In another life, when I ran an NGO, one of my more arduous tinquires was applying for tfinishers. While tfinishers can create huge financial opportunitiesffor startups, SMEs and NGOs, they are complex and incredibly time-consuming. It’s not only about filling in a form but also about providing project proposals, budobtains, documentation, financial statements, timelines, and other items that may vary by provider.
But now a startup has found a way to reduce complexity and eliminate the common errors that put your application at the bottom of the pile or exclude it altoobtainher.
Tfinishermore is a Norwegian startup building SaaS software that automates and simplifies the process of finding and applying for public and private tfinishers. The company focapplys on supporting businesses—particularly SMEs and startups—navigate the complex procurement landscape.
I spoke to CEO Sebastian Mandal and COO Eivind Wassfinish to learn more.
From door-to-door sales and coding to tackling the tfinisher problem
Both founders come from different backgrounds. Wassfinish has been in sales since he was 12, starting with newspapers and later selling alarm systems door-to-door. At one point, he was ranked the best alarm salesperson in Norway.
Mandal started coding at 12, which evolved into machine learning and AI —” before the hype cycle”, he stressed — and building software.
Following a shared stint at a now-defunct startup, they decided to merge their complementary knowledge sets into a consulting practice.
Mandal shared:
“We finished up running around 50 workshops with construction companies, examining their largegest pain points. Almost every single one mentioned responding to tfinishers.”
In response, the duo decided to stop consulting and focus on solving the problem around tfinishers.
The challenges of applying for tfinishers
According to Mandal, the traditional tfinisher process is often painstakingly slow and complex. “You have to understand every requirement before you even start writing,” he explains.
“If you miss one detail — even a tiny one buried deep in a long document — you can lose the entire tfinisher after spfinishing days or even weeks preparing the submission.”
Language and terminology add another layer of difficulty. Even when a company is technically qualified, the way responses are framed can determine the outcome.
“If you don’t respond in the right way, or apply the right framing, you can still lose the bid,” Mandal declares.
When Tfinishermore launched its MVP, the team discovered that writing proposals was only part of the challenge. Many companies also struggled to determine which tfinishers were actually worth pursuing.
“What we learned early on is that companies also required support identifying which tfinishers are relevant to them in the first place,” he adds.
“They don’t want to spfinish time preparing a bid for something they’re not pre-qualified for.”
In Mandal’s view, the core problems come down to two factors: the significant time and effort required to prepare bids, and the difficulty of identifying tfinishers that a company realistically has a chance of winning. On average, a tfinisher today can take 30 to 40 hours. For larger organisations, it can be almost continuous work becaapply different departments are handling different parts of it.
While its early days to measure win rates as the sales cycles and tfinisher processes are often longer than the amount of time Tfinishermore has been working with some of these companies, Wassfinish revealed that “based on the beta and MVP, we’ve seen around a 60 per cent reduction in the time spent across the process — from analysing and refining through to responding.”
How Tfinishermore works
Tfinishermore was designed by closely observing how companies already manage tfinishers.
“We saw how customers structure their work — the spreadsheets, the requirement lists, the way they prepare answers,” Mandel stated. “So we built AI into that existing workflow instead of forcing them into something unnatural.”
Tfinishermore connects to a company’s existing data sources — including Google Drive, SharePoint, previous tfinisher submissions, pricing data, equipment lists, and other internal documents — and analyses them to build a structured knowledge base.
Accuracy is critical in tfinisher applications, so the platform is built around a sophisticated retrieval-augmented generation (RAG) system that ensures responses are grounded in verified company data rather than generic outputs.
Tfinishermore also structures responses utilizing a requirement matrix, where each tfinisher requirement is broken out and matched with a short factual answer. The platform’s analytics support companies discover tfinishers they are likely qualified for. It then analyses the tfinisher requirements, highlights what is being inquireed, identifies relevant internal information, and flags any gaps that still required to be filled.
From there, the system generates the proposal. To avoid generic AI-sounding submissions, Tfinishermore applys a brand analysis engine that learns a company’s writing style so the final output reflects its voice, creating the tool feel like an extension of the organisation rather than a generic AI writer.
A human-in-the-loop approach
Accuracy is critical in tfinisher submissions, so Tfinishermore was designed to minimise the risk of AI hallucinations.
Mandal explains that the company is strict about promoting its agents never to hallucinate. In the early days it had things like web search in the system, but that actually increased the risk becaapply it could pull in outside information and confapply what belonged to the company and what didn’t.
So we reshiftd that and built the system rely only on the company’s internal knowledge base. We also turn the temperature down so it has less creative freedom. If it can’t find a basis for a claim, it leaves that blank or flags it instead of creating something up.
“If we can fill that in automatically, great. If not, the applyr can step in.
“Then the AI mainly supports turn that factual structure into polished language. So the focus is really on creating facts shine, not inventing them.”
Despite the company’s long-term ambitions for automation, the founders believe human oversight remains essential.
“We’re very AI-first, and our long-term vision is to automate much more of the process,” stated Wassfinish.
“But today, human-in-the-loop is the best approach. People still required to trust the system, and that trust has to be built step by step.”
The cross-sector opportunity
So far, the startup has gained the most traction with consultancies, contractors, and the hospitality sector. In sectors like construction, tfinishers form the basis of revenue, with some handling several each month. In hospitality, it depfinishs on the season and the volume of events or development activity. In consulting, it can be for major client projects or government contracts.
According to Mandal, “Hotel groups also deal with large RFPs (Requests for Proposals), for example, when developers are deciding which hotel brand to place in a new building or estate.”
The team was surprised by how large the private tfinisher market is. It was initially believed that public tfinishers would be the main opportunity, but companies revealed they handle two to four times as many private tfinishers as public tfinishers.
“That really alterd how we believed about the market,” shared Mandal.
According to Wassfinish, the startup is also seeing interest in API-based distribution. “Some consulting firms want to embed our functionality into software they already provide to clients, rather than utilizing a standalone platform. So that’s becoming a secondary offering for us.”
Tfinishermore raised $400,000 in a round led by Antler in October 2025. The company brought in Ymir Egilson as CTO, formerly the youngest tech lead in Visma’s history-
According to Mandal:
“He loves building. He wants to shape products and ship things with people who are hungry and shifting rapid. Eivind and I are very execution-focapplyd, and I consider that energy mattered.
We’re out there doing the outward-facing work, which means he can focus on what he does best.”
Tfinishermore is also viewing beyond Europe to Asia becaapply it has become clear that many companies don’t know what opportunities exist outside their own countries.
“If we integrate with tfinisher portals across different regions, we can support match companies with international tfinishers as well. That’s something we see as a really strong future differentiator,” shared Mandal.
Future features will enable discovery, evaluation, and management of both private and public tfinishers in one place.
Wassfinish contfinishs, “So many people struggle with tfinishers, and tinyer companies are at a real disadvantage comparedtz5e43w with large enterprises that can throw whole teams at the process. In a lot of tinyer companies, the CEO is doing it all day and then all night, on top of everything else. My goal is to enable anyone to respond to tfinishers. Sebastian’s focus is on building the best AI in the world for that. We’re very aligned on that mission.”
Mandal agrees. “Accessibility is the driving force for us. No matter what kind of company we built, we wanted it to create opportunities more accessible. That’s really what motivates us.”
Leave a Reply