ChatGPT prompt: "A cartoon in the style of the New Yorker, showing an inventor obtaining a patent from a device which resembles an ATM". My suggested caption: "And all for only R1500!"
21 September 2026
Note: this became a two-parter. Part I is here. Last time, we saw that the large firms are in for what might be some fairly wrenching changes. Let's turn now to the small firms, and see if they are faring any better. Firstly, however: whenever I mention 'automation', what I mean is some combination of simple process automation, specialised scripts and software to create or populate documents, and AI. When I talk about 'AI', I mean both in terms of LLMs and the broader use of machine learning to spot patterns, organise data, control processes and generate documents.
Lesson 2: all the way down
Let's get into the mindset of the smaller IP practitioner. You have a lot of limitations versus the larger firms - you can't sell on brand-name recognition, can't charge astronomical fees and often can't offer full service (or have to contract out to do so). But you also have a bunch of advantages, chiefly lower overheads and greater flexibility. So while the existence of the large firm is that of the large tree which sucks up 80% of the resources (and looks mainly at other large trees as competition), the small firm is the gribbly little forest creature that can scuttle around in darkness and live off of leaf mulch. It's a more flattering analogy than it sounds like, I promise.
The smaller firms that I am aware of have, by and large, embraced automation and AI with extreme enthusiasm. This is partly because, being unable to charge high prices in any case, they must instead lower their cost of production. As an example from my professional background, and using representative rather than accurate numbers: when a client comes in with an invention, the patent attorney can expect to conduct at least two meetings, read anything from half-a-dozen to dozens of pages of text, perform an hour or so of background technical research,and then draft a four to fifty-page specification over the course of a few days (plus one to a dozen figures, which often need the services of a draughtsman). The spec then needs to be reviewed, edited, taken to client for input, edited again, finalised, and then filed as a provisional patent application. Let's conservatively call this 12 hours of professional time in total. Ignoring correspondence, filing fees, admin costs or review time by a senior partner, the drafting done by the associate is 'worth' something like R 30 000 - 60 000 at a large firm. The going rate for a provisional patent application in South Africa is, needless to say, not this much. Say R 15 000 - 45 000 depending on the reputation of the firm, the field of technology and how much material is going into the specification. Which is why larger firms rely on the associated follow-on work (complete application filings, foreign applications in the same patent ring, overseas examination and renewals) to actually make money off of the whole process. The small firms and sole practitioners can push closer to their own personal cost of living, bringing their personal time-based costs to R 10 000 -15 000. Which means that small firms can conceivably survive entirely on walk-in inventors and provisional/complete filings, albeit at almost zero margin.
Now, however, the small firm goes full-bore for automation and builds a complete pipeline from receiving invention disclosure documents all the way to filing. This can cut professional time down to something like the initial meeting/s and half an hour of review, with the proviso that the end product may end up being a bit one-size-fits-all compared to the fully manual version as written up by an experienced patent attorney with an appropriate technical background. This means that they can now either charge the going rate and pocket the difference in margin, or undercut somewhat and take in more of the market. The cumulative result of this is that the 'going rate' for patent filings is likely to shrink over the next few years to under R10 000.00 for a provisional filing, with large firms more or less only surviving on loyal legacy clients and those that are willing to pay multiple times over market rate for quality*. Or, alternatively, the large firms also go heavy on automation, cut staff and downsize their offices and facilities to slash overhead. But then they aren't large firms anymore.
This all sounds good for the little guys, but there's a sting in the tail. When discussing this issue, a friend and former colleague floated the idea that the patent office could just as easily set up its own pipeline. Locally, at least, the CIPC already assists inventors at the door. So this is either already happening behind the scenes, or else is just a matter of someone putting two and two together and then getting it to fit inside the existing filing process**. If a patent office does so, they can link the system up to also provide an initial inventiveness search report, as well as drafting narrow claims that the patent office feels could withstand examination (examiners themselves have been known to suggest claims, so this isn't unprecedented). Once done, you could end up going to the patent office, inputting your invention disclosure documents, filling out a few forms and, potentially, walking away the same day with a patent application studded with prima facie valid (but narrow) claims. If this happens, then a big chunk of the ecosystem goes away. Because now there is a floor price (potentially government-subsidised) to the patent application. Since the rule is that something 80% as good but for 10% of the price tends to drive out the rest of the market, this leaves both large and small firms fighting over a tiny pool of applicants willing to pay for 'premium' service, which might in any case be less than an ordinary patent application costs them right now.
So a pretty grim prediction, albeit one that interacts oddly with my understanding of how the long-delayed coming of patent examination into South Africa might play out (perhaps some other time). Until then, I expect that the little guys will do surprisingly well for themselves, for the simple reason that they can implement in a few months what the larger firms take years to do. The local firm to watch here (although I worry that invoking them will prove divisive) is, I think, S&Z. Who got onto the automation bandwagon early (albeit with a very specific use case) and have made great waves in the process. The next few years will reveal others.
* There's a secret third category of clients here: companies which have large patent portfolios, and which are willing to spread work around in order to make representing against them in litigation a bit more dicey due to lconflict of interest considerations. But these clients tend to be very rare.
** This is actually a bigger ask than I'm making it sound like. Not least because, as a matter of law in South Africa, a registered, practicing patent attorney currently needs to sign off on a complete patent application as part of the filing requirements. I think that this is also broadly the case overseas. But then there's nothing stopping a patent office from keeping a roster of registered patent attorneys around on a consultative basis to quickly do some review work (15 minutes each would do, with perhaps some consideration of handing the work out evenly so that nobody objects too hard). And rumour has it that the long-mooted Patent Bill will be bringing back patent agents, so there might be no shortage of warm bodies to fill the manual review slot in future, and potentially without needing to call in an outside firm.
Enquiries
Phone: +27 73 158 7995
Email: tthomas@schmidtconsult.co.za
LinkedIn: https://www.linkedin.com/in/thomas-schmidt-za/
All Rights Reserved Copyright 2026 Thomas Schmidt