One Search. More Sources.

When I started building AiForm Procure, one of the ideas behind it seemed relatively straightforward: South African businesses should not have to spend so much time searching for public procurement opportunities.

The opportunities exist. The problem is often finding them.

That idea became the basis of the first Procurement Notes article: South Africa doesn’t have a tender shortage. It has a tender discovery problem. But building the product has forced me to think more carefully about what “discovery” actually means.

Initially, it is tempting to think of the problem as aggregation. Find procurement notices from different places, bring them into one database, add a search box, and you have solved a meaningful part of the problem.

The further we build, the less convinced I am that aggregation alone is enough. Because the difficult part is not simply collecting tenders.

The hard part is making procurement information published by different institutions behave like one searchable system.

The supplier sees one market. The information does not.

From the perspective of a business, public procurement can look like one market. A construction company may be interested in work from a municipality, a national department, a development institution or another public entity. An IT company may have exactly the same problem. So might a training provider, manufacturer, engineering consultancy, cleaning company or equipment supplier.

Their fundamental question is remarkably simple:

Is anyone in the public sector currently buying something my business can provide?

The information needed to answer that question, however, does not necessarily begin in one place. Different institutions publish procurement information through different channels and structures. The language used to describe similar requirements can vary. Dates, categories, documents and statuses may be represented differently.

This means the supplier often has to understand the structure of the procurement landscape before they can even search it effectively.

That is backwards.

A capable supplier should not need to know which collection of portals and institutional websites to check before discovering that an opportunity exists. The discovery system should carry more of that burden.

Adding another source changed the problem

This became much more tangible while expanding the procurement sources feeding AiForm Procure. What began primarily around opportunities sourced through eTenders.gov.za has started expanding to additional institutional sources.

The current platform already demonstrates what happens when that source layer becomes broader. In a recent platform snapshot, AiForm Procure contained 2,149 open RFQ records, of which 149 were classified as live and accepting bids. Fifty-seven were closing within seven days and six had been recently posted.

But the number that interests me most is not 2,149. It is the variety behind it.

The same discovery interface is now surfacing opportunities from multiple procurement sources, including SANRAL, DBSA and municipal procurement sources. And the opportunities themselves can be remarkably different.

Recent examples have included the restoration and management of public art and heritage objects, professional end-user computer training, business intelligence and data analytics, railway maintenance, motorcycles and all-terrain vehicles, sanitation infrastructure, professional advisory services and major construction work.

That is when “more sources” stops being a simple ingestion problem. You have to decide how all of this information becomes useful.

More data can create a worse product

There is an uncomfortable truth about aggregation products:

Adding more information can make discovery worse.

Imagine taking thousands of procurement notices from across South Africa and placing them into a single enormous feed. Technically, you have centralised the information. Practically, you may simply have moved the fragmentation problem.

Instead of a supplier searching six different places, they now search one place containing thousands of poorly organised records.

The location changed. The cognitive burden did not.

This distinction matters enormously for AiForm Procure. I don't want success to be measured simply by how many tender records we can put into a database.

Of course, coverage matters. A discovery platform that misses relevant opportunities has an obvious limitation. But volume without structure quickly becomes noise.

The real objective is to make a growing pool of procurement information increasingly navigable.

That means thinking about how opportunities are categorised, how sources are identified, how closing dates are interpreted, how active and historical records are distinguished, how duplicates or related notices are handled, and how suppliers eventually move from broad searching towards relevance.

The question isn't only: How much procurement information can we collect?

It is: How quickly can a business move from all of that information to the small number of opportunities deserving its attention?

Standardisation is invisible when it works

There is a class of product work that users should barely notice.

If two procurement sources describe dates differently, the supplier should ideally still be able to filter by closing date. If institutions organise procurement into different categories, the supplier should still be able to search according to what their business does. If opportunities originate from different organisations, their provenance should remain clear even when they appear together.

The complexity belongs underneath the interface. The experience above it should become simpler.

This is one of the things building AiForm Procure is teaching me: normalisation is not merely a data-engineering concern. It is a user-experience problem. Every inconsistency that the platform can responsibly resolve is one less inconsistency the supplier has to interpret.

But there is an important boundary here. AiForm Procure should make procurement information easier to discover and navigate. It should not obscure where that information came from. The original procurement notice and its supporting documents remain authoritative for the supplier's decision.

Aggregation should reduce the journey to the source. It should never pretend to become the source.

Discovery is only the first decision

This connects directly with another principle we've been developing while building AiForm Procure:

A match is not a recommendation to bid.

If a tender contains words associated with your industry, that does not mean your business should pursue it. Finding the opportunity merely creates the possibility of a decision.

The supplier still has to establish whether it meets the mandatory requirements, whether it has the technical capability and supporting evidence, whether it can perform the work in the required location and timeframe, whether it has sufficient capacity, and whether pursuing the contract makes commercial sense.

That is why I increasingly think about procurement discovery as a funnel rather than a search result.

At the top might be thousands of public procurement records. Then come the opportunities that are actually open. Then the opportunities relevant to a supplier's industry, geography and capabilities. Then the opportunities for which that supplier appears eligible.

And eventually, perhaps, a very small number where the business can reasonably say:

This deserves our time.

That last step matters because bidding is not free. Preparing a serious tender response consumes time, staff capacity and sometimes money. A business that indiscriminately pursues every vaguely relevant tender can waste significant resources.

Better discovery should therefore not encourage more bidding for its own sake.

It should support better bid decisions.

One search is the beginning, not the destination

This is why “One search. More sources.” means something slightly different to me now than it did when we first started adding procurement sources.

At first, it described convenience. Instead of searching multiple places, search one. That remains valuable.

But the larger opportunity is not the search box. It is what becomes possible once fragmented procurement information can be organised through a common discovery layer.

A supplier should eventually be able to express something closer to:

This is what my business does. This is where we operate. These are the kinds of contracts we can realistically deliver.

And the system should do more of the work of identifying what deserves attention.

That is fundamentally different from expecting a supplier to know which institution might need their services, find that institution's procurement channel, understand how it publishes opportunities, construct the right search and repeat the process elsewhere.

The future I am interested in building moves more of that complexity away from the supplier.

Not by making the bid decision for them. By getting them to the right decision faster.

Where are you, and what does your business do?

There is another part of this problem that cannot be solved by better data pipelines alone. We need to understand the businesses on the other side of the search.

It is easy to sit behind a product and think in categories: construction, ICT, professional services, transport, catering, manufacturing.

Real businesses are more specific than that. A company may technically operate in construction but specialise in a particular type of work. A technology company may only provide certain services. A supplier may have the capability to deliver nationally, while another is commercially viable only within one province.

Those differences determine whether an opportunity is useful.

So as we continue building the source side of AiForm Procure, I want to spend more time understanding the demand side of discovery too. And that starts with two very simple questions:

Where are you? What does your business do?

Not which tender websites you already know. Not which procurement portals you remember to check every morning.

Your province. Your business. Your capabilities.

Because perhaps that is what procurement discovery should have started with all along. The supplier should tell the system who they are. The system should get better at finding the opportunity.

One search. More sources.

South Africa's public procurement landscape will not suddenly become less complex because we build another platform.

The institutions will remain different. Procurement requirements will remain contextual. Suppliers will still need to read the documents, understand the requirements and decide whether an opportunity makes commercial sense.

That complexity is real. But finding the opportunity does not need to carry all of it.

If we can bring more procurement sources into a common discovery layer, structure that information responsibly, preserve the path back to the original source, and increasingly understand what individual businesses are actually capable of doing, then we can remove some unnecessary friction from the beginning of the process.

That is what we're building towards with AiForm Procure.

Not every tender. Not more bids for the sake of more bids.

A clearer path from fragmented information to relevant opportunity.

One search. More sources. Better decisions about what comes next.

AiForm Procure

Find more. Search less.