Back to all proofs

From a standing start to an AI venture serving small businesses

Client

Date

Abu Dhabi Department of Economic Development (ADDED), in partnership with Mohamed bin Zayed University of Artificial Intelligence (MBZUAI)

08/2026

The challenge

Abu Dhabi wanted the AI research talent already inside the country turned into companies serving its small businesses. Nothing existed to do it with: no cohort, no venture, and SME guidance that reached businesses only through advisors handling one enquiry at a time.

The solution

FifthRow ran the programme end to end as one pipeline: an AI talent pool and a policy mandate in, a soft-launched product with real users out. ADDED held the go/no-go at every stage gate.

2 weeks

no pipeline to 200+ screened

27%

beta sign-up rate
University outreach10+ universities across the region
Candidate pipeline200+ screened in 2 weeks
Ventures in market1 product soft-launched to SMEs
Beta sign-up rate27%
Time to launch4 months vs 12–18 typical
01 — TALENT SOURCING & OUTREACH

The programme opened with a direct recruitment push into universities across the UAE and the wider region rather than an open call. Applications were screened against a single standard so a cohort could be assembled in weeks rather than an academic season.

Core Activities

  • Ran outreach at 10+ regional universities
  • Administered applications end to end
  • Screened 200+ candidates against one standard
  • Shortlisted the strongest for an in-person bootcamp

The Results

More than 200 applications inside two weeks, narrowed to a shortlist of 40+ candidates. ADDED went from no pipeline to a full cohort before the first teaching session.

200+

candidates sourced in two weeks
02 — SELECTION BOOTCAMP

The shortlist went through a working bootcamp, not an assessment centre. Candidates formed teams, took ideas through validation and market testing, and built toward an MVP – so the selection decision was made on what they produced, not on what they proposed.

Core Activities

  • Ran ideation and validation sessions
  • Market-tested ideas before any build commitment
  • Facilitated team formation
  • Ran MVP build fundamentals, judging output not pitches

The Results

One winning team emerged from the bootcamp with a concept ADDED backed: an AI chatbot serving small businesses across the emirate. That team entered a four-month mentored build.

40+

founders judged on builds
03 — VALIDATION & MVP BUILD

Before the build started, the concept was tested with the people it was meant to serve. Focus groups validated the product concept, and the datasets that would carry regional context into the answers were assembled in parallel with development.

Core Activities

  • Validated the concept in SME focus groups
  • Narrowed chatbot concepts against that feedback
  • Structured the datasets carrying regional SME context
  • Mentored the build: features, testing, bug fixing
  • Fed the client's written prototype review back in

The Results

A working MVP chatbot in user acceptance testing, built on validated concepts and regionally grounded data, plus a finished landing page and launch collateral ready for release. A structured client review then reset the product before launch, on the points that mattered most to the authority: how completely it covered its remit, how directly it answered, and how it presented itself.

4 months

mentored build to user-acceptance testing
04 — SOFT LAUNCH & BETA TESTING

The MVP went out to real small business users as a soft launch, with a product demo run in parallel to collect structured feedback on answer quality and satisfaction.

Core Activities

  • Soft-launched the MVP into a structured SME beta
  • Ran a demo capturing response quality and satisfaction
  • Turned beta feedback into a prioritised iteration list
  • Built the commercialisation strategy: endorsements, alliances, distribution
  • Handed over the product and roadmap

The Results

A soft-launched product with a 27% sign-up rate and actionable feedback from beta testers. Users rated the responses and the natural language handling highly, and asked for the specific changes that shaped the next release, from how an answer is paced to how the product asks a clarifying question.

27%

sign-up rate from beta traffic

What it proved

Outreach at 10+ universities produced 200+ candidates in two weeks. Forty-plus went through the bootcamp; one team came out. Four months of mentored build produced an MVP chatbot in user acceptance testing, validated in focus groups and grounded in regional data. It soft-launched to real small businesses at a 27% sign-up rate, with next-release feedback in hand.

The authority finished with three things it lacked at the start: a working product with users, a commercialisation strategy backed by government endorsements and partner alliances, and a repeatable recruitment-to-launch pipeline. Feature expansion and distribution were scoped against beta evidence, not assumption. Advisory content it already published now answers a small business directly, at any hour, in regional context.

Results

MetricBenchmarkResultImprovement
Cohort recruitment time8–12 weeks2 weeks4–6× faster
Concept to soft launch12–18 months typical4 months mentored build3–4× faster
Candidate pipelinenone200+ sourced, 40+ shortlistedNew capability
Ventures soft-launched01, with beta users0 → 1
Beta sign-up rate4.61% landing-page benchmark27%5.9× better