Streamlining BPO Onboarding with Research-Led Automation

context

SquadStack is a tele-sales marketplace that connects businesses with trained tele-callers. For companies, the platform offers access to a reliable supply of callers to scale sales operations. For tele-callers, it provides consistent remote earning opportunities and professional development through structured training and performance support. I was part of the User Product Team, focused on improving the experience of supply-side users (tele-callers).

We were testing out the new BPO model wherein we connected businesses with distributed BPOs across India. Most of these BPOs were large centers spread across tier-3 and tier-4 cities, employing hundreds of tele-callers. For the model to succeed, SquadStack needed to ensure not just acquisition of BPO partners, but also a seamless system to onboard new tele-callers consistently and at scale.

At this stage, onboarding was only partially productized. Much of it still relied on manual interventions by center managers, who were responsible for verifying KYC, conducting interviews, updating tele-caller information, and tracking progress. Training and onboarding were handled in fragmented, inconsistent ways, making the process heavy on operations and light on user experience.

Because onboarding was structured use-case by use-case (rather than a one-time setup), it became a recurring and critical touchpoint for tele-callers and managers. This meant that any friction here could directly affect adoption, retention, and client fulfillment.

problem

The manual, fragmented workflows created bottlenecks for all three stakeholders in the system:

• For tele-callers (end users): The experience was slow and inconsistent. Many dropped off mid-way or switched to competing BPOs where the process was smoother and faster. This eroded motivation and impacted the quality of tele-callers entering the system.

• For center managers (internal users): Onboarding new tele-callers consumed disproportionate time. Instead of focusing on improving performance and supporting sales, managers were stuck in repetitive validation tasks like KYC checks, interviews, and manual data updates.

• For SquadStack (business): Acquisition and fulfillment were directly at risk. Without efficient and consistent onboarding, SquadStack couldn’t reliably meet client demand. The lack of standardization also raised compliance risks, while manager inefficiency translated into higher operational costs.

  
  

In short: the system was too fragile to scale. To make the model sustainable, onboarding needed to be redesigned as a productized, streamlined, and user-friendly process.

goals

The project set out to make onboarding:

• Faster and smoother for tele-callers, ensuring motivation and adoption were not lost.

• Less manual for center managers, freeing their bandwidth for performance-related work.

• Consistent and compliant for SquadStack, enabling reliable fulfillment of client requirements.

  
  

To achieve this, we broke the problem into two parallel objectives:

• Fulfillment Matching: ensuring the right tele-callers were onboarded for the right use cases, so client demand could be met reliably.

• Operational Efficiency & Cost Savings: reducing dependency on center managers for acquisition-related tasks by automating repetitive workflows, allowing them to shift their focus to performance and sales support.

  
  

While both objectives were critical, this case study will focus on the second objective, Operational Efficiency & Cost Savings, and the systems we designed to achieve it. But it would be helpful to note that each of these objectives were co-dependent on each other for the success of the entire system.

success criteria

Primary Metrics:

• Reduction in center manager time spent on acquisition-related tasks (target: 50–60%).

• Reduction in cost per acquisition through role redefinition and automation.

• Reduced tele-caller drop-off during onboarding (target: <20% across all stages).

• Faster onboarding time (target: from 3 weeks to 1.5 weeks).

  
  

Qualitative Metrics:

• Smoother, more motivating onboarding experience for tele-callers.

• Clear and consistent workflows for center managers.

• Higher client confidence in fulfillment and compliance quality.

  
  

Guardrails:

• Maintain compliance and vetting standards despite automation.

• Ensure scalability without degrading user experience.

approach

We approached the problem in two dimensions:

• User Understanding: We engaged with both end users (tele-callers) and internal users (center managers) to uncover frustrations and inefficiencies in the onboarding process. This included mapping user journeys, observing sessions, and interviewing stakeholders to highlight pain points such as tele-caller drop-offs, inconsistent vetting, and center manager bandwidth strain.

• System Thinking: We broke down the existing workflows into discrete steps, documented the time and effort required for each, and prioritized which could be automated or productized without compromising compliance.

  
  

Rather than removing center managers from onboarding, the strategy was to redefine their role: automate repetitive acquisition-related tasks, while enabling center managers to focus on higher-value activities such as performance and sales coaching.

This would be ideally done by focusing on phased automation to de-risk implementation and immediately return time and cost savings, while ensuring user experience improved at each step.

process

We followed a phased, research-driven approach rooted in the double diamond framework; starting broad to understand users, then narrowing into solutions that balanced automation with human oversight.

Phase 1: Discovery & Mapping

We began by immersing ourselves in the environment of our internal users (center managers) and external users (tele-callers).

Using contextual inquiry and shadowing, we observed how center managers spent their day across onboarding cycles—shadowing them through tasks such as KYC validation, conducting interviews, updating partner details, and maintaining spreadsheets. We also conducted in-depth interviews with all active center managers(7) to capture perceived pain points and cross-validated through direct observation.

On the tele-caller side, we spoke with 5–7 recently onboarded tele-callers to uncover their pain points, particularly the reasons behind drop-offs. To strengthen our perspective, we supplemented this with backend data on drop-offs at multiple onboarding stages, which helped us tie subjective feedback to hard evidence. In parallel, a comparative analysis of onboarding practices at other BPOs contextualized why our process felt disproportionately friction-heavy for tele-callers.

Through thematic analysis of this combined data, we surfaced two critical insights:

• The most time-intensive center manager tasks (e.g., repeated KYC checks, fragmented status updates, manual test facilitation) were also the points where tele-callers most often abandoned the process.

• The operational burden on center managers was largely invisible to higher-level stakeholders, making it harder to justify resourcing until we quantified the effort in hours lost.

Phase 2: Defining Needs & Design

From this analysis, we reframed the challenge as a dual-user problem:

• Tele-callers needed a faster, smoother experience with minimal uncertainty about next steps.

• Center managers needed bandwidth relief so they could focus on higher-value performance coaching rather than repetitive operational work.

This reframing led us to our core design principle: automate wherever rule-based decisions existed, but preserve human oversight in nuanced, judgment-heavy steps (such as case-specific sales interviews).

To move from problem to design direction, we mapped all center manager tasks and layered in three lenses:

• Time intensity for center managers; validated through shadowing and time-tracking to quantify weekly hours per task.

• Impact on tele-caller drop-offs; informed by tele-caller interviews that revealed friction points leading to attrition.

• Feasibility of automation; assessed by classifying tasks into fully automatable, partially automatable, and human-only.

  
  

We used an Effort-Impact Matrix to prioritize which tasks to automate first, balancing ROI for center managers with measurable improvements in tele-caller experience. This process was co-defined with center managers themselves and validated by their leadership, ensuring alignment across stakeholders before moving ahead.

Phase 3: Co-creation & Iterative Pilots

We broke workflows into atomic tasks and evaluated them against two criteria: (1) potential for rule-based automation, and (2) impact on user experience. To keep solutions grounded, we co-created automation concepts with both center managers and tele-callers through design workshops.

Each automation or process change was tested in small-scale pilots (10–15 candidates at a time). By rolling out one change at a time, we reduced risk and isolated the impact of each intervention.

Success was measured against previously defined metrics:

• Center manager time saved per task (benchmark vs. automated).

• Tele-caller NPS and drop-offs, with reasons for discontinuation tracked.

• System reliability and adoption, especially among center managers resistant to change.

• Qualitative feedback via interviews, usability checks, and live workshops after each pilot.

  
  

Through this iterative cycle, we surfaced key learnings:

• Partial automation was more effective than full automation in judgment-heavy areas.

• Tele-caller NPS and completion rates rose when onboarding steps were consolidated and made more transparent.

• Adoption resistance clustered among long-tenured center managers, requiring targeted training and reinforcement.

  
  

This approach created a live feedback loop; refining workflows with every cycle while de-risking implementation and capturing early time and cost savings.

Phase 4: Scaling & Integration

With a lot of center manager workflows automated, center managers transitioned from execution-heavy roles to oversight and exception-handling. This freed significant bandwidth for performance coaching while improving tele-caller experience.

To institutionalize adoption, we:

• Conducted training workshops and created internal documentation for center managers.

• Shared research readouts with stakeholders at each milestone, making both user and business impact visible.

• Embedded continuous feedback loops with center managers to evolve workflows.

• Established ongoing monitoring to ensure balance between efficiency and user experience.

The end result was a productized onboarding process—consistent, scalable, and far less dependent on manual intervention, without compromising the human touchpoints that mattered most.

insights & impact

The new onboarding system drove both measurable and qualitative impact across users and the business:

• 68% reduction in center manager time spent on acquisition-related tasks.

• ₹3M in cost savings, largely by shifting center manager responsibilities away from repetitive acquisition work.

• Tele-caller drop-off rates improved significantly—from 30% down to 10%—while NPS scores held steady, indicating that automation did not compromise the tele-caller experience.

• Hiring cycle reduced from ~3 weeks to ~1 week.

  
  

Qualitatively, Center Managers reported feeling more visible, valued, and able to focus on the higher-value coaching responsibilities they were actually hired for. tele-callers experienced smoother onboarding indirectly reflected in lower drop-offs, even if they lacked a before/after comparison. At the leadership level, confidence in scaling operations grew considerably, as fulfillment and compliance became more predictable and less dependent on operational bottlenecks.

Together, these results delivered on the three core goals: cost savings, reducing drop-offs, and reducing center manager dependency, without compromising user experience.

reflections & takeaways

This project surfaced key lessons in balancing automation with human oversight:

• Partial automation > full automation: Not every process benefits from end-to-end automation. The real value often lies in freeing humans from repetitive work so they can shine in high-judgment tasks.

• Dual-user perspective is critical: Shadowing revealed invisible burdens for center managers, while comparative analysis and tele-caller interviews contextualized tele-caller struggles. Looking at both groups together unlocked the right direction for solutions.

• Scaling requires finishing the “smaller” tasks too: While high-impact automations solved the immediate problem, some lower-priority operational fixes were left behind. In hindsight, addressing these earlier could have prevented friction at larger scale.

For me as a PM, the biggest learning was how phased automation and prioritization frameworks can de-risk implementation while still delivering immediate ROI. From a UXR lens, this reinforced the value of embedding with users, surfacing unseen struggles, and anchoring design in dual-user needs.