What a bid writer learned about AI that most consultants never will
Maybe* Team Partner Stories April 2026
Jake Wells spent 17 years writing public-sector bids and R&D grant applications for clients. He was not an early adopter of AI. He still takes notes with a pen in every meeting. And yet he built something with Maybe* that completely changed how his team works and delivered results that speak for themselves.
The numbers first
In the six months since deploying AI agents on the Maybe* platform, Jake’s team have delivered the following:
€1m
Horizon Europe CSA grant secured, in partnership with a leading UK university
£700k
Innovate UK grant awarded under the Investor Partnership Programme
77%
Success rate on public sector bids over the last six months, across 9 submissions securing up to £5m in funding
32–40hrs
Saved per proposal. A typical 160-hour engagement now runs 20 to 25% leaner
One of the bids that did not win scored 91% in evaluation. Another result is still pending. The 77% figure will almost certainly move. See the case study.
The problem every bid professional recognises
Bid writing looks, from the outside, like a writing job. It is not. Before a single word of a response is drafted, a bid professional needs to absorb anything from 10 to 25+ documents, research the buying authority, understand their priorities and build a picture of what this particular client actually cares about. That work takes days. And it takes days every single time.
Then the timescale gets compressed. It almost always does. And when it does, the stages that consume the most time are the ones that get cut. The research becomes shallower. The review is delayed. The submission becomes rushed as the deadline looms. Compliance is assured, but quality takes the hit.
“In an ideal world, you would have the full 4 or 6 week timescale to go through the process step by step. In reality, that timescale gets squeezed. And when it does, it is some of the most time-consuming elements that end up being missed out.”
Jake had lived with this for nearly two decades. He knew exactly which parts of the process were eating the time. He knew exactly what it would look like if those parts could be done faster. What he did not have was a way to do it.
Why did generic AI not solve it
Jake describes himself as a laggard when it comes to technology. He tried AI tools. His honest assessment was that the tools available required too much from the user and delivered too little that was specific enough to actually help a bid professional do better work.
The problem with generic AI in a specialist field is that it does not know the field. It can write. It can summarise. But it does not understand the logic of a public sector tender, the nuance of an R&D funding application, the structure of a winning bid narrative, or why buyer research matters as much as the response itself. Without that understanding, the output requires as much work to fix as it would have taken to do from scratch.
Using the wrong AI tool can be more damaging than using no tool at all. The difference is whether the expertise is built in from the start.
Building agents that actually work
When Jake encountered the Maybe* platform, what stood out was not the technology. It was the structure. The platform was built to carry real domain knowledge, encode it into agents, and deploy it in a way that made the expertise accessible rather than the technology visible.
Jake built AI agents that absorbed the tender documents, researched the authority, and handled the front-end stages of the bid process that had always consumed the most time. The agents follow the logical stages of bid management. They work the way a bid professional works, not the way a generic AI tool works.
“The ability to have a colleague in effect is how it functions and how it feels. One who has read all of the documents, who can answer all of the questions as you go through the process.”
The time saving of 32 to 40 hours per proposal is real. But the more important shift is qualitative. With the front-end work handled, the team can follow the full process on every bid, even when the calendar is tight. Reviews happen earlier. Writers get more feedback. The strategic, creative work that actually wins bids gets the time it deserves.
“Our experienced writers and consultants can spend more of their time on the strategic, quality-focused elements of the process. The agents handle the research, the re-reading of call text and specifications, and the early drafts.”
What building it taught him
Experiencing the Maybe* platform from the inside changed how Jake thought about AI adoption. He had seen, in his own professional domain, exactly how the right tool with the right expertise behind it could transform a workflow. A €1m Horizon Europe grant. A £700k Innovate UK award. A 77% bid win rate. These are not outcomes of deploying AI carelessly. They come from encoding 17 years of hard-won expertise into a platform that knows how to use it.
That experience is what Jake now brings to other businesses. Working alongside the Maybe* team, he helps organisations identify where their own domain expertise, properly encoded into the right platform, can change how they work. His approach is grounded in the same principle that shaped his own agents: understand the work deeply first, then let the platform do the rest.
“AI won’t take your job. But somebody using AI probably will. The question is whether you have the right expertise and the right platform working together.”
Work with Jake and Maybe* to optimise AI in your business
If you want to understand where AI can genuinely work in your workflows, talk to the Maybe* team and Jake.
