xBites — Due Diligence Report | MISK Communication

Section 1

What this business actually is — in plain words

Stripped of the jargon, here is the entire business in five sentences.

The one-paragraph version Real-estate developers and brokers in Egypt collect enormous piles of leads from Facebook and Instagram ads — people who filled in a form saying “I’m interested in an apartment.” Most of those leads are never called back, or are called back two days later when the person has moved on. xBites sells a robot that chats with those leads on WhatsApp instead of a human: it answers instantly, in Arabic or English, at any hour, knows the prices and payment plans of every project, and works out who is a serious buyer and who is a tyre-kicker. It also does the reverse — it can message thousands of old, cold leads sitting in a CRM and try to wake them up. The broker pays a monthly subscription, roughly $400–$1,000, and gets a dashboard of who is worth calling.

Why anyone would want it

Their own pitch states the core problem accurately: a human agent making 2-minute calls can handle roughly 300 calls a day, which is about 260 hours a month of pure dialling. Response times run from 2 hours to 2 days. Every hour of delay kills conversion. A machine that replies in 4 seconds, 24/7, in the customer's language, to unlimited people at once, is a genuinely sensible product for this market. The problem they are solving is real.

How they make money

Model	What it means	Evidence it is actually being sold
Monthly SaaS subscription	Tiered by number of “inquiries” (one lead conversation = one inquiry). Deck shows $400 / $600 / $1,000 per month.	Confirmed — real signed contracts and paid tax invoices exist
Annual enterprise licence	Big clients pay upfront yearly. Coldwell Banker was quoted 520,000 EGP/yr + optional 420,000 EGP implementation.	Never closed — their CRM says “He will not continue with xBites”
Commission on closed sales	Take ~3% of property sales the AI sources. Appears in the financial model from year 3 at $175k/quarter.	Pure projection — zero commission revenue has ever been earned
“AI Interviewer” (HR)	Same engine pointed at recruitment — screens job candidates instead of property buyers.	Confirmed — $747 total from one client (Shaghalni)

The thing to understand before anything else This company did not start as a real-estate business. Its original product was called “Talk to your Data” — a chatbot that sits on top of a company’s internal documents so staff can ask it questions. That is what was sold to Fawry and Mobica in April 2025. The platform still runs on the subdomain ttyd.xbites.io — literally t-t-y-d, Talk To Your Data. The “MENA’s First AI Broker for Real Estate” positioning is a re-skin of a generic document-chatbot, pointed at property inventory. That is not disqualifying — plenty of good companies pivot — but it explains why the moat is thin and why the same engine can serve HR, real estate, or anything else.

Section 2

What they have actually built

We did not rely on the slides. We pulled apart their three demo videos frame by frame and read the live product screens. This is the most positive section of the report.

✓ What is genuinely there

A working multi-tenant platform at ttyd.xbites.io, organised into projects with tabs for WhatsApp, Integrations, Campaigns, Data Sources and Dashboard.
A real Egyptian property database — one table, unit_availability, holding 13,013 rows across 24 columns: developer name, project, unit type, price, finishing price, BUA, land area, garden/roof space, bedrooms, maintenance fees, parking, floor, city, governorate. This is a genuine data asset and took real work to assemble.
A functioning WhatsApp agent. In the demo it handles a live buyer conversation, compares two North Coast projects, and generates and sends a branded PDF comparison report inside the chat. That is a well-executed feature, not a mock-up.
Genuine white-labelling. The agent runs under client brands — “Darin AI” for one deployment, “REDi” for RED. The client’s customers never see xBites.
Real Arabic/English handling and a produced TV-style ad campaign that RED ran to promote their REDi agent.

✗ What is weaker than presented

The “analytics dashboard” is not their product. The Lead Buying Signals screen in their own demo is an embedded Google Looker Studio report (lookerstudio.google.com/reporting/…). It is a free Google BI tool with their data piped into it.
The demo dashboard contains 54 leads. Total. The pagination reads 1 – 54 / 54. The deck claims 40,000 inquiries and 2,000 qualified leads.
Every lead in it is graded “Cold.” Not one warm or hot lead appears in the flagship demo.
One of the 54 “leads” is inbound spam — an events agency pitching exhibition-stand design for Cityscape. The AI’s own summary reads: “no interest in real estate… showed no engagement in real estate discussions.”
The phone numbers are Saudi and UAE (+966, +971), not Egyptian — so the showcase deployment is not even in their claimed home market.

Fair assessment of the engineering Whoever built this is competent. The RAG-over-structured-inventory approach, the in-chat PDF generation, and the white-label tenancy model are all sensibly done. The product is not vapourware. The issue is not capability — it is scale, commercial traction, and honesty about both.

Section 3

Independent verification — our own research, not their files

Everything in this section was established from public sources with no reliance on any document they supplied: domain registries, the Singapore corporate register, Wayback Machine archives, Meta’s own developer documentation, Fawry’s published earnings release, university press, app stores and Egyptian job boards. Where a source was blocked, we say so rather than infer.

What survives independent verification — the complete list 1. XBITES PTE. LTD. is a genuine, live Singapore company, UEN 202451512W, incorporated 2024-12-24. 2. The Egyptian tax registration and ETA e-invoicing are genuine. 3. Google for Startups Accelerator: AI First MENA & Turkey, Class of 2025 — confirmed against Google Cloud’s own blog, which describes them as “the MENA region’s first AI-powered real estate sales assistant.” This is their one non-purchasable credential. 4. Egypt PropTech Challenge win at Cityscape 2025 — corroborated by two independent Arabic outlets. 5. The Antler investment is confirmed — xBites appears in Antler’s own public portfolio, and Preqin records “USD 180,000 in a pre-seed funding round from Antler on February 20, 2025.” 6. DarE Cohort IV selection is confirmed via a January 2026 announcement. 7. The website is real, professionally built and actively maintained. 8. The signed client contracts and ETA invoices are real. That is the entire list.

3.1 · The website tells a completely different story to the deck

Deck claim	What is actually on xbites.io
“Self-service platform… from setup to lead conversion in minutes”	There is no login, no dashboard, no signup, no docs, no API and no demo anywhere on the site. The single call-to-action on every page is a WhatsApp deep link to one Egyptian mobile number. There is no product surface a buyer can touch without going through a person.	Contradicted
Pricing: $400 / $600 / $1,000 per month	The words “price”, “pricing”, “plan”, “tier”, “subscription” and “/month” appear zero times on the homepage or the /for-business page.	Absent
Company website is `xbites.ai` (their own competitor sheet, and their TikTok handle)	That domain has never been registered. Authoritative registry returns “Domain not found”; RDAP returns 404 against a working control; no DNS records of any kind; zero Wayback captures ever. Only `xbites.io` exists. Anyone could register it tomorrow and impersonate them.	Does not exist
Named clients / case studies	Zero client names, zero case studies, zero testimonials on the entire site. Coldwell Banker — their flagship named pilot — appears zero times.	None published

The developer logo wall — read this carefully, it is not quite what it looks like Their homepage displays the logos of SODIC, Palm Hills, Orascom, Mountain View, Hassan Allam, Al Ahly Sabbour, Hyde Park and Madinet Misr. Crucially, these sit under the heading “Verified Data — From 130+ Developers.” They are being presented as sources of inventory data, not as customers — which is defensible and probably accurate. We should not accuse them of claiming these as clients on the website.

The problem is what happens when the same SODIC logo is then filed inside a folder named “Paying Clients” in the data room, backed by a “Living Expenses” mobile transfer. The website is honest about the distinction. The data room deliberately erases it.

3.2 · This is the third business run on the same shell

Wayback holds 70 captures of xbites.io going back to January 2023 — and the domain itself was registered on 2022-06-24, two and a half years before the Singapore company existed.

Date	What the site said	What the business was
2022 – 2024	“The TikTok for career advancers”	Edtech — bite-sized career videos
2025-01-02 9 days after incorporation	“xBites: Bite-Size Steps for Your Data Transformation Journey” — plus an AI Interviewer marked “COMING SOON”	Data/AI consulting + HR tech
2025-07-12	“AI Agents for Real Estate”	The pivot goes live
2025-10-08 → now	“MENA’s First AI Broker” → “Your Trusted AI Partner in Real Estate”	Current positioning

The real-estate business did not exist publicly until roughly six to seven months after the company was incorporated to pursue it. The January 2025 snapshot also shipped with unreplaced placeholder blog authors — “John Doe”, “Dina Doe”, “Adam Smith” — live in production.

The logo-wall technique is a documented repeat pattern The January 2025 version of the site displayed the logos of Zalando, Talabat, Xebia, Banque Misr and Intervu alongside a roster of ten “experts.” Those were the former employers of the consultants listed — not clients. The 2026 site uses the identical device with property-developer logos. In both cases the wording is technically defensible and the visual impression is engineered to read as a customer wall. Twice is a method, not an accident.

3.3 · Governance is materially worse than the data room suggests

The co-founder question is now settled

Mahmoud Khodier has left. His live LinkedIn profile lists him as “Head of AI Enablement & Productivity at Lucidya,” based in Amsterdam, with xBites removed from his profile entirely.

He nonetheless remains a registered director of the Singapore company on the public register, and the data room still ships his Founders’ Agreement, IP Assignment, Non-Compete and Form 45 as though he were present — while the team slide omits him and the cap table zeroes him.

Note on precision: the cap table supplied is an undated current snapshot. It shows Dina holding all 5,000,000 founder shares. It does not establish when the transfer occurred, and no transfer instrument was provided. Khodier was still signing client contracts as CEO on 25 February 2026. See §3.7 for the full timeline.

The board is hollow

The only two registered directors are the departed co-founder and a rented nominee — Lydia Chin Shi Ni, who is a Senior Operations Manager at Sleek, the corporate-services firm that incorporated the company.
Dina Helmy holds no board seat — despite being CEO, 79% shareholder, and the entire operating business.
Registered office is a virtual mailbox at 160 Robinson Road #14-04 — an address shared with roughly 9,690 other entities.
Paid-up capital: SGD 50.
A board resolution of 2025-01-23 to change the business activity code never appeared on the public register; ACRA still classifies them as generic software development.
The corporate documents supplied are a March 2025 snapshot — roughly 17 months stale.

3.4 · The leadership history in the deck is wrong

Verified by us directly against the primary source The Egyptian University of Informatics published an article on 24 April 2025 about a guest lecture by the xBites founders. It titles them, verbatim:

“Mrs. Dina Helmy, Founder & Chief Data Officer of xBites” · “Mr. Mahmoud Khodier, Founder & CEO of xBites”

The pitch deck given to MISK presents Dina Helmy as “Founder & CEO” and does not mention Khodier at all. The APACHE contract — created January 2026 — names “xBites PTE. LTD., represented by its CEO Mr. Mahmoud Khodier.” The company has rewritten its own leadership history without disclosing the change.

The same article is also the only public mention anywhere of the Fawry and Mobica relationships — and it is xBites describing itself, not either company acknowledging anything.

3.5 · “META Tech Provider” — true, but it is a licence, not a credential

Per Meta’s own developer documentation, Tech Provider is the bottom tier of four. Obtaining it requires a Meta app, business verification and app review — no revenue test, no headcount test, no badge, and no public directory listing. Every functioning competitor either holds it or cannot legally operate. Meta publishes no public list, which is precisely why the claim is unfalsifiable from outside and why it works on a slide.

In their favour — the claim is almost certainly genuine Their Coldwell Banker proposal states that the client “will pay these charges directly to Meta using the credit card associated with their Meta / WhatsApp Business Account.” That is exactly the Tech Provider billing model; a higher-tier Solution Partner would resell on its own credit line. Their commercial structure fingerprints as a real Tech Provider. They under-evidenced a true claim and over-marketed a weak one — a judgment signal, not fraud.

3.8 · The funding, checked against public databases

Claim	Independent finding
Antler investment	Confirmed twice over. xBites appears in Antler’s own public portfolio (one entry among 1,243 companies). Preqin states: “The company successfully raised USD 180,000* in a pre-seed funding round from Antler on February 20, 2025, which marks its first deal.”*	Verified
“$230,000 raised to date”	Across Preqin’s entire profile, the figure 230,000 appears zero times and 50,000 appears zero times. The only dollar figure any public database carries for this company is $180,000. Preqin’s total-raised field is paywalled, so this neither confirms nor refutes the combined figure — but no public source supports $230,000.	Unsupported publicly
Dar Ventures $50,000 grant	Programme selection is confirmed — DarE Cohort IV, powered by Dar Ventures and operated by Acasia, announced January 2026, thanking Zeina Mandour (the same person who executed their grant agreement). But no public source states any amount. The $50,000 rests on the executed DocuSign agreement and the two $25,000 tranches visible in their own books — which we consider adequate evidence.	Adequately evidenced
Antler ARC “follow-on”	Antler discloses no amount anywhere on its site. Consistent with our reading that the ARC is a conditional letter, not committed capital.	Not capital

Two details worth noticing Antler tags them as a Saudi Arabia company, not Egyptian — Antler’s portfolio filter has no Egypt option at all, which fits the KSA entity in their books. More curiously, Preqin lists their website as xbites.lovable.app — a Lovable subdomain, the AI app-generation tool. That may simply be a stale field, but it is consistent with an early product assembled on a no-code AI builder, and it is worth asking about.

3.7 · The people — and what happened to the CEO

The single most important correction to the story we were told Mahmoud Khodier was not a departing minor co-founder. He was the CEO and the entire commercial engine of the business. He signed the Antler SAFE. He signed the Dar Ventures grant. He signed every single client contract in the data room — Fawry, Mobica, Melee, Coldwell Banker, APACHE — the last of them on 25 February 2026. Dina Helmy was Chief Data Officer, corroborated independently by the EUI article, by her own CV, and by the Founders’ Agreement.

He is now Head of AI Enablement & Productivity at Lucidya, in Amsterdam, with xBites removed from his profile.

Pinning the departure — and a detail our own reading of their books confirms

Marker	Date	Source
Last post on the xBites company page — then silence	2025-12-24	LinkedIn
Last xBites-related post on his own profile	2026-01-27	LinkedIn
Last client contract signed by him as CEO	2026-02-25	Melee LOI, data room
Anomalous payroll spike — $20,000 against an ~$8,000 monthly run-rate	May 2026	Their own bookkeeping sheet — we verified this line ourselves
First post at his new employer, Lucidya	2026-06-29	LinkedIn

The departure therefore falls between late February and late June 2026 — most likely March–May, with that $20,000 May payroll spike the probable settlement. We flag this as a well-supported inference, not a documented fact.

Why he may have forfeited the shares — and why that is not reassuring The Founders’ Agreement in the data room defines a “Bad Leaver” as a founder who resigns, with forfeiture consequences, and Khodier also signed a non-compete. If his move to Lucidya triggered that clause, his 2,500,000 shares would revert — which would neatly explain the cap table MISK was shown.

That is a coherent story. It is also entirely undocumented. The data room contains no ACRA cessation filing, no share transfer instrument, no leaver notice, and no board resolution appointing Dina as CEO or even as a director. The last authoritative public filing still shows Khodier holding 50% and sitting as the sole substantive director, with Dina not a director at all.

Put bluntly: if Khodier still holds those 2,500,000 shares, the cap table MISK was given does not exist. A fresh ACRA business profile costs about SGD 5.50 and settles the question in five minutes. Nothing should be signed or paid before that document and the executed transfer instrument are in hand.

How many people actually work here

Every independent signal converges on one to three

Three named team members appear in the last public roster (January 2026): Dina Helmy, Ahmed Hossam, Abdelkareem Elkhateb.
$400 total payroll in July 2026, $0 forecast from August — from their own books.
Zero ESOP has been allocated to anyone — the cap table shows the entire 10% pool unallocated.
Cursor, Langchain and Sentry subscriptions were all cancelled during 2026 — we verified these lines going to zero ourselves. Development tooling is switched off.
One public code repository, untouched since April 2025.
No company post in roughly seven months; no job listings anywhere; no careers page.
LinkedIn shows a 2–10 band with 6 associated profiles.

There is no engineering team. There is one founder, and freelancers.

Two specific names worth raising

“Mahmoud Ali, Sales Director” is one of only two faces on the pitch deck’s team slide. He appears in none of the three public team rosters, has no verifiable employment history at xBites, and is flagged part-time in their own team sheet. He is confirmed to be a different person from Mahmoud Khodier. Ask who he is.

Abdelkareem Elkhateb, their one full-time AI engineer, is genuinely strong — 34 published models, active as recently as 23 July 2026. But his public profile is marked “Open to Work” and lists no employer.

Seven further roster members could not be verified either way — LinkedIn blocks anonymous profile access. We report them as unverified, not as absent.

A pattern, not an accident The January 2025 site listed a ten-person team; every one of those people has since vanished from the company’s web presence, and xbites.io names no humans at all today. Across four pivots in three and a half years, the constant is a bench of loosely-attached freelancers presented as a team. The eleven-name team sheet in this data room — eight of eleven marked part-time, with two rows literally annotated “Used to work with us” and “freelance” — is the same device, one pivot later.

3.6 · The silence

What we could not find, anywhere

Zero press of any kind after October 2025 — nine months of silence while raising.
Zero lifetime coverage in MENAbytes, Wamda or WAYA — the three principal MENA startup outlets. All three returned explicit no-results.
Zero public record of any funding. Neither the Antler SAFE nor the Dar Ventures grant appears in any database. Antler’s portfolio page is paginated and could not be fully enumerated — absence there is inconclusive and we do not treat it as evidence.
Zero job postings on Indeed Egypt, Wuzzuf, Jobzella or Tanqeeb; no careers page exists.
No mobile app on either store — consistent with a WhatsApp-native design, but it means no independently observable usage signal exists.
No complaints, disputes or litigation found in English or Arabic. Nothing negative surfaced at all.

LinkedIn: real but dormant

Followers	1,976
Employees on LinkedIn	6
Total posts	9 in ten months
Most recent post	~late Jan 2026

Of nine posts, essentially all are accolade announcements — accelerator selections, award wins, summit appearances. Zero customer wins, zero traction metrics, zero hiring. For a company whose entire external narrative is awards, the awards stopped in January and so did the posting.

A trap to avoid if anyone else researches this A Glassdoor page for “Xbite” (2.3/5, 21 reviews) and a Trustpilot page for xbite.co.uk both belong to Xbite Ltd, an unrelated UK e-commerce company that dominates the search results for this name. Other live collisions include Bits x Bites (a China food-tech VC, which also has a Singapore entity currently contaminating registry-aggregator data for xBites) and XBites Jerky in Ohio. None of these are relevant. Do not let a negative review of a British retailer end up in a decision about an Egyptian startup.

Section 4

The money — what is actually real

This is the heart of the matter. We reconciled three independent sources: their internal bookkeeping, every Egyptian Tax Authority e-invoice in the folder, and the payment screenshots. Two of the three agree closely. That is what makes the conclusion safe.

Important credit where due — the invoices are genuine Every invoice in the data room is a real Egyptian Tax Authority (ETA) e-invoice carrying a government-issued UUID (e.g. E3C61BW998AM7VQXSEDHD2SK10), a verifiable status field, submission and issuance timestamps, and the issuer’s tax registration #772184283. These are filed with the state and are not trivially faked. Their bookkeeping is honest. The dishonesty is in the pitch deck, not the ledger.

Every invoice they have ever issued

Client	Issued	EGP	USD @47	ETA status	Product
Shaghalni	2025-11-17	11,700	$249	Valid	AI Interviewer (HR)
RED — Real Estate Domain	2025-11-19	59,245	$1,261	Valid	AI Sales Agent (REDi)
Shaghalni	2025-12-08	11,821	$252	Valid	AI Interviewer (HR)
Shaghalni	2026-01-19	11,807	$251	Valid	AI Interviewer (HR)
Platinum Real Estate	2026-05-20	50,000	$1,064	Valid	AI Sales Agent
Connect Homes	2026-05-20	50,000	$1,064	Valid	AI Sales Agent
Mastermind Real Estate	2026-06-10	20,000	$426	Cancelled	AI Sales Agent
UREEG	2026-06-18	20,000	$426	Valid	AI Sales Agent
TOTAL — valid invoices only		214,573	$4,565

Cross-check against their own books

Their Historical Financials.xlsx bookkeeping sheet records, month by month, from January 2025 to September 2026:

AI Sales Agent — subscription	$4,668
AI Interviewer — HR	$747
Lifetime revenue, 21 months	$5,415
Average per month	$258

Two independent sources — government invoice records ($4,565) and internal books ($5,415) — agree within 16%. The residual gap is explained by an unpaid September RED invoice and FX timing. This number is reliable.

Only 12 of the last 21 months had any revenue at all

The revenue line reads, in dollars, month by month:

Jan–Oct 2025: 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
Nov 2025: 1,499 · Dec 2025: 249
Jan 2026: 249 · Feb–Apr 2026: 0, 0, 0
May 2026: 218 · Jun 2026: 2,800 · Jul 2026: 400

The company earned nothing at all for its first ten months, and its single best month in its entire history was $2,800.

The part that should end the conversation: they are out of money

From their own cash model — not our estimate

$771

projected bank balance,
September 2026

Cash has fallen from $127,500 (Dec 2024) to $10,325 (Nov 2025), recovered briefly on the second Dar Ventures tranche, and then collapsed. Critically, their own payroll line runs:

Jun 2026 $2,840 → Jul 2026 $400 → Aug 2026 $0 → Sep 2026 $0

They have modelled their own team to zero from August 2026. That is this month. Read alongside the $771 closing balance, this is not a company raising to grow — it is a company that has run out of runway.

Where the money actually went — a governance problem Their books show the Egyptian company bank account (CIB) holding 0.00 in every single month except December 2025 ($2,602). Meanwhile the “proof of payment” screenshots show client money arriving by personal InstaPay transfer into the founder’s private account, categorised as “Living Expenses.” Small Egyptian B2B deals do often settle by InstaPay, so this is not automatically improper — but revenue bypassing the corporate account, combined with a cancelled tax invoice for the matching amount (see Mastermind, §5), is a real governance defect that any investor or acquirer would have to resolve before signing anything.

Section 5

The client list — real customers, logos, and templates

The data room presents a folder literally named “Paying Clients” containing eight companies, plus four “Pilots.” We tested each one against contracts, invoices, payment evidence and their own CRM. Only two survive as genuine, currently-paying, real-estate customers.

Presented as	Client	What the evidence actually shows	Verdict
Paying	RED (Real Estate Domain)	Valid ETA invoice 59,245 EGP. White-labelled “REDi” agent. RED produced their own TV-style ad campaign for it. Genuine, marketed, referenceable relationship — their best customer by far. But their CRM (2026-07-01): “currently developing an application for them… no time to discuss anything new.”	Real — churning
Paying	Shaghalni	Three valid invoices, ~11.8k EGP each, Nov 25 – Jan 26. Genuine paying customer. But this is the HR product, not real estate — and payments stopped after January 2026.	Real — lapsed
Paying	Platinum Real Estate	Valid ETA invoice, 50,000 EGP, 2026-05-20. No contract in the folder, no payment proof.	Invoiced
Paying	Connect Homes	Valid ETA invoice, 50,000 EGP, 2026-05-20. No contract, no payment proof.	Invoiced
Paying	UREEG	Valid ETA invoice 20,000 EGP. Underlying document is an explicitly non-binding pilot LOI with a free first month and a 40% discount, unsigned signature blocks.	Discounted pilot
Paying	APACHE	Signed AI Sales Agent Agreement (first month free, then 50% off). Payment screenshot of 10,900 EGP matches their Tier-3 discounted price exactly. But their own CRM says: “No more business with them.”	Churned
Paying	Mastermind	ETA invoice for 20,000 EGP is CANCELLED. The matching 20,000 EGP arrived as a personal InstaPay transfer from the owner, Ahmed Wael, to the founder’s private account, tagged “Living Expenses.” Underlying LOI is a free, non-binding pilot.	Not a clean sale
Paying	SODIC	No contract. No invoice. The only evidence is a screenshot of a 20,000 EGP personal InstaPay transfer from “MAHMOUD A” to the handle `mar.moha@instapay`, dated 2026-07-18, categorised “Living Expenses.”** Nothing in it references SODIC. SODIC appears in xBites’ product only as a developer name inside the property database (“June by SODIC”).	Not a client
Pilot	Coldwell Banker Egypt	Genuine e-signed free-pilot LOI (2,000 inquiries, 4–6 weeks, non-binding) and a 520,000 EGP/yr proposal. Their CRM: “He will not continue with xBites.”	Lost
Pilot	Fawry EGP 8.65bn revenue FY25	A genuinely executed contract — Dropbox Sign audit trail, 15–16 Feb 2025, countersigned by Alaa Abd Elgwad Ghania, Fawry’s serving Chief Product Officer. But it is for the old “Talk to your Data” product at $300/mo, and it produced $0 across its full one-year term. Fawry appears zero times in xBites’ own 285-row CRM. Decisively: Fawry’s FY2025 earnings release (5 Mar 2026) tells the market it expects to launch its “proprietary LLM-powered chatbot by year-end.”	Signed, never activated
Pilot	Mobica ~2,500 staff, 9 factories	Also genuinely executed, same date, same product, same $0 outcome. Countersigned by Farouk Mohamed Farouk — whom the contract calls “Deputy CEO” but who is actually Vice Chairman. Mobica’s live site runs no chat AI of any kind and lists no WhatsApp contact at all. Marked “Freeze” in xBites’ own pipeline.	Signed, never activated
Pilot	Melee	E-signed free non-binding LOI, up to 1,000 leads, no fee. CRM: “called CEO 2/6 no reply.”	Free pilot

The SODIC problem, stated plainly SODIC is one of Egypt’s largest listed developers. It is the single most impressive name in the data room, and it is placed in a folder called “Paying Clients.” The supporting evidence is a peer-to-peer mobile money transfer between two private individuals, labelled “Living Expenses,” with no reference to SODIC anywhere in it. Presenting that as a customer relationship is not an accounting slip — it is a deliberate choice about how to dress a data room. This single item should govern how much weight you give every other claim they make.

What their own CRM says — the most honest document in the folder

The Pipeline.xlsx file appears to have been included by mistake. It is their internal sales tracker: 44 “Active,” 6 “WON,” 1 lost to a competitor. The notes are candid and directly contradict the data room around them.

“He will not continue with xBites.” — Coldwell Banker

“No more business with them.” — Apache, filed under status WON

“They are currently developing an application for them. No time to discuss anything new.” — RED, their biggest client

“He does not understand the products, he says that it is the same as chatbot, he is completely lost.” — Alashraf, prospect

“Trying a similar product now (Not AI), check back by September.” — Gate Real Estate

Source: 13-Clients/Pipeline/Pipeline.xlsx, 84 rows, exported 2026-07-25.

Section 6

Corporate structure and the missing co-founder

The corporate paperwork is professionally done — and it contains a contradiction the data room never explains.

The structure (all verified from filings)

HoldCo: XBITES PTE. LTD., Singapore, UEN 202451512W, incorporated 2024-12-24, status LIVE. Registered at 160 Robinson Road #14-04 (a corporate-services address).
OpCo: xBites, Egypt. Tax registration 772184283, activity code 6209 (other IT services), activity start 2025-02-08, Nasr City First tax office. Registered at a virtual office in Capital Mall, Fifth Settlement — the folder contains the virtual address contract.
Directors at incorporation: Mahmoud Khodier (Netherlands) and Lydia Chin Shi Ni (Singapore, nominee).
A Saudi entity was also set up — the books show a $4,650 formation cost in Sept 2025 and $533/month running cost thereafter.

Note: the Egypt tax card records no commercial-register number and no social-insurance number, i.e. no formally employed staff at the Egyptian entity. A separate commercial register document dated 2025-03-04 does exist in the folder.

The contradiction

At incorporation (2024-12-24), the company was a clean 50/50:

Certificate	Holder	Shares
No. 1	Mahmoud Khodier	2,500,000
No. 2	Dina Mohamed	2,500,000

The cap table given to MISK shows:

Holder	%	Shares
Dina Helmy	79%	5,000,000
Antler MENAP	11%	696,203
ESOP (unallocated)	10%	632,911

Khodier’s entire 2,500,000 shares have moved to his co-founder, and he does not appear anywhere in the current materials.

Why this matters more than a normal founder departure Founders leave; shares get bought back. That is ordinary. What is not ordinary is the evidence trail. Mahmoud Khodier was: the holder of Share Certificate No. 1; the sole executive director of the Singapore entity; the signatory on the Fawry agreement, the Melee LOI and the Coldwell Banker LOI; and is named as “xBites PTE. LTD., represented by its CEO Mr. Mahmoud Khodier” on the APACHE contract, a document created as recently as January 2026. Six months later he has been erased from the cap table.

The data room contains no share transfer form, no board resolution approving the transfer, no separation agreement, and no ACRA filing after January 2025 to evidence any of it. The Founders’ Agreement in the same folder defines a “Bad Leaver” as a founder who resigns, with forfeiture consequences — so a mechanism exists. But whether it was properly executed is undocumented, and an unresolved 50% founder claim is the kind of thing that blocks a funding round or a partnership outright.

Personal identifiers (passport numbers and both founders’ residential addresses, which appear in the Founders’ Agreement and share certificates) have been deliberately REDACTED from this report.

Section 7

Funding — what “$230,000 raised” really means

Both funding instruments are genuine executed documents. The headline number is nonetheless overstated, and one important claim is presented as money when it is only an option.

Claim	Document	What it actually says
$180,000 SAFE from Antler	Antler SAFE Note, dated 15/02/2025	Genuine, and independently confirmed in Antler’s public portfolio and on Preqin (“USD 180,000… on February 20, 2025”). Purchase amount $180,000, post-money valuation cap $1,636,363.64. But: the SAFE states payment is “contingent upon the Company entering into Investor’s standard services agreement with Antler Ventures Ltd.” Their own balance sheet shows only $127,500 cash at Dec 2024 — implying roughly $52,500 (29%) was retained as programme fees rather than received as cash.
$50,000 grant from Dar Ventures	DocuSign-executed Grant Agreement, 5 Nov 2025	Genuine and fully received — both $25,000 tranches appear in the books (Dec 2025, Apr 2026). Non-equity. Programme term ended 10 March 2026. Carries a call option for Dar to invest up to $250,000 (inclusive of the grant) for up to 5.5% at a $4.5M cap — exercisable only if noticed before 10 May 2026, a date that has now passed.
Antler “follow-on” (ARC)	ARC Term Sheet & Letter of Support	Not committed money. It is an expressly conditional letter: Antler “may” invest up to $250,000, only alongside a round led by third-party professional investors, subject to confirmatory diligence, and terminable at Antler’s sole discretion. It should never be counted as capital.
“Google for Startups Accelerator Grad — $250,000 in credits”	Pitch deck claim	See §3 for our independent verification. Note that cloud credits are not capital — they cannot pay salaries. They do, however, explain why the books show $0.00 LLM cost every single month, which in turn means the company has never paid its true cost of goods.

The honest funding number Against a headline of “$230,000 raised to date,” the actual cash that ever reached the company is approximately $177,500 ($127,500 net from Antler + $50,000 from Dar Ventures). Of that, roughly $176,700 has been spent, overwhelmingly on payroll, to generate $5,415 of lifetime revenue. That is a burn-to-revenue ratio of about 33 : 1.

Section 8

The market claims versus their own spreadsheet

We re-derived their market sizing from their own file rather than accepting the slide. The deck does not match the model behind it.

Claim: “$100M Obtainable Market”

Their own Market Size.xlsx bottom-up model produces:

Egypt	$36,369,000
UAE	$2,183,760
Saudi	$3,821,580
Total	$42,374,340

The deck states $100M. Their own workbook says $42.4M. That is a 2.36× overstatement of the headline market number.

The model rests on an impossible input

The Egypt figure is built as: 100,000 active brokers × 5 projects/day × 10 questions/project = 5,000,000 inquiries per day, or 150 million per month.

For scale: that would be roughly 1.4 inquiries per month for every man, woman and child in Egypt — purely about buying new-build property. The same sheet elsewhere states Egypt has 500,000 brokers and separately 40,000 brokerage firms, contradicting its own 100,000 figure.

Three incompatible pricing models in one data room

Source	Entry tier	Inquiries included	Implied price / 1,000 inquiries
Pitch deck	$400/mo	2,000	$200
Market Size model (Egypt)	$299/mo	10,000	$29.90
APACHE signed contract	7,650 EGP (~$163)/mo	500	~$326

The TAM is calculated using the middle row — the cheapest per-inquiry pricing, which they do not actually charge anyone. The contracts they really sign are up to 11× more expensive per inquiry than the model assumes, which would collapse the addressable-volume assumption if applied consistently.

The cost assumption that invalidates the margin story The model assumes a fully-loaded cost of $0.003 per inquiry including WhatsApp, producing headline gross margins of 83%. Yet their own Coldwell Banker proposal concedes that Meta charges approximately 5 EGP (~$0.106) per outbound conversation — 35× their modelled cost — and solves the problem by pushing that fee onto the client’s own Meta account. On the 1,000,000-lead reactivation they proposed to Coldwell Banker, that is roughly 5,000,000 EGP (~$106,000) of Meta fees the client pays directly, against a 520,000 EGP licence. The economics only work because the largest cost line is someone else’s problem.

Section 9

Could we just build this ourselves?

The question you actually asked. Short answer: yes — most of it, quickly. But one component is genuinely hard, and it is not the AI.

The honest one-line answer The technology is almost entirely commodity in 2026. The property database is not. A competent 2–3 person team could rebuild the software in roughly 6–10 weeks. Rebuilding the 13,013-unit inventory database, and keeping it current as developers change prices and payment plans every few weeks, is the part that would actually take real time and ongoing cost — and it is the only thing here resembling a moat.

Component-by-component

Component	What it takes in 2026	Difficulty
The conversational AI itself	An off-the-shelf frontier model with a good Arabic capability. No training, no fine-tuning, no proprietary model. Their own cost sheet lists `Langchain` at $80/month — a standard open-source orchestration framework.	Commodity
Property knowledge / RAG	Retrieval over a structured table. Their own schema is a single 24-column table. This is a well-trodden pattern with mature tooling.	Commodity
WhatsApp channel	WhatsApp Cloud API. Requires registering as a Meta Tech Provider — which, per Meta’s own documentation, is a self-service application requiring business verification and app review. No revenue test, no headcount test, no badge. It is a form, not a barrier.	Admin, ~2–4 weeks
Campaign / broadcast engine	Template management, scheduling, throughput handling against Meta’s messaging tiers. Genuinely fiddly, well-documented, no novel engineering.	Moderate
Analytics dashboard	They did not build one. Their own demo shows an embedded Google Looker Studio report — a free tool. We would do exactly the same, in a day.	Free
In-chat PDF comparison reports	The nicest feature in the product. Templated HTML → PDF → WhatsApp media message. Genuinely good product thinking; a few days of work.	Easy
Multi-tenant white-labelling	Per-client config, branding, isolated knowledge bases. Standard SaaS architecture.	Moderate
The 13,013-unit property database	This is the hard part. Developer-by-developer collection of inventory, pricing, payment plans and availability across the Egyptian primary market — then keeping it fresh. It is unglamorous, manual, relationship-dependent, and never finished.	The real barrier
Distribution into developers & brokers	The second hard part — and the one MISK already has and xBites does not.	The real asset

MISK engineering estimate — labelled as an estimate, not a verified figure Based on the component breakdown above, a credible internal build:

Scope	Team	Elapsed
Working MVP — one client, inbound WhatsApp qualification over a supplied inventory sheet	2 engineers	3–4 weeks
Production multi-tenant SaaS with campaigns, white-labelling and dashboards	2–3 engineers	6–10 weeks
Inventory database at xBites’ current scale, plus a maintenance process	1–2 data staff, ongoing	3–6 months, then permanent

This is our judgement from the architecture we observed, not a quoted figure. Treat it as directional.

The economics of the channel — and why the margin story is fragile

Per Meta’s own pricing documentation, WhatsApp moved to a per-message billing model effective 1 July 2025, deprecating the older per-conversation model. Marketing template messages — exactly what a cold-lead reactivation campaign consists of — are billed individually.

Run their own Coldwell Banker proposal through this They proposed reactivating Coldwell Banker’s database of “over 1 million contacts.” Their own document discloses Meta charging ~5 EGP per outbound conversation initiation, payable by the client directly. At that disclosed rate, messaging one million leads once costs roughly 5,000,000 EGP (~$106,000) in Meta fees alone — against an annual licence of 520,000 EGP. The channel cost is ~10× the software fee, and under per-message billing a multi-touch sequence multiplies it further.

This is not a criticism of xBites specifically — it is the structural economics of WhatsApp outbound in 2026, and it applies equally to anything MISK builds. It is the single most important number in any business case here, and it is why the vendor’s modelled cost of $0.003 per inquiry is not a rounding error but a category mistake.

The legal problem nobody in the data room has priced

Egypt’s Data Protection Law reaches full enforcement on 31 October 2026 — about 14 weeks from now We verified this ourselves against primary sources. Egypt’s PDPL executive regulations took effect 1 November 2025 with a 12-month grace period and full enforcement on 31 October 2026.

The regulator’s own Electronic Direct Marketing Guidelines, published on pdpc.gov.eg, state verbatim:

“Transitional Provisions for Existing Customers: the ‘soft opt in’ mechanism… creators may rely on the ‘soft opt in’ mechanism to lawfully continue sending EDMCs to existing customers. This approach is an exemption to obtaining valid consent in situations where there is an existing customer relationship established through a prior sale of a product or payment for a service…”

Two things follow, and both are fatal to the campaign xBites proposed to Coldwell Banker:

It is transitional, not permanent — it expires with the grace period on 31 October 2026.
An unconverted lead is not a customer. The exemption requires a prior sale or payment. A database of “over 1 million contacts” who once enquired about an apartment contains only a small fraction who ever transacted. For everyone else, explicit consent is required.

And there are two separate licences — one of which falls on the platform operator, not the client The same regulator guideline requires, verbatim:

“Obtain an EDM license or permit titled ‘EDM for Self’ prior to initiating any EDMC” · “Obtain an EDM license or permit titled ‘EDM for others’ prior to initiating any…”

In the structure xBites proposed, Coldwell Banker is the “creator” and xBites is the “sender.” xBites therefore needs its own “EDM for others” licence and independent consent record-keeping. This is not a client-side problem the vendor can disclaim in a proposal — and their proposal does exactly that, addressing only the Meta fees.

They are pitching a one-million-lead outbound blast roughly fourteen weeks before full enforcement, with no evidence of holding the licence that makes them lawful to send it. That is a direct question to put to them.

⚠️ This lands on MISK too — and it changes how we would build If MISK builds this capability and runs broadcast campaigns on behalf of property clients, MISK becomes the “sender” and needs the same “EDM for others” licence. It does not change the build-versus-buy answer, but it is a lead-time item on the critical path, not a paperwork afterthought — and it is far cheaper to obtain than to acquire a company that may not hold one.

The architectural conclusion — and it is the most useful output of this whole report: build on Click-to-WhatsApp, not on cold broadcast. CTWA is user-initiated, which means it survives both Meta’s opt-in policy and the PDPC; it opens the 24-hour service window plus a 72-hour Free Entry Point in which messages are free; and it captures explicit, logged consent at first contact.

The correct design makes the six-figure messaging bill largely disappear. Which reframes the whole evaluation: the reason to decline the xBites proposal is not that it is expensive — it is that the one mechanism it sells is simultaneously the least compliant and the most costly way to do this.

Verified by us directly: the enforcement date, the soft opt-in text, and both licence names, from the PDPC’s own published guideline and a law-firm regulatory update. Relayed but not independently verified by us, and to be confirmed with Egyptian counsel before relying on it: the specific penalty bands (reported as EGP 500,000–5,000,000 for licensing violations), personal liability of the responsible manager, mandatory publication of convictions, potential imprisonment under Cybercrime Law 175/2018, and a reported EGP 10m civil damages precedent.

Is any of it innovative?

Where they are genuinely ahead

The Egyptian primary-market dataset. Real, laborious, and not downloadable from anywhere.
Arabic-dialect handling tuned for property sales — not novel technically, but non-trivial to get right and they have 18 months of conversation data behind it.
Compliant local invoicing through the Egyptian ETA e-invoicing system — genuinely annoying to set up and frequently underestimated.
The in-chat PDF comparison report. The one feature we would straightforwardly copy.

Where the “first / only” framing does not hold

Their own competitive sheet names four regional rivals — Coraly ($2M pre-seed), PropAI ($1.5M pre-seed), Seqoon ($500K+) and Realiste AI ($1M Series A) — all of them better funded than xBites, three of them founded the same year. Globally the category is mature and well-capitalised.

The claim “MENA’s First AI Broker” is a marketing line, not a defensible position — and Google’s accelerator blog repeats it because xBites supplied the description. Nothing in the technology is proprietary.

What this means commercially for MISK If MISK wants this capability, the build is well within reach and the result is owned outright. The rational reason to engage xBites at all is not the software — it is the inventory database and the 18 months of hard-won knowledge about why Egyptian developers do and don’t buy this. Both are acquirable as assets. Neither requires taking equity in a company with $771 in the bank and an unresolved founder claim.

Section 10

What they want from MISK — and what they are offering

No written ask accompanied the folder, so we inferred the intent from what they chose to send. The composition of the data room is itself the answer.

What the shape of the folder tells us They did not send a product brochure, a rate card, or a partnership deck. They sent a complete investor data room: cap table, corporate structure, incorporation documents, share certificates, board resolutions, ESOP, founders’ agreement, IP assignments, non-competes, transaction documents, historical financials, a 3-year financial model, market sizing, competitive landscape — and a founder’s passport. That is a fundraising package, not a sales package. Nobody assembles this to sell a $400/month subscription.

The three possible asks, assessed

Possible ask	How likely	What it would mean for MISK	Our read
1 · Investment / equity	Most likely	Cheque into a company with $771 of cash, $258/month of revenue, an unresolved 50% founder claim, and payroll modelled to zero from this month.	Decline. At any price. The governance gap alone is disqualifying before the numbers are considered.
2 · Partnership / reseller	Likely secondary	MISK resells or white-labels the AI agent into its own client base — which is exactly the distribution xBites cannot buy and desperately needs.	This is the only ask with any logic to it — but the logic runs in their favour, not ours. MISK would supply the scarce asset (client relationships) in exchange for software we could commission ourselves. See §9.
3 · MISK as a customer	Least likely	MISK buys the agent for its own lead handling at $400–1,000/month.	Possible but odd — you do not send your cap table and passport to a prospect. If this is the ask, the price is defensible but the vendor risk is not: a supplier with $771 in the bank may not exist in six months.

What they genuinely offer

Assets that have real value

The 13,013-unit property database — developer, project, pricing, payment and unit-level detail across the Egyptian primary market. Assembling and maintaining this is genuinely laborious and is the closest thing here to a moat.
A working, white-labelled, Arabic-capable WhatsApp agent that is already deployed under at least two client brands.
A live Egyptian tax-registered entity able to invoice local clients compliantly through the ETA e-invoicing system — non-trivial and often underestimated.
Real category experience — they have run the sales motion into Egyptian developers and brokers for 18 months and know exactly why it stalls.

What they do not have

Distribution. 44 “active” pipeline entries, almost all stalled on “follow up,” “no reply,” or “waiting for internal decision.”
Proof the product retains customers. Every substantial client has churned, lapsed, or is churning.
A team. Payroll goes to zero next month; 8 of 11 listed staff were already part-time or freelance.
Defensible technology. See §9.
Runway to negotiate. This is the single most important commercial fact in the report.

Section 11

Red flag ledger

Ranked by severity. Every item is sourced from documents they supplied themselves.

SODIC presented as a paying client on the evidence of a personal “Living Expenses” transfer

No contract, no invoice, no reference to SODIC anywhere in the supporting file. The most recognisable logo in the deck rests on a peer-to-peer mobile payment between two individuals. This is the item that reframes everything else.

A 50% co-founder has been erased from the cap table with no supporting documentation

Mahmoud Khodier held Share Certificate No. 1 (2,500,000 shares), was sole executive director, and was named CEO on a contract created in January 2026. No transfer form, board resolution, separation agreement, or post-January-2025 ACRA filing explains his disappearance. An unresolved founder claim is a transaction-blocker.

The company is out of cash and has modelled its own payroll to zero

$771 projected closing balance; payroll $2,840 → $400 → $0 → $0 across Jun–Sep 2026. They are not raising to scale; they are raising to survive, and the data room does not say so.

Traction claims overstated by roughly three orders of magnitude

Deck: “40,000 inquiries, 2,000 qualified leads.” Their own demo dashboard: 54 leads, all Cold, one of which is inbound spam. Deck: “Product Market Fit.” Reality: $258/month average revenue and universal churn.

Churned and lost accounts filed as current clients

Apache is filed under “Paying Clients” while their CRM reads “No more business with them.” Coldwell Banker is presented with a full financial proposal while their CRM reads “He will not continue with xBites.”

Two dead contracts with blue-chip counterparties presented as live “pilots”

The Fawry and Mobica agreements are genuinely executed — real Dropbox Sign audit trails, real C-suite signatories — but they are for the discontinued “Talk to your Data” product and produced exactly $0 over a full one-year term, confirmed against xBites’ own books. Fawry does not appear once in their 285-row CRM; Mobica is marked “Freeze.” Fawry has since told the market it is building its own LLM chatbot in-house. Filing these under a traction narrative is the misrepresentation.

Revenue routed around the corporate bank account

The Egyptian company account shows 0.00 in every month but one, while client payments arrive by personal InstaPay into the founder’s private account tagged “Living Expenses” — in the Mastermind case, against a tax invoice that was cancelled.

Pitching a 1,000,000-lead outbound blast ~14 weeks before Egypt's data-protection law reaches full enforcement

Egypt's PDPL is fully enforced from 31 October 2026. The regulator's "soft opt-in" is transitional and requires a prior sale or payment — unconverted leads do not qualify. As the "sender," xBites requires its own "EDM for others" licence; nothing in the data room evidences one, and their proposal addresses only the Meta fees. Verified by us against the regulator's own published guideline.

Market size overstated 2.36× against their own model

Deck says $100M obtainable; the workbook behind it computes $42.4M, itself built on an implausible 5,000,000-inquiries-per-day input and contradicted by two different broker counts in the same sheet.

Gross margin built on a cost assumption 35× below their own disclosed figure

Model: $0.003/inquiry. Their Coldwell proposal: ~5 EGP (~$0.106) per Meta conversation. Books show $0.00 LLM cost every month — they have never paid true COGS, having run on Google credits.

“$230,000 raised” overstates actual cash by ~23%

Approximately $177,500 reached the company; the Antler SAFE was partly consumed by programme fees. The Antler ARC “follow-on” is a discretionary, terminable letter of support and is not committed capital.

Data room assembled the day before it was sent, with a section missing

All eight invoice PDFs were generated within a four-minute window on 2026-07-25; the deck and product slides were built in Canva on 2026-07-22. Innocuous on its own. Less so alongside the numbering gap — folders run 01–04 and 06–14, with 05 absent.

Mitigating: the underlying bookkeeping is honest

In fairness — their internal financials, invoices and CRM all tell the same, accurate, unflattering story. They did not falsify the ledger. They dressed the narrative on top of it, and then included the documents that contradict it.

Section 12

Conclusion and recommended response

Recommendation

Do not invest. Do not sign a reseller agreement on their terms. If the product is genuinely wanted, wait — and buy the asset, not the equity.

The honest summary is that xBites is a competent piece of engineering wrapped in a dishonest pitch, attached to a company that is out of money. Amr’s instinct on first reading was directionally right, though the reason is more interesting than “it’s a scam.” It is not a scam. It is a real, failing startup that has chosen to present its failure as traction — and, revealingly, included the very files that disprove the presentation.

Where they were straight with us

The company, tax registration and corporate filings are genuine.
The invoices are real government e-invoices, and the bookkeeping matches them.
The Antler SAFE and Dar Ventures grant are real executed instruments.
The product works and is better engineered than the traction suggests.
They disclosed the Meta per-conversation cost honestly in the Coldwell proposal.

Where they were not

A non-client (SODIC) presented as a paying client.
Churned accounts (Apache, Coldwell) presented as live.
Unsigned templates (Fawry, Mobica) presented as pilots.
54 cold leads presented as 40,000 inquiries and “Product Market Fit.”
A vanished 50% co-founder, undocumented.
Imminent insolvency omitted entirely from the narrative.

If MISK still wants the capability

The leverage position Their weakness is precisely MISK’s strength. They have software and no distribution; MISK has client relationships and no software. But they are the ones with $771 in the bank and payroll at zero this month — which means MISK negotiates from a position of near-total leverage, and there is no reason whatsoever to move quickly. If the capability is genuinely wanted, the sensible structures, in order of preference:

Wait 60–90 days. On their own model they cannot operate past August. Distressed terms — or an asset purchase of the database and codebase — will be dramatically cheaper than anything on offer today.
Buy the asset, not the company. The 13,013-unit database plus the codebase, on a clean asset-purchase basis, sidesteps the unresolved founder claim entirely. Never acquire the equity while Khodier’s 2,500,000 shares are unexplained.
Commission the build. See §9 — a small team reproduces the core for a fraction of any realistic equity ask, and MISK owns it outright.
Revenue-share pilot only. If they must be engaged now: no fee, no equity, no exclusivity, one MISK client, 60 days, MISK owns the client relationship and the data. Terminable at will.

Conditions precedent — if any deal is contemplated at all

Written explanation and full documentation of the Mahmoud Khodier share transfer, plus a signed release from him. Non-negotiable, and everything else is moot until it is produced.
A current ACRA business profile (theirs is 18 months old) and current Egyptian commercial register extract.
Bank statements for both the Singapore and Egypt entities for the last 12 months — to establish what revenue actually landed in the company rather than in personal accounts.
Direct reference calls with RED and Shaghalni, arranged by MISK and not by xBites.
A written retraction or substantiation of the SODIC, Fawry and Mobica claims. Their answer to this question is more informative than any document they could supply.
Disclosure of the missing folder 05.
Evidence that they hold an Egyptian “EDM for others” licence — or a written acknowledgement that they do not. See §9; this is the fastest way to test whether they understand their own regulatory exposure.

Suggested reply to them — short, courteous, non-committal

“Thank you for sharing the materials — the product demo is genuinely impressive and the inventory database is a real asset. Before we can discuss any form of partnership or investment, we need a few things clarified: the current shareholding position and the documentation around Mr. Khodier’s exit; up-to-date bank statements for both entities; and confirmation of the current status of the SODIC, Fawry and Mobica relationships as presented in the folder. We would also want to speak directly with RED and Shaghalni as references. Once we have those, we can take a view.”

This is a fair, professional request. It costs MISK nothing, closes no doors, and their response to it will settle the matter faster than further analysis.

Methodology & sources All 78 files in the shared folder were extracted and read: 57 PDFs (text layer), 9 Excel workbooks (all sheets), 1 PowerPoint, 1 Word document, 4 payment screenshots (OCR), 4 Arabic registry documents (visual reading), and 3 demo videos (frame-sampled and read). PDF metadata was examined for every document. Financial figures were independently recomputed from source rather than taken from summary slides, and cross-checked between the internal ledger, the ETA e-invoice records and the payment evidence. Independent verification (§3) was conducted separately against public sources with no reliance on materials supplied by the company. Personal identifiers have been redacted.

Prepared by: MISK Communication · Performance Marketing & Growth · 2026-07-26