Trust.

isaga · a probability exercise

Every company has an About Us page. Who we are, what we do. But honestly, why should you even care about that? Plus, the objective of such pages is to help customers build trust.

The paradox

Trust is a paradox.

You can only really trust us once you work with us. But if you have worked with us, we know you trust us and you like us. We've interacted in so many ways. There's no way you're reading this page.

So there's a decent probability that you don't work with us yet. And that's exactly why you're here. You're doing your due diligence.

Here's the thing. We can give you 5 reasons to trust us. It's still not going to work. No amount of us telling you is going to make you trust us. Whether you trust us or not is not something we can control.

But we still have to write an About Us page. So why just add photos and a company profile? Let's get creative.

See, we're engineers too, and you're one of the logical people out there. That's why you're doing due diligence in the first place. So let's add some math to it.

What is the probability that you should trust us?

The math we'll use: Bayes' theorem

Bayes' theorem tells you how to update a belief when new evidence comes in.

The easier way to run it is one piece of evidence at a time:

The likelihood ratio (LR) is simple. How much more likely is this evidence if we're trustworthy, compared to if we're not? An LR of 1 means the evidence tells you nothing. An LR of 2 means it's twice as likely for a trustworthy company.

To move between % and odds, you only need two lines:

• Odds = Trust ÷ (100% − Trust). So 25% trust is 25% ÷ 75% = 0.33.
• Trust = Odds ÷ (1 + Odds). So odds of 0.40 is 0.40 ÷ 1.40 = 29%.

And an LR in % terms is just how much it lifts the odds. An LR of 1.2 lifts the odds by 20%. An LR of 2 doubles them, a 100% lift.

Let's start.

Where we start: 25%

To be simplistic, let's say the probability that you can trust us is 50%. A coin toss.

But we'll borrow an idea from Warren Buffett and Charlie Munger: margin of safety. We take a margin of safety of 50%. So we start at 25%. It's always good to be cautious and not trust strangers :)

01 Who we are

We're a group of nerds from IITs, BITS, NITs and IIMs. This only means we are not dumb.

It could be a fluke that all of us had a great exam day. That's why we're giving this only a small weight. An LR of 1.2 means we assume it improves the odds by just 20% over the base rate.

02 The banks

Our group companies (gonuclei.com) work with all the banks in India. Pick your choice.

HDFC? Yes. ICICI? Yes. IDFC? Yes. HSBC, Kotak, AU Bank? Yes.

And not just private sector banks. Public sector banks as well. SBI, the biggest of them. Union Bank, Canara Bank, and the list goes on and on.

Every one of these banks has done due diligence on us.

03 Our partners

Our companies' partners include the likes of Mastercard, Visa, AWS, Google Cloud and Infosys.

They have all done their due diligence on us too. Having reached this scale, working with the likes of these, means one thing. We are not a fraud, for sure.

04 Deutsche Telekom

Across our group companies, we launched in Europe with Deutsche Telekom, the largest telecom operator in Europe.

Do you think that's easy? Your first customer, and it's the largest telco. Of course not.

But we've already talked about enterprises, so this one only gets a small weight.

05 Companies like yours

Again, we've talked about large conglomerates. There's a good chance you can't relate to them at all, because your business is very different from theirs.

So let us give you more examples. Across our group companies, we work with listed companies like Mamaearth. Spirituality companies like AppsForBharat. EV companies like Ultraviolette. Agri-tech companies like Fasal. Marketplaces like DriveX. Health tech companies like HealthPlix. Fintechs like FinBox.

And this is all in the last 6 months.

06 This page itself

Again, the fact that we're writing this page differently says something. Not every company writes like this. We're pretty sure you won't find it anywhere else.

It shows we're serious about this. We want to work hard, but at the same time, we want to have fun. And it shows we're decently smart.

And we're proud of trying something new. Steve Jobs talked about wanting to put a dent in the universe. This is our small dent: trying it our way. So this one gets a bigger weight.

The scorecard

A quick note for the engineers. The LRs are our assumptions. Change them if you like. We've also treated each variable as independent, which isn't fully true, so we kept the numbers modest. And no amount of evidence gets you to 100%. That's just how the math works. One more thing: we carry the unrounded odds from step to step, so a rounded % may look off by a hair.

So, should you trust us?

You get the logic. Directionally, we can keep adding variables and the number will keep going up.

But we'll still say: don't trust us.

Try us instead. Experience us. We're not going to ask you to commit to anything. You're free to choose. Trust comes from repetition, after you've experienced us.

We stand behind what we say

To make this matter of trust more serious, we stand behind what we say.

We're not just going to come and recommend that you use a Savings Plan. If there's a loss because of underutilization, we will cover that loss for you. We call it our Loss Coverage, No Risk Policy, and it's part of our agreement.

So we're not just saying. We're committing to you.

One caution

Does this mean we work with just about any company?

We also do our due diligence on every company before we decide to work with them.

We want to build a community where people have fun, but at the same time, share knowledge. We run a book club. We use mental models a lot internally.

We will never go after the masses or try to be the largest company. We believe in running a company that's fun to work with, but also delivers significant value to the customer.

Don't take our word for it. Ask Claude.

Our group companies are isaga and Nuclei. Every fact on this page should be verifiable independently.

And we know you. You're engineers and CTOs. We know you still trust Claude more than you trust us. Fair enough.

So to exceed your expectations, here's a prompt we've written for you. Copy it, paste it into Claude, and verify everything above yourself. Go ahead and ask if the math checks out directionally.

I'm evaluating isaga and Nuclei as possible vendors. Search the internet and open the pages yourself. Don't answer from memory. Give me a link for every finding. Where a page on isaga.ai supports a point, link to it.

The two companies
- isaga: https://isaga.ai
  Legal name: ThinkSaaS Technologies Private Limited (CIN U47413KA2025PTC205621)
- Nuclei: https://www.gonuclei.com
  Legal name: CDNA Technologies Private Limited (CIN U72200KA2018PTC152932)

How to grade each claim
- Verified: a third party confirms it (a bank, partner, telecom operator, company registry, news outlet or official directory). For a person's role or education, their own LinkedIn or a profile site such as The Org is enough.
- Partly verified: only the company's own website shows it.
- Could not verify: you found nothing.

1. Start with their websites
Visit both websites. In a few lines, compare what each company does, who it serves, and how the two are connected.

2. Are they one group?
Look up both legal names on Tofler, Zauba Corp or IndiaFilings. Check for common directors, a shared registered address, and shareholding, where it is public.

3. The people
Find the directors of both companies, plus the CEO, the CTO, the Chief of Staff, the head of product, the leader of the SaaS business, and senior engineering and security leaders. Use LinkedIn, The Org, company records and news. For each person, give their name, role, and where they studied (engineering college and MBA, if any). Then check this claim: the team includes alumni of IITs, BITS, NITs and IIMs.

4. The banks
Check HDFC Bank, ICICI Bank, IDFC FIRST Bank, HSBC, Kotak, AU Small Finance Bank, SBI, Union Bank and Canara Bank. Search each bank's own website and PDFs for "Nuclei", including product pages, terms, brochures and procurement documents. Also read the bank case studies on the Nuclei blog. Give each bank its own verdict.

5. The partners
Check each one and give each its own verdict:
- Mastercard: the Mastercard newsroom
- Visa: https://partner.visa.com/site/partner-directory/nuclei.html
- Infosys: Infosys Finacle news
- AWS: https://partners.amazonaws.com/partners/001aq00000EGCBeAAP/
- Google Cloud: https://cloud.google.com/find-a-partner/partner/thinksaas-technologies-private-limited

6. Deutsche Telekom
Check whether Nuclei launched in Europe with Deutsche Telekom. Look at the Magenta Moments gift cards from T-Mobile Czech Republic, T-Mobile Poland and Magyar Telekom, and at Nuclei's European entity.

7. The customers
Check Mamaearth, AppsForBharat, Ultraviolette, Fasal, DriveX, HealthPlix and FinBox. Start with the logos and testimonials on https://isaga.ai/aws-services.

Put points 2 to 7 in one table: claim, verdict, link. Use one row per point, plus one row per person in point 3. In the bank, partner and customer rows, list each name with its own verdict. Keep it to 15 rows or fewer.

8. The Bayes' theorem math
Their Trust page uses Bayes' theorem to estimate whether you should trust them. It starts at a 25% prior and applies likelihood ratios of 1.2, 2, 1.5, 1.1, 1.2 and 1.5 in odds form, ending at about 70%. Of course, I understand the likelihood ratios are their own assumptions. This is an illustration, not a measured score, so I'm not asking for a statistics audit. Tell me two things: is the arithmetic right, and is the logic directionally right?

9. Your take
Is this a clever way to write a trust page? Why or why not? How does it compare with a typical About Us page? End with a 3-line verdict on whether isaga is worth a first conversation.

isaga · ThinkSaaS Technologies Private Limited