Q&A: This tech exec drives value with a customized AI approach

As artificial intelligence moves from Silicon Valley boardrooms to construction sites and engineering offices, the industrial sector faces a unique challenge: how to harness AI’s potential without getting lost in the hype. We sat down with Josiah Shelley, CEO of ForwardPath, to discuss why custom AI solutions are becoming the go-to strategy for forward-thinking industrial companies.

SiteNews: When starting ForwardPath, why did you focus on the Industrial Sector?

Josiah Shelley: It’s interesting how this came about. I built my first AI project six years ago, well before the current AI boom. Through my involvement in PEO Leadership, I regularly spoke about AI and its potential and I noticed a pattern emerging. Most companies had no idea where to get started or what the playbook should be to get something done with AI. And I kept having Construction and engineering companies approaching me with genuinely compelling needs and use cases for AI implementation. They weren’t just curious about the technology; they had specific problems they wanted to solve and the willingness to make a bet and invest in solutions. The industrial sector had this unique combination of complex workflows, valuable data, and motivated leadership that made it the perfect fit for what we wanted to build at ForwardPath.

Why has your approach leaned towards custom AI applications over off-the-shelf solutions?

There are essentially three main ways companies can adopt AI today: personal AI assistants like Copilot or ChatGPT, vertical SaaS products designed for specific industries, and custom builds. While each has its place, custom builds offer the highest level of accuracy and customization for organizations that deploy them. The recent MIT study validated this approach. This is particularly crucial for the industrial sector because every company has different workflows, unique data structures, and specific operational needs.

A general contractor’s proposal process looks different from a mechanical contractor’s. An engineering firm’s document review needs vary significantly from a construction manager’s. When you try to force-fit these diverse workflows into a one-size-fits-all solution, you lose the precision and value that makes AI transformative rather than just another piece of software collecting dust.

Where are you seeing AI drive the most tangible value in construction and engineering right now?

The biggest wins we’re seeing are in areas where professionals spend significant time on repetitive, knowledge-intensive tasks. Document review and comparison: imagine automatically identifying discrepancies between contract versions or quickly extracting submittal requirements from a 500-page specification document. We’re seeing 60-80% time savings in these workflows. Proposal and estimating work is another hot spot. Companies have decades of historical bid data sitting in folders, and AI can help them leverage that institutional knowledge to create more accurate, competitive proposals in a fraction of the time. One of our clients reduced their average proposal creation time from two weeks to three days while actually improving their win rate. When we spend a day with a client on strategy and use case identification, the opportunities are vast, but choosing the right opportunities is an indicator of whether the project will be successful or not.

What’s the biggest misconception about AI in the industrial sector?

That it’s going to replace workers. The reality is quite different. The most successful AI implementations we’ve seen augment human expertise rather than replace it. Think of it as giving your best estimator the ability to review 10 times more proposals, or enabling your project managers to spot risks before they become problems. More proactive, less reactive.

The industrial sector’s deep expertise and judgment aren’t going anywhere. What AI does is eliminate the mundane tasks that prevent these professionals from applying their expertise where it matters most. When a senior engineer spends less time searching through documents and more time solving complex design challenges, everyone wins.

You’re co-sponsoring the State of AI in AEC Survey. What trends are you tracking and what sort of insights are you hoping to get?

We’re particularly interested in understanding the gap between AI interest and actual implementation. Everyone talks about AI, but we want to know who’s actually using it, where they’re applying it, and what’s getting in their way of being successful.

One particularly interesting insight is that the companies who are investing in AI, want to invest more. Suggesting that the value is there and it’s the companies who lack strategy, training and where to find the value, they are getting stuck. More to come on insights as we analyze the data.

What separates successful AI adoptions from the failures?

Three things consistently separate winners from those who struggle. First, successful companies start with a clear problem, not a technology fascination. They say, “we lose $2 million annually on change order disputes” rather than “we need to use AI somewhere.”

Second, they involve the actual users from day one. We co-build solutions with the teams who will use them daily. If your superintendents aren’t bought in on that new AI tool, it doesn’t matter how sophisticated the technology is; it won’t get used.

Third, they think in terms of workflows, not point solutions. AI works best when it’s integrated into how people actually work, not when it requires them to completely change their processes to accommodate a new tool.

What should industrial companies be thinking about as they plan their tech strategy for 2026?

Stop waiting for the perfect moment or the perfect solution. The companies gaining competitive advantage right now are the ones experimenting, learning, and iterating. Start with one workflow, that takes too much time and where you have clear success metrics. Build something simple, measure the impact, and expand from there.

Also, invest in AI literacy across your organization. This starts with training people to use tools; it’s about helping them understand what’s possible so they can identify opportunities in their own work. Some of the best AI use cases we’ve implemented came from someone who understood enough about AI to say, “Hey, could we use this to solve this problem we face every day?”

Where do you see the industrial sector’s AI adoption in five years?

I believe we’ll look back at 2024-2025 as the inflection point when AI went from experimental to essential in the industrial sector. In five years, AI-assisted workflows will be as common as CAD software is today. The companies that start building their AI capabilities now (not just buying tools, but developing the organizational muscle to identify, implement, and scale AI solutions) will have a significant competitive moat.

The beauty of the industrial sector is that small efficiency gains translate to massive value at scale. A 10% improvement in proposal win rate or project margin might not sound revolutionary, but for a company doing hundreds of millions in revenue, that’s transformative. The firms that recognize this and act on it now are the ones that will be leading the industry in five years.

What are the top things you advise industrial enterprises to do today to start their AI journey?

We get this question all the time, and we’ve developed a systematic approach based on what actually works in the field.

First, deploy a strategy and methodology for pinpointing AI use cases within your business. Most companies struggle here: they know AI could help somewhere, but they don’t know where to start or how to evaluate opportunities. We’ve designed a one-day educational workshop that brings together different stakeholders to derive, define, and align on practical AI use cases. It’s amazing what happens when you get your divisional leaders, back office, and leadership in the same room talking about pain points through an AI lens.

Second, with use cases in hand, determine the right tool for each job. Some problems are perfectly solved by deploying Copilot, ChatGPT, or Claude. Others might already have existing software solutions that just need to be implemented. But many of the highest-value use cases will require custom tools. Don’t try to force everything into one approach.

Third, when you do need custom tools, build them right. Custom development offers the most accuracy and customization to drive maximum value. It allows for integration directly into your existing software stack and can be deployed in your own IT infrastructure, giving you the highest level of data security and control.

Fourth, invest seriously in training and enablement. The best AI tool in the world is worthless if your team doesn’t know how to use it effectively. We deploy training programs that take teams from AI-curious to power users who can operate at multiple times their previous capacity.

Finally, measure everything and iterate relentlessly. Track adoption rates, time savings, error reduction, whatever metrics matter for each use case. Use that data to refine your tools, expand successful implementations, and build excitement. AI transformation isn’t a one-and-done project; it’s an ongoing evolution that gets more powerful as your organization’s AI maturity grows.

This is exactly what we do at Forwardpath: from strategy, to custom tool development to training and beyond. Learn more about how we partner with industrial enterprises to turn AI potential into measurable business value at forwardpath.ai

ForwardPath is currently analyzing results from the State of AI in AEC Survey to understand where AI is driving real value in the industrial sector. The company works with organizations like Rimkus Engineering, Kinectrics, Tiree, and Laframboise Group to develop custom AI solutions that accelerate workflows and unlock organizational knowledge.