA only solve simple questions?
No. In addition to absorbing initial support and recurring questions, the operation also evolved with integrations and flows that expanded the AI's capabilities within support.
