Apr 9, 2026 · 47m · neon-show
Why Construction still Runs on Spreadsheets (And How AI Fixes It) | Sneha & Graham, Merlin AI
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of The Neon Show, host Siddharth Ahluwalia interviews Merlin AI leaders Sneha and Graham on how vertical AI replaces legacy ERPs and manual spreadsheets to transform construction management. They explore the company's journey from forward-deployed field immersion to building an agentic operating system that orchestrates complex project workflows.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. How this is scored →
Siddhartha as informed peer 2.4 Guest teaching 4.4 Guest disagreement 0.6 Siddhartha pushing back 0.1
speaking balance: gold is Siddhartha, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 31:20 Graham dismisses legacy construction software as wasted effort
Graham bluntly argues that the last ten years of construction tech merely digitized and cemented flawed existing behaviors instead of improving building delivery.
Hardest push from Siddhartha ▶ 45:28 Host questions construction vs manufacturing market scaleThe host steps in to challenge Sneha's comparison by asking whether construction is genuinely larger than manufacturing in the US market.
Biggest teaching moment ▶ 27:01 Graham educates on capturing trillion-dollar waste over IT budgetsGraham corrects the narrow framing of technology budgets by explaining that Merlin's true market opportunity lies in monetizing systemic margin waste and risk padding.
Siddhartha holds their own ▶ 45:56 Host cites customer insight calling previous software a black boxThe host demonstrates deep familiarity with the portfolio company by recalling early site visits where customers rejected incumbent software as impenetrable black boxes.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Siddhartha as informed peer | Guest teaching | Guest disagreement | Siddhartha pushing back | Why |
|---|---|---|---|---|---|---|
| The Failure of Generic Software in Construction | 3 | 5 | 1 | 0 | The host asks why general enterprise tools like NetSuite fail in construction. Sneha educates him on how every construction project is fundamentally a financial project requiring domain-specific ERP rather than fragmented point solutions. | |
| Graham's Background and the Differentiator of AI | 2 | 5 | 1 | 0 | Graham details the harsh economics of the production layer and trade contractors where thin margins reside. He explains how AI uniquely handles massive coordination volume compared to earlier web apps. | |
| Sneha's Transition from Manufacturing to Merlin AI | 2 | 5 | 0 | 0 | Sneha outlines how offsite manufacturing insights exposed severe operational and financial friction in prefab shops. She shares how customer reliance on messy SmartSheets and Dynamics validated building Merlin. | |
| Meet the Merlin AI Founding and Leadership Team | 2 | 4 | 0 | 0 | The host asks about the leadership team, and Sneha explains how deep tech expertise combined with field-level empathy led to voice agents replacing manual mobile data entry for workers. | |
| Real-Time Budgeting and Instant Quality Control in the Field | 2 | 4 | 0 | 0 | Sneha shares a practical case study where an electrical layout mistake was caught in real time via an uploaded photo before drywall installation, illustrating live budgeting and QC. | |
| Reinventing Construction ERP as an End-to-End OS | 3 | 6 | 1 | 0 | Sneha and Graham unpack why incumbent ERPs and Procore fall short, explaining that Procore functions primarily as a GC project management tool while Merlin acts as an operational operating system across fragmented SMB subcontractors. | |
| Navigating US Construction Personas and the Production Layer | 2 | 5 | 0 | 0 | Sneha and Graham categorize industry personas from offsite modular builders to subcontractors and developers, noting Merlin's specific focus on the execution-oriented production layer. | |
| Software Spend vs. Unlocking Trillion-Dollar Industry Waste | 3 | 5 | 2 | 0 | When the host asks about typical software spend percentages on a $100M project, Graham gently reframes the question, arguing that the real prize is capturing the enormous margin waste and inefficiency rather than existing IT budgets. | |
| The Transformative Influence of Y Combinator and Neon Fund | 3 | 1 | 0 | 0 | A collaborative investor-founder reflection where Sneha recounts the value of YC and credits the host's fund for backing vertical construction tech before mainstream validation. | |
| Moving Beyond Digital Policing to Agentic Orchestration | 2 | 6 | 2 | 0 | Graham offers an incisive critique of the previous decade of construction tech, arguing Procore cemented legacy behaviors rather than transforming delivery, and explains the shift toward proactive agentic orchestration. | |
| Scaling Merlin to $100M ARR via Organic Network Effects | 2 | 4 | 1 | 0 | Sneha playfully acknowledges the classic VC scaling question, then details the path to $100M ARR through embedded transactions and viral network effects across trade contractors on project sites. | |
| Overcoming Delivery Constraints and the Forward-Deployed Strategy | 3 | 4 | 0 | 0 | The host notes Merlin's unique delivery-constrained dynamic, prompting Sneha to explain how hands-on forward-deployed engineering and building Excel pivot tables for clients created early adoption. | |
| Founding Conviction and Bringing Deep Tech into the Field | 2 | 4 | 0 | 0 | Sneha discusses her conviction in serving non-tech blue-collar workers and how co-founder KS was convinced after living on customer job sites for months. | |
| Advice for Vertical SaaS Founders and Choosing Construction | 3 | 4 | 1 | 1 | Sneha advises vertical SaaS founders to do hands-on factory work rather than academic research. The host queries whether construction is bigger than manufacturing in the US, and Sneha emphasizes global market size and readiness for disruption. |
Statements from this episode (23)
Kumari: Every construction project is fundamentally a financial project
“Every construction project is a financial project at the end of the day. Right? So what, An owner, a builder cares about is end of the day. And did I make money on that project or not?”
Blake: Subcontractors bear the construction industry's greatest operational and financial pain
“This production layer and in the sector, regardless of asset class is where the pain lives, like subcontracting trade contractors, they own the pain. It's thin margin, high risk, everything flows downhill to them.”
Blake: AI is the only technology capable of coordinating complex construction project data
“The volume of data required to coordinate all those things from the light fixtures to every other aspect of the project. There's so much volume and complexity and dependency there. And quite frankly, you know, AI is the only thing that can solve for that.”
Kumari: Construction's operational workflows are worse than manufacturing
“And while manufacturing still needs a lot of work, right? Construction was worse.”
Kumari: Prefab operations failed despite using Dynamics, Smartsheet, and Excel
“So when I was running an ops floor for a prefab trust and panel panelization company, they had bigger problems, not just on the manufacturing floor, but on the financials as well, and they were constantly using smart sheets, and I believe Dynamics was their so…”
Kumari: Merlin AI bypassed market education because OpenAI established AI's value
“We did not have to do a lot of convincing when it came to AI, because I would say OpenAI did a great job there. People understand the value.”
Kumari: Merlin AI uses voice agents for construction site status reporting
“So how about deploying voice agents, which is what we are doing, right? So people on site just pick up their phone, talk about their status and leave, right?”
Kumari: Modular construction companies became Merlin AI's core customers within six months
“I think our very early on customers again for, I have to say modular guys, I mean, they have been with us. We, it's our very soon, very fast, within six months, it was our bread and butter. Also given my network, it was easy to acquire those customers.”
Blake: Construction tech stalls from operational contradictions, not adoption resistance
“I think that's, it's no longer an adoption challenge. It's not an objection. It just there's some contradiction in, in the way these businesses operate today. That means they need to adjust to really get the benefit.”
Kumari: Merlin AI operates as an end-to-end construction operating system
“Ultimately what we want to call ourselves is we are an end-to-end enterprise operating system for you know, for any, you know, company that exists today. But because ERP is a more known term than calling an enterprise software, it just made sense that we put i…”
Kumari: Procore is not a true enterprise solution for construction
“Procore is not really an enterprise solution. They're really very popular and very uniquely positioned to solve for general contractors from a project management, a little bit of financial side, and that's what the role and, you know, I will of course have Gra…”
Kumari: Merlin AI has replaced NetSuite, Dynamics, SAP, Epicor, Syspro, and Odoo
“As far as NetSuite and, you know, Dynamics and SAPs and Pro and Epicore, SysPro, all of which we have replaced, Odoo also, by the way the latest to join the bandwagon, like all of these were not built for construction.”
Kumari: Construction projects allocate only 1% to 5% of budgets to technology
“Clearly you can see two to three percent of the overall project is going to technology and doesn't have to be just Merlin. There's a lot of other stuff, but then people who are recurring building this Technology sits at the core of what they are building, righ…”
Kumari: Few venture investors understand vertical SaaS for the construction industry
“Not a lot of investors
Understand the play of vertical SaaS and forget about understanding the, in the world of construction. A lot of investors would come us to build their own thesis.”
Blake: Procore digitized construction but locked in outdated legacy workflows
“Procore without doubt has finally broken through a level of, you know, digitization of a section, but at the same time, what it's done is actually crystallized and cemented and locked in and embedded, you know, the existing behavior.”
Blake: Kojo successfully targets practical construction supply chain workflows
“So the ones that the ones that excite me, I would just say, you know, like, I, you know, Kojo, I think is a good example. People that are you know, are active in the supply chain side who are thinking about how the work actually gets done, how to enable that p…”
Blake: Proactive agentic orchestration is the next phase in construction tech
“No one's there yet, but this step towards true, agentic, you know, proactive orchestration that has to be next, right?”
Kumari: Merlin owns transaction flow but leaves marketplace monetization untapped
“We are talking, we own the whole transaction piece, right? And we are not even doing anything about it. That's right there for us money on the table. And we will get there. So there's a whole marketplace right there where people are coming to us, where we coul…”
Blake: Construction tech expands via project networks exposing 10x more users
“Network is the, is one of the very positive things about our sector. You know, the pervasiveness of having one customer and that customer touches 10 or 50 projects a year and every one of those projects, you know, ultimately incorporates or includes another Yo…”
Kumari: Word-of-mouth remains the dominant growth driver in construction software
“And then as we highlighted, word of mouth is phenomenally a very big perpetrator of your value in construction, even today. So if you are my competitor, I'll go to your website, I'll see who you're talking to, and I'll go use that if that has, if I'm seeing yo…”
Kumari: The construction industry is underserved by top tech talent
“I feel we are underserved in terms of there aren't very many intelligent people who are trying to solve for us.”