The startup idea did not arrive on a whiteboard. It arrived as repetition. Melanie Fellay was working at RealtyShares, the real-estate investment startup, when she became responsible for helping people use Salesforce. The software had cost serious money. The instructions lived in PDFs, PowerPoints, Google Docs and training sessions. Yet the same questions kept coming back, as faithfully as calendar reminders.
This is the sort of nuisance an operator can either tolerate or interrogate. Fellay had moved through investor relations, customer success, servicing, business operations and eventually chief of staff. She knew the questions were not evidence that colleagues were careless. The system required people to leave the work, find a repository, remember the right search terms and translate generic guidance back into a live task. Each step was a tiny tollbooth.
Her question was plain: what if the information found the employee instead?
A product still needed a product builder. Fellay learned that her coworker Zari Zahra was resigning. She met Zahra by the elevators and pitched the idea. They were not childhood friends rehearsing a dorm-room destiny. They were colleagues with complementary evidence: Fellay understood the operational pain; Zahra understood how to shape software around it. The hallway conversation lit the fuse for Spekit, founded in 2018.
Watch the workaround, not the wish list
Spekit’s origin contains a useful discipline for would-be founders. Fellay did not begin by asking employees what futuristic training platform they wanted. She watched them struggle with a process everyone had already accepted. The recurring questions, scattered documents and forgettable sessions were a shadow product. People had constructed it themselves because the official product did not fit the moment.
Find the repeated question
Repetition often marks a system failure disguised as a people problem.
Reduce retrieval friction
Every login, tab and hunt competes with the work a person was hired to do.
Bring context
The same instruction lands differently when it appears beside a real decision.
Make it memorable
A useful product still needs a shape, voice and story people can recall.
The word Spekit is short for a “snippet” of knowledge. The company embedded those snippets across workplace tools, turning training from a place into a layer. It combined elements of a knowledge base, a learning system and digital adoption software. For an employee, the promise was simpler: keep doing the task; the answer is nearby.
Three countries in the operating system
Fellay’s taste for precision and color has a geographic backstory. She was born in Geneva and spent her early childhood in a Swiss village with her younger sisters, Izzy and Audrey. School was in French. English came at home from her Colorado-born mother and during summers with American relatives. At fifteen, her father’s work took the family to Nassau.
She has described her leadership style as a mix of Swiss perfectionism and quality obsession with Bahamian joie de vivre and family values. It is a neat formula, but the interesting part is the tension. Enterprise software rewards rigor. Young companies need warmth, improvisation and the courage to look unlike the procurement checklist.
At St Andrew’s International School in the Bahamas, Fellay graduated as valedictorian. At the University of Colorado Boulder, she studied accounting and finance, minored in economics and Spanish for the professions, and finished in the top five percent of her Leeds School of Business class. She also led business-statistics recitations. Years later she would tell the story of a student who struggled with probability until she reframed the lesson around roulette. The formulas had not changed. Their context had.
Joins RealtyShares and moves through customer, Salesforce and business-operations roles.
Cofounds the company with Zari Zahra around knowledge inside the workflow.
A $12.2 million Series A follows rapid growth during the shift to remote work.
A $45 million Series B brings reported funding to roughly $60 million.
Fellay publishes Just-in-Time: The Future of Enablement in a World of AI.
Purple, yellow and eight arms
Spekit did not dress like a company hoping to disappear into a Gartner report. The palette was bright. The mascot was an octopus. Eight arms suited software meant to reach into the many applications surrounding a modern worker, but the creature also did essential memory work. Fellay believed brand should make the company recognizable before a prospect had decoded the feature list.
That belief was tested early. The team booked a booth at Dreamforce before it had customers. At another early conference, the booth materials failed to arrive. Fellay and the team ran to nearby stores, bought what they could and assembled something presentable in a few hours. An enterprise-software booth built with retail improvisation is a better founder photograph than a gong beside a funding announcement.
Fundraising did not always provide the polished photograph. In 2019, Spekit had six team members, about as many customers and little revenue. Fellay was living with her grandmother. She and Zahra were not taking salaries. Professional investor conversations had been discouraging. Then a friend’s post led Fellay to Dan Scheinman on Twitter. His profile mentioned early investments in companies including Zoom. She reached out; the connection became important to Spekit’s next chapter.
Two rounds, ten months apart
In March 2021, Spekit announced a $12.2 million Series A. Ten months later came a $45 million Series B led by Craft Ventures, bringing total funding to about $60 million. Forbes placed Fellay on its 2022 30 Under 30 list in Marketing & Advertising. The customer list grew to include companies such as Southwest Airlines, Uber Freight, iHeartMedia and Equifax.
Candor travels upward
Fellay’s most instructive management anecdote is not about an inspired all-hands. It begins with an uncomfortable email. After a meeting, a leader told her, candidly, that she had a habit of entering without full context and steering the discussion too quickly. The phrasing was earthy. The substance was unmistakable.
Fellay shared the email anonymously with the company. Her point was that radical candor could not be a value reserved for feedback flowing down an org chart. “To win together, we must learn together,” she said, adding that she was as fallible as anyone else. The response turns a private correction into public permission: if the CEO can show the bruise, perhaps everyone else can stop armoring the mistake.
There is a parallel with the product. A healthy company and a useful knowledge system both shorten the distance between the problem and the information needed to fix it. Neither asks people to pretend they already know.
The AI chapter is a context chapter
By 2026, Spekit’s language had moved from digital adoption toward revenue enablement and AI. The original argument, however, remained intact. A sales representative does not need another impressive destination. The representative needs current, approved material beside the buyer, deal and task at hand.
Fellay calls the gap between abundant AI investment and stubbornly unchanged behavior a readiness gap. General-purpose systems can connect tools and generate fluent text. The harder task is knowing which company answer is approved, what a particular deal requires and what action should follow. Grammatically correct can still be operationally wrong.
Her 2026 book, Just-in-Time: The Future of Enablement in a World of AI, turns the company thesis into a broader framework. The aspiration she states publicly is both product-minded and personal: become the best leader she can be, build a company customers care intensely about, scale a distinctive team and share the truth of the journey. She also advocates for more women in technology and helped launch a Diversity at Dreamforce scholarship for underrepresented founders.
On Spekit’s eighth birthday, Fellay wrote with less gloss than anniversary posts usually permit. The company had known uncertainty, sudden growth, funding, product-market-fit questions and painful scaling years. She also wrote that it had found its footing again. The honesty gives the story its proper dimensions. Eight years is long enough for the elevator pitch to become infrastructure, and long enough to discover that the founder must keep learning in public.
The training manual was never really the enemy. Distance was. Fellay’s work has been an extended campaign against the distance between a question and its answer, a buyer and the right material, a leader and useful criticism. Her product philosophy can be reduced to a courteous design principle: do not make people leave the moment in order to understand it.