Atma Energy

To kick off the work, I facilitated a brand focus workshop with leadership to align on positioning and values, then delivered a Brand Focus book documenting the strategic foundation.

From there I designed the logo mark, wordmark, and logotype, defined the color palette, and crafted the visual direction:

The company's primary typeface is Area Normal by Blaze Type. Area Normal is a warm sans-serif font, geometric with neo-grotesque attributes that give it a feeling of stability, clarity, and humanity that matches the brand voice.

I workshopped and selected the company typeface with the goal to make the company typography as simple and clear as their products and services.

The full brand guidelines were delivered first as a PDF. When I established the company playbook using Outline, I moved the brand guidelines to this living document to increase their accessibility for the team as well as external marketing stakeholders:

I also created the company's big idea. The big idea captures the story of the brand in three simple words:

With the brand established, I created the company design system in Figma and designed the marketing site components and website. The system covered both the public-facing marketing site (Statamic + Tailwind) and the internal admin panel (Filament).

Proposal tool

Summary

This custom solar proposal system replaced a manual, Excel-based workflow. It initially decreased proposal creation time by 78%, and a later automation integration reduced the time needed by another 40%. The proposal system supported annual sales growth from under $2M to over $10M.

Context

Atma sales reps create thousands of solar proposals for homeowners. Each proposal contains details about system design, savings projections, financing options, and equipment specs. When I joined, reps were working from a complex Excel spreadsheet filled with calculations. The manual workflow included preparing the data, populating the sheet, removing irrelevant sections, exporting a PDF, uploading it to Google Drive, and emailing it to the customer. The spreadsheet was time-consuming to fill out, difficult to adjust, and increased the risk of error.

Problem

The process couldn't scale, and it produced proposals with inconsistent quality and presentation. This destabilized the connection with homeowners at a critical stage of the sales funnel. Homeowners lacked a structured path from initial proposal delivery to contract signing, and reps handled responses ad-hoc as homeowners had no self-service option for reviewing options or taking next steps.

Approach

There was no off-the-shelf solution for this. Atma’s custom requirements and dedication to transparency with customers meant that no third-party proposal solution would suffice. Because a unique system was needed to accurately reflect Atma's sales process and product offering, there was no real blueprint to reference which made the design phase critical. I walked stakeholders through flowcharts and prototypes repeatedly to validate decisions and flush out hidden requirements before anything was built.

A key architectural requirement lay in separating the two stages of the proposals: the Initial Proposal (IP) and the Engineered Proposal (EP). The IP is generated early in the sales process after a discovery call with the lead. Before the lead is committed, speed is the most important factor. The IP is like a rough draft – panel placement is approximate and measurements aren't exact, but it's good enough to give the homeowner a realistic picture. When the lead chooses to move forward, Atma schedules a drone flight to capture detailed roof imagery. Engineers use the drone data to design a precise system – this is the EP, the proposal the homeowner signs against.

On the customer side, I designed a tabbed proposal template that guided homeowners through a structured flow: Meet Atma, Your energy needs, Your solution, Savings, Financing, Benefits, Process, and Configure. The Configure tab served as the conversion point, where homeowners could move toward agreement creation on their own terms.

Scanifly integration

After the initial launch, the system went through multiple iterations and improvements based on feedback from sales team, engineers, and customers. As AI-powered design solutions came to market, our engineering team identified an opportunity to automate IP creation using a tool from Scanifly. I worked with engineering to translate this into a production feature.

The integration connected Monday (lead management), the Atma panel (Filament-based proposal system), and Scanifly (third-party solar design tool) into an automated pipeline. When a rep qualifies a lead and changes a status in Monday, the system automatically creates the user, generates the proposal, and creates a Scanifly account – all without manual data entry between systems. The rep then opens the proposal tool, triggers the automatic design in Scanifly, finalizes it in minutes, and hands off to the homeowner.

This system reduced initial proposal creation time by an additional 40%, eliminated manual data entry between three systems, and ensured data consistency across the pipeline.

Build

The proposal system was built in Laravel with a Filament admin panel for reps, Vue components for the interactive customer-facing views, and Tailwind for styling. Chart.js handled savings and energy visualizations. The Scanifly integration used Monday webhooks and the Scanifly API, with validation logic to catch missing data before proposal creation.

Outcome

• Proposal creation time decreased by 78%

• Scanifly automation reduced IP creation time by an additional 40%

• Proposal quality became consistent across all reps regardless of experience

• Homeowners gained a self-service path from proposal review to agreement

• The system contributed to annual sales growing from under $2M to over $10M

Reflection

The most difficult part of this project was the lack of reference points. Every feature was custom to Atma's process, and most problems didn't have existing patterns to borrow from. This was true for nearly every project at Atma. I learned quickly to take a hard and fast approach to validation, with multiple flowcharts, prototypes, and requirements docs requiring approval from stakeholders before any code was written. The cost of building the wrong thing in a bootstrapped company with a small team is high, so the up-front investment in alignment paid for itself many times over.

Service platform

Summary

This subscription-powered service platform replaced manual ticket management with an automated workflow. It reduced service response time from weeks to hours and established a new recurring revenue stream.

Context

Atma monitors customer solar systems for issues. This includes panels underperforming, batteries not reporting, inverters failing, and more. When I joined, the operations team managed service entirely through third-party monitoring, email, phone calls, and a Monday board. Tickets were created and updated manually. Customers had no visibility into their ticket status beyond whatever they were told on a call. The oldest unresolved ticket was several months old.

The company had validated demand for a service subscription product, but the existing subscription model contained design flaws: the pricing tiers were confusing to customers and the structure would have been difficult to implement technically.

Problem

There were several problems that compounded each other.

1. Manual ticket management didn't scale. Every status update depended on someone remembering to do it, and the risk of error was high.

2. Customers had zero self-service visibility, which meant every status inquiry required a phone call or email that pulled the ops team away from their work.

3. The validated subscription product couldn't ship in its current form without creating technical debt and customer confusion.

Approach

The subscription needed to be fixed before the platform could be built. I spent several weeks working with stakeholders to demonstrate the flaws in the existing subscription design – the tier structure was ambiguous and the pricing logic would have been fragile to implement. Eventually, after reframing the problem from the customer's perspective, stakeholders agreed to simplify. The final model was a single annual subscription at $365 with clear on-site service pricing: $150/hour for the first two hours, $100/hour after that.

For the platform itself, I designed a 20+ status ticket workflow that automated as much of the process as possible. The system used the Enphase API to monitor customer systems every 15 minutes, checking for a defined set of error statuses. When an issue was detected it triggered an alert. For Premium subscribers, detected issues automatically created a service ticket and started the workflow with no human intervention needed.

A key product decision was the subscription paywall for automatic ticket creation. Premium subscribers got automatic monitoring and ticket creation. Basic users saw system alerts on their dashboard but had to initiate a ticket themselves. This added clear value to the subscription without leaving basic users in the dark about their system health.

For customers, I designed a service section in the dashboard with ticket tracking, subscription management, and warranty information. For admins, I created a service management interface in Filament with role-based access: operations manager, service manager, service representative, and service technician — each with appropriate permissions and views.

Build

The platform was built in Laravel with:

• Filament for the admin panel

• Stripe for subscription billing and payment management

• Enphase API for automated system monitoring.

The ticket workflow included:

• Automated email notifications at key status changes

• Calendly integration for scheduling service appointments

• Image upload for documentation

• Service report PDF generated at ticket close

Outcome

• Service response time reduced from weeks to hours

• Automated monitoring checks all customer systems every 15 minutes

• Subscription model created a new recurring revenue stream

• Customers gained real-time visibility into ticket progress, reducing inbound calls and emails

• Ticket workflow automated status transitions, notifications, and handoffs that were previously manual

Reflection

The hardest part of this project wasn't the design or the technical build – it was adoption. The operations team knew the existing process was broken, but they were time-poor and deep in daily firefighting, which made adopting new practices difficult even when the new system was objectively better.

Every improvement met resistance at some point: the subscription model needed weeks of stakeholder alignment, Calendly required education and convincing people of its merits, and the in-app messaging system (designed specifically to replace email and phone for ticket communication) was initially bypassed in favor of the old methods.

After launch and for the first few months, the platform was seen internally as underwhelming. Then adoption caught on: the ops team integrated the workflows into their daily process, subscription revenue grew, and leadership's assessment shifted. What looked like a slow start was actually the normal cost of changing how people work.

The lesson learned was that shipping the product is not the end of the story. A well-designed system still requires patience, presence, and sometimes just time before the people it serves make it their own.

Contract automation

Summary

This project moved contract creation from a manual, rep-driven process into the customer's hands. It removed the friction that was losing deals at the final stage of the sales funnel.

Context

After a homeowner reviewed and approved their Engineered Proposal (EP), the next step was signing a contract. This was the last stage before becoming a customer, and it was where Atma was losing deals they'd already won. Reps estimated they lost up to 10% of deals at the contract stage. It wasn’t because homeowners had changed their mind about solar – it was the process itself that created discomfort and delay.

Problem

The contract process was entirely manual and rep-controlled. Homeowners wanted to understand what they were signing, but the only way to see the contract was to ask the rep to generate one – which felt like a commitment before they were ready. Reps tried to work around this by sharing blank contracts or walking leads through the terms over email and phone, but this created protracted back-and-forth discussions that drained momentum.

When a contract was finally ready, the rep created it manually in DocuSign. If the homeowner wanted to change their financing option or add a subscription, the rep would void the contract and recreate it from scratch, which generated more emails, more calls, more manual work, and more opportunities for the deal to stall.

Approach

The core insight was that the contract stage failed because it took control away from the homeowner at the last moment they most needed it most. We had already created a self-service proposal to support and increase efficiency for all users. We just needed to align the contract feature with the rest of the template. The solution was to invert the process: let the homeowner configure and generate their own contract.

I designed a contract page, separate from the proposal, where the homeowner could select their financing type (express loan, traditional loan, or cash), choose subscriptions (battery and service plans), enter a down payment amount, and generate the contract themselves. The contract was then presented for review and signing via an embedded DocuSign experience that didn’t require email attachments or back-and-forth discussions.

Key decisions:

• Separated the contract page from the rest of the proposal to create a clear psychological shift from "reviewing options" to "making a decision"

• Restricted contract creation to Engineered Proposals only, ensuring the homeowner was working from verified system specifications

• Added a 30-day EP expiration – if the proposal was older than 30 days, the homeowner was redirected to contact their rep for an update before proceeding, protecting against outdated pricing or specs

• Built five distinct states for the Configure tab so the homeowner always knew exactly where they stood: ready to configure, contract generated, homeowner signed, both signed, or proposal expired

On the rep side, I designed automated notifications at each stage — contract created, homeowner signed, ready for rep signature — and Monday board updates that moved the lead through pipeline stages automatically. After both signatures were collected, the system promoted the proposal to Contracted status, created an installation record, and updated the homeowner's role from Lead to Customer.

Build

The system was built in Laravel with a new DocuSign service for contract generation, embedding, and webhook handling. Financing and subscription selections fed into the contract template dynamically. Stripe handled subscription billing setup. Monday API integration automated pipeline stage transitions and status column updates. The Configure tab used state-driven rendering in Vue, with each of the five states pulling from proposal and contract status data.

Outcome

• Homeowners can configure and generate their own contract without rep involvement

• Contract revisions no longer require manual recreation — homeowners can reconfigure before signing

• Automated notifications and Monday updates eliminated manual pipeline management

• The full flow from contract generation to installation creation runs without manual intervention after both signatures

• Addressed the contract friction that reps estimated cost up to 10% of deals at the final stage

Reflection

This project reinforced something I'd seen in several of the major feature builds at Atma. The biggest gains often come not from building new capabilities but from removing the friction in processes people are already trying to complete. In this case, homeowners wanted to sign, and reps wanted to close, but the old process stood in the way of them both. The solution lay in handing over control to the person who needed it most at that moment.