Using AI to transform everyday work at Coherent Solutions

AI results at a glance

• Since early 2025, over 600 employees (more than 30% of our total workforce) from technical and non-technical roles have completed our internal AI training.

• AI-driven automated training provides feedback in hours instead of days. Technical trainers spend less time repeating reviews.

• HR team members conducting retention and exit interviews can produce structured documentation in minutes rather than an hour.

• Sales preparation, which once required extensive internal coordination, now benefits from faster access to verified information through Coherent’s AI platform, the Searchable Platform for Assets, Resources, and Knowledge (SPARK).

• Recruitment teams use AI to standardize job descriptions, help candidates prepare for interviews, and automate routine tasks, while keeping communication personal.

Our AI transformation approach

At Coherent Solutions, we didn’t approach AI use in our organization with a big platform launch or a top-down order. Instead, we began with a clear idea of how artificial intelligence should work to support our business goals and deliver measurable value across teams.

Our focus was simple: AI needed to be integrated into actual workflows and have clear boundaries to promote safe, responsible use.

We also realized that AI couldn’t be applied indiscriminately, nor could everyone use the same tools. Each team was encouraged to use approved AI tools where it made sense for their work, supporting their team-specific goals, as long as they followed the AI guidelines.

Identifying the problems AI could solve

Our goals were to use the tools to cut down on low-level tasks and rework, while preserving important context, and supporting human judgment. Individual teams began by identifying the most significant problem areas where AI could be useful. Project documentation and preparation took too long and often needed to be redone. Reviews depended on who was available. Important knowledge was scattered across emails, folders, and individual team members. When someone changed roles, crucial knowledge was often lost. We brought AI into these situations on purpose, not to make decisions, but to help keep important context from being lost.

From the beginning, we set clear rules for data, permissions, and oversight. AI outputs had to come from trusted sources and be easy for people to understand.

This wasn’t about launching a new framework. It was integrating AI into daily work to improve efficiency.

How AI changed daily work

When AI became part of our teams’ daily workflows, adoption approaches weren’t uniform— which we expected.

After all, different teams picked different tools. Some adopted them quickly, while others took more time. What mattered most were the results, not the individual adoption approaches.

Gradually, we noticed teams using AI to prepare and review work. Tasks that used to stall due to a lack of context proceeded faster with automated information from previous projects. Team members were able to make decisions on new projects with a clearer understanding of what occurred in previous projects. And with guidelines for data use in our AI tools, they could trust the information they received and focus on what mattered most.

AI as human support, not replacement

Across all teams, we noticed the following: AI was used as a support, not a replacement for their effort and expertise. People still made decisions, but context was retained across projects, and manual effort decreased as teams automated basic tasks.

Ultimately, team members came to trust the AI tools because they helped them do their jobs better, not just automate tasks.

At Coherent, we call this approach “AI as fabric.” Ideally, AI should be built into everyday workflows rather than added as a separate layer at the beginning or end of a process. Building AI into workflows allows for learning, context, and execution to build over time. Different teams may use different tools, but this approach keeps the focus on continuity and reliability.

How AI shows up across our organization

Sales and Marketing preparation

For Sales and Marketing, AI didn’t change accountability, but it dramatically improved how teams got ready.

Before AI, preparing for a client meeting meant piecing together information across sources. Teams searched shared folders, reviewed old presentations, and checked with delivery managers to confirm which projects they could reference.

The information was there but finding it took considerable time and coordination.

Once AI was added to the preparation process, research time shortened, context was available sooner, and teams could focus more on engaging with new clients.

As AI reduced friction in the Sales and Marketing workflows, preparing proposals and presentations went from hours and days to mere minutes. Sales teams were able to spend less time verifying information and more time building relationships with potential clients.

AI didn’t replace human judgment or creativity. Instead, it reduced uncertainty and helped teams move faster and with more confidence.

Recruitment and talent selection

In recruitment, we decided to use AI only where it made sense for our organization’s needs.

Our goal wasn’t to automate the entire hiring process. Instead, we wanted to strategically use AI for the routine tasks that took time away from talking to and assessing candidates. These tasks included writing job descriptions, preparing CVs for client interviews, planning, analytics, and market research.

Automating these tasks helped the hiring team expand its recruitment efforts and evaluate more candidates.

For this team, we set clear boundaries: Candidates would continue to communicate with real people, not an anonymous chatbot.

AI helps us scale and stay consistent, but the candidate experience is shaped by their interactions with our team.

HR and people operations

For HR teams, we focused AI adoption on the processes that took the most time.

Retention interviews were important, but they often created hours of extra manual work, including transcription, translation, organizing notes, and preparing summaries.

The team began using tools to automate these tasks. Now, interviews can be summarized and organized quickly in a consistent format. This reduces the burden of post-interview work, freeing team members to focus on other tasks.

We also noticed repeating patterns in day-to-day HR requests. Employees often asked the same questions about policies, processes, and internal documentation. The answers were already available, but many employees didn’t know where to find them. People either searched across systems or asked HR directly, which created a steady stream of interruptions. To address this, the HR team decided to set up an internal Knowledge Assistant, giving employees a fast, consistent way to find answers on their own, without needing to message HR for routine information.

Training and enablement

Training was one of the first areas where our organization used AI to change both our operating model and our workflows. The shift began in summer 2024, when a trainer built an automated grading system to handle repetitive code reviews. Now, we use AI agents to evaluate submissions against course rubrics and return structured feedback every three hours instead of days, similar to how we’ve applied AI-driven QA automation in delivery environments.

Today, 80%-90% of our initial evaluations are handled by AI. Trainers focus on resolving edge cases and refining instructions. As one trainer put it, "When you do the same thing repeatedly, you naturally simplify your feedback or rush through it. AI doesn't get tired." The system applies the same standard every time, allowing for greater consistency and accuracy.

AI training programs and knowledge sharing

Our team reinvested the time saved from grading into building a structured AI training program. In the first fourteen months, 878 employees across 28 practices and 11 offices enrolled in the courses. So far, roughly one-third of the company has participated, including those in non-technical roles.

During this time, the training team also began applying AI to knowledge curation—our processes for maintaining and sharing engineering data. The team used a multi-agent system to help generate practice-level educational digests, transforming a manual process into a fifteen-minute task.

Overall, we found that the embedded agentic AI absorbed the operational load, allowing the training team to scale its reach without compromising quality.

The impact of our internal AI projects

The impact of our internal AI initiatives became measurable in the way everyday work shifted across the organization.

Human Resources

In HR, they saw a reduction in follow-up work. Retention and exit interviews utilized AI to automatically transcribe, translate, and summarize information. Once conversations were captured and organized, that work could be quickly prepped and shared.

Training

In training, feedback no longer depended on when a trainer could address their project backlog. With AI, feedback was delivered almost instantly, and evaluation rules and standards stayed the same for all submissions. With these guidelines in place, trainers didn’t have to repeat the same judgment for every learner.

Sales and business development

In sales preparation, teams stopped rebuilding context for every new opportunity. Case studies, outcomes, and supporting materials were readily available in a standardized format, so there was no need to confirm details with delivery teams.

In all these areas, our teams saw real, practical changes in their efforts and productivity.

These changes weren’t planned — they happened organically as AI became part of daily work.

SPARK: Coherent’s embedded AI solution

Across teams, AI is applied differently depending on role and workflow. SPARK is one of the clearest examples of how that shift took shape.

What began as a focused initiative within Sales and Marketing evolved into a company-wide platform embedded in daily work.

SPARK stands for Searchable Platform for Assets, Resources, and Knowledge. It is an internal AI-powered system designed to make Coherent Solutions’ institutional knowledge structured, accessible, and usable in real time. By connecting client history, delivery context, staffing insight, and marketing assets, SPARK turns fragmented information into shared operational intelligence.

Over the years, we have built experience across clients, industries, and delivery teams. That knowledge was spread across systems, documents, and employees. Accessing delivery experiences and historical company data often meant piecing things together manually and checking sources with several team members.

A single source of (AI-supported) truth

SPARK was created to change that. The platform centralizes trusted data and embeds AI directly into workflows, ensuring consistency in how information is captured, structured, and reused. Teams access verified context in one searchable environment rather than rebuilding it each time.

Our teams also built SPARK to improve over time. As teams document work, generate materials, and update client information as part of their normal processes, the system updates. Team members also provide feedback from using SPARK — highlighting gaps, clarifying what needs refinement, and guiding the platform’s evolution.

We see SPARK as a living, breathing system, not a final project.

Moving toward an AI-native delivery model

As the benefits of AI emerged across teams, we began considering how the technology would impact delivery.

The traditional Software Development Lifecycle (SDLC) model assumes work moves in distinct, linear phases. Context is passed along at each handoff, and significant feedback was often left to the end of the cycle when it was difficult (and expensive) to change the product.

While we integrated AI in various delivery tasks, we hadn’t yet adjusted the delivery model it supported. Consequently, we saw an increase in speed, but not value or understanding for our teams.

We realized that in order to really connect AI with our delivery model, we had to rethink the entire framework. We considered how information and learning moved between team members and delivery phases. As we made adjustments, we spent more time on preparation and research, moved feedback earlier in the process, and ensured the team had all the context they needed to make clear, informed decisions. In each of these phases and decision changes, we considered how AI could support the new process.

The benefits of CDL

Over time, these changes became a completely new delivery model, which we call the Continuous Delivery Loop (CDL). As with most of our AI projects, CDL developed organically, an effort from various delivery teams to thoughtfully implement AI and improve workflows.

CDL is a delivery that learns as it goes. Work doesn’t wait for formal checkpoints to improve. Feedback shapes the next step. Human judgment is centralized and supported by AI-driven systems that remember what teams have already learned.

In developing the CDL framework, we saw a clear path from simply using AI in isolated tasks to allowing it to shape how delivery actually works.

A blueprint for organic AI adoption

At Coherent Solutions, we’ve seen AI transform daily work for many of our departments and teams. From reducing the work that needs to be redone at every project handoff to helping our sales team build better client relationships. We saw the most success when we gave teams the ability to adopt tools as they needed. As a result, we saw productivity gains and positive organic changes in how teams plan, collaborate, and keep work moving.

Many organizations are trying out AI tools. Few see lasting changes in productivity and employee satisfaction once those tools are in place.

To see real value in AI adoption, companies should focus on how the technology can be used to improve how information, learning, and human judgment are exercised throughout the organization over time. When that happens, AI is no longer an add-on, but an integral part of how employees, teams, and an entire organization function.